Do Share Repurchases Crowd Out Internal Investment in South Africa?

Abstract

Researchers in developed countries have questioned whether share repurchase activity influences internal investment. The aim of this study was to investigate the relationship between share repurchases and internal investment (defined as capital expenditure, employment expenditure, and research and development) in South Africa, as little was known about this relationship in developing countries. A quantitative research methodology was followed, employing the data of South African listed companies during the 2002–2017 period. A significant negative relationship was noted between share repurchases and employment expenditure when considering all companies, while high-growth companies exhibited a significant negative relationship between share repurchases and capital expenditure. The negative relationships could indicate that companies increase share repurchases to the detriment of internal investment (especially employment). Alternatively, it may imply that share repurchase and internal investment decisions are determined simultaneously, with companies decreasing internal investment and increasing share repurchases in the absence of identifiable profitable projects (or increasing internal investment and decreasing share repurchases when growth opportunities are available). These findings could be useful to shareholders, corporate governance regulators and activists. Given the high unemployment and income inequality in South Africa, the results support a call for the improved regulation of share repurchases to ensure effective monitoring.

1. Introduction

Globally, researchers are questioning whether share repurchase activity curtails internal investment (Almeida et al. 2016; Wang et al. 2021). Several earlier studies conducted in developed countries have reported a negative relationship between share repurchases and internal investment (Almeida et al. 2016; Apergis et al. 2021; Chasiotis and Georgantopoulos 2022; Iyer et al. 2017; Wang et al. 2021). In South Africa, it is especially important to study the relationship between share repurchases and internal investment in view of the high unemployment, high income inequality, and low economic growth in the country, which could be alleviated by higher internal investment. Thus, the aim of the present study was to investigate the relationship between share repurchases and internal investment (defined as capital expenditure, employment expenditure, and research and development) in South Africa to determine whether the additional regulation and disclosure of share repurchases are needed in the South African context.

Share repurchase activity has escalated in the last two decades. Share repurchases are often used to increase shareholder return when a company has excess cash flow available or to signal undervaluation to the market. However, share repurchases can also be employed by executive directors (executives) to increase the company’s share price and earnings per share (EPS) figure (Wang et al. 2021). As the variable portion of executive remuneration is often dependent on share price and EPS metrics, share repurchases can be abused by executives in a bid to increase their remuneration. Share repurchases often benefit the wealthiest families in an economy (including those of executives) as well as institutional investors (like hedge funds) (Lazonick 2014) to the detriment of other company stakeholders and reduce corporate sustainability (Vaupel et al. 2022). As a result of share repurchase activity, less money might be available for investing in employees, capital projects, innovation, and sustainability initiatives (Almeida et al. 2016; Lazonick 2014). The aforementioned initiatives are important to all company stakeholders, including employees and customers, and enhance corporate sustainability.

Because of South Africa’s significant economic disparity, which already leads to serious societal issues and puts pressure on government expenditure (Hundenborn et al. 2019), it is essential that share repurchase activity be carefully watched. From its legalisation in 1999, the share repurchase activity of South African listed companies showed a steady increase, reaching its highest peak during the global financial crisis (2007–2009) (Wesson et al. 2015). After the financial crisis, the amount spent on share repurchases decreased to a lower level and remained relatively constant during the period 2010–2017, although a larger number of companies were engaging in share repurchases (Steenkamp and Wesson 2020). During the 2000–2015 period, South African listed companies spent an average of 14% of their profit on share repurchase activity, which corresponds closely to the 19% average reported for European companies during the same period (Steenkamp and Wesson 2020).

South Africa has a dual economy, characterised by high unemployment, high income inequality, and low economic growth on the one hand and a sophisticated stock exchange and high-quality corporate governance regulations on the other hand (Wesson et al. 2018). When specifically considering share repurchases, however, South African regulations trail behind those of many other countries (Steenkamp and Wesson 2020). The Johannesburg Stock Exchange (JSE) requires general (open-market) share repurchases to be announced only once 3% of outstanding shares have been repurchased (Wesson et al. 2018). This decreases the transparency of South African share repurchases and leads to some share repurchases never being announced (Steenkamp and Wesson 2020). Owing to the unique regulatory environment, a complete record of South African share repurchase activity is not available on financial databases. However, Steenkamp et al. (2023) recently constructed a database containing a complete record of share repurchase activity for South African listed companies for the period 2002–2017 by hand-collecting data from annual financial statements and then performing reconciliations to calculate the amount spent on share repurchases. With complete South African share repurchase data now being available for the 2002–2017 period, the present study investigated the effect of share repurchases on internal investment during this period.

The results of the present study demonstrated that increased share repurchases were associated with decreased employment expenditure (or smaller increases in employment expenditure). For companies with higher levels of available growth opportunities, a negative relationship was found between share repurchases and both capital expenditure and employment expenditure. The inverse relationship identified could indicate that share repurchases are crowding out internal investment. Alternatively, it could imply that companies decrease internal investment when they do not have profitable projects available, and then simultaneously increase share repurchases to distribute the free cash flow to shareholders. The results support a call for improved corporate governance measures to enable the continuous monitoring of share repurchase activity in South Africa. Specifically, it is recommended that the JSE amends its announcement rule for general share repurchases to require more timeous announcements of all share repurchase activities.

The remainder of the paper is structured as follows. First, a review of the prior literature is provided, which leads to the development of two hypotheses. Second, the methodology employed is explained. Third, the results are presented, after which conclusions are drawn and recommendations are made.

2. Literature Review and Hypothesis Development

The literature review, firstly, focuses on share repurchase activity and its interaction with internal investment. Next, the share repurchase landscape in South Africa is considered. The literature review concludes by discussing previous global and local studies on the relationship between share repurchases and internal investment and articulating the hypotheses to be tested in the present study.

2.1. Share Repurchase Activity and Its Interaction with Internal Investment

A share repurchase occurs when a company buys back the shares it previously issued (Steenkamp and Wesson 2020). Share repurchases are carried out for various reasons, including to signal undervaluation to the market, to fend off a takeover threat, to offset the shares issued to executives and other employees as part of their share-based remuneration, and influence metrics such as the share price and EPS (Dhanani 2016; Dittmar 2000; Wesson et al. 2018). However, one of the primary reasons for executing a share repurchase is that it facilitates the return of excess cash to shareholders, in line with the free cash flow hypothesis of Jensen (1986). According to the free cash flow hypothesis (Jensen 1986), a company should distribute the cash earned during a certain financial period to shareholders (through paying dividends or repurchasing shares) after the company has invested internally (in employees, capital outlay or research and development). A company would be engaged in ‘empire-building’ and vulnerable to exploitation by executives if it kept large quantities of cash on hand (Cook and Zhang 2022; Dhanani 2016), and thus companies with high free cash flow should distribute dividends or repurchase shares to mitigate agency risk (Cook and Zhang 2022; Wesson et al. 2015).

Lazonick (2014) argued that dividends, when compared to share repurchases, are a more ethical way to distribute free cash flow to shareholders as dividends reward shareholders for holding (rather than selling) their shares and benefit all shareholders equally. However, share repurchases have begun to take the role of dividend increases (Grullon and Michaely 2004; Wesson et al. 2015). Many companies prefer not to raise dividend levels since this creates an expectation that the raised level will be maintained and because the market reacts unfavourably when companies reduce their dividends (Andriosopoulos and Hoque 2013). Companies therefore frequently prefer to maintain dividend levels while adopting share repurchases as a more flexible form of distribution (Apergis et al. 2021; Brav et al. 2005; Jagannathan et al. 2000). The aforementioned explains the increased popularity of share repurchases globally.

In line with the free cash flow hypothesis, share repurchases may impact the cash available for internal investment (Chao and Huang 2022). Because of the large amounts being spent on share repurchases (especially in the United States), stakeholders have recently grown concerned that companies might be overspending on share repurchases while retaining too little for internal investment (Almeida et al. 2016; Steenkamp and Wesson 2020; Schneider and Kohlmeyer 2015). Commentators are worried that executives are not considering all available investment opportunities but are instead buying back shares to boost the share price and EPS while also increasing their personal compensation (Lazonick 2014). Internal investment can generate profits over a long period of time, whereas share repurchases only result in a single boost to the share price and EPS metrics. Excessive share repurchase activity could thus negatively affect the long-term financial health of a company.

While the previous paragraph considered the financial effects of share repurchases, Vaupel et al. (2022) further explained that share repurchases could affect the social and environmental sustainability of companies. Share repurchases could maximise shareholder value to the detriment of environmental and social interests; as such, the ethics of engaging in share repurchases should also be considered (Vaupel et al. 2022). Bassier and Woolard (2021) found that income inequality in South Africa had increased after the global financial crisis of 2007–2009, partially because of large returns on capital investments by the wealthy. Share repurchasing is a financial tool that could enrich the wealthy at the expense of other company stakeholders (Lazonick 2014). As a result, it has become imperative that research in South Africa focuses on the impact of share repurchases on internal investment (especially since internal investment affects all company stakeholders, including employees and customers). Research findings could inform corporate governance policies relating to share repurchases, which could include improved disclosure and increased transparency. Currently, South African regulations do not allow for the real-time monitoring of the share repurchase activity of listed companies—as will be explained in the following section.

2.2. The Share Repurchase Landscape in South Africa

Share repurchases were legalised in South Africa in 1999 (Wesson et al. 2015). Over the period 2000–2009, the number of companies engaging in share repurchases, the number of shares repurchased and the amount spent on share repurchases showed a steady increase (Wesson et al. 2015). After reaching a high point during the global financial crisis of 2007–2009 (share repurchase value was almost 45% of company profit in 2009), the amount spent on share repurchases dropped to a lower level (ranging between 4% and 14% of profit during 2010 to 2017) (Steenkamp and Wesson 2020).

Shares of South African holding companies may be repurchased by the holding company itself and/or its subsidiaries. Shares repurchased by subsidiaries are not cancelled but instead are recorded as treasury shares, in contrast to shares repurchased by the holding company, which are cancelled and return to authorised share capital (Steenkamp and Wesson 2020). Subsidiaries were the preferred repurchasing entity between 2000 and 2017 (Steenkamp and Wesson 2020; Wesson et al. 2015). Because treasury shares can be resold, used in business combinations, or distributed to employees at a later date, they provide greater flexibility than other types of share repurchases, which may have contributed to the preference for repurchasing shares through subsidiaries (Steenkamp and Wesson 2020).

For South African listed companies, share repurchases fall into one of two categories: general (i.e., carried out on the open market at the prevailing market price) or specific (i.e., the repurchase of a specific number of shares from certain shareholders at a predetermined price) (Vermeulen 2014). In contrast to worldwide trends, general (open market) share repurchases were not South Africa’s favoured form of share repurchase (Wesson et al. 2015). Based on the amount spent, specific repurchases accounted for more than 57% of all share repurchases between 2000 and 2009 (Wesson et al. 2015) and 62% between 2010 and 2017 (Steenkamp and Wesson 2020).

Specific repurchases need to be announced via the JSE’s stock exchange news service (SENS) as soon as the terms have been agreed upon (JSE 2017), providing adequate information to enable monitoring. General repurchases in South Africa, however, only need to be announced on SENS once more than 3% of the outstanding shares of that class have been bought back cumulatively (JSE 2017). Because of the more forgiving and simpler rules for general repurchases, certain share repurchases (those that total less than 3% cumulatively) are not announced via SENS (Steenkamp and Wesson 2020). SENS announcements therefore do not provide a comprehensive real-time account of every general share repurchase carried out by listed companies in South Africa (Steenkamp and Wesson 2020). Prior research found that a substantial percentage of general share repurchases are not announced via SENS (Steenkamp and Wesson 2020; Vermeulen 2014; Wesson et al. 2015). During the 2000–2009 period, 41% of general repurchases were not announced on SENS, while 79% of general share repurchases were not announced during the 2010–2017 period—illustrating that the transparency of general share repurchases had decreased over time. The current announcement rules in South Africa differ from those employed by most other countries with sophisticated stock exchanges (where daily, weekly, monthly, or quarterly announcements relating to the actual number and value of shares repurchased are the norm), leading to a lack of transparency relating to share repurchase activity in the South African environment (Steenkamp and Wesson 2020).

The IRESS financial database, which is commonly used by South African researchers to gather financial data, also does not provide a comprehensive record of the share repurchase activity of South African listed companies (Steenkamp and Wesson 2020). This is probably the result of inconsistent disclosure in annual financial statements, given the unique South African regulatory environment relating to share repurchases—which is not catered for by the reporting requirements contained in the International Financial Reporting Standards. Steenkamp and Wesson (2020) found that, during the 2010–2017 period, the share repurchase value reported by IRESS only constituted 66% of the actual net repurchase value (represented by share repurchases excluding intragroup repurchases where the holding company repurchases treasury shares) and 44% of the actual gross repurchase value.

This section explained the unique regulatory requirements pertaining to South African share repurchases. The resultant lack of transparency, caused by the JSE’s announcement rules and inconsistent disclosure in annual financial statements, emphasises the importance of studying the relationship between share repurchases and internal investment. If such a relationship exists, it would support a call for improved corporate governance relating to share repurchases in South Africa, including the real-time announcement of share repurchases on SENS and improved disclosure requirements in annual financial statements.

2.3. The Relationship between Share Repurchases and Internal Investment

Several global studies have reported a significant negative relationship between share repurchases and certain internal investment variables. Using data from the United States (US) from 1996 to 2005, Bhargava (2013) found that companies that spend more on share repurchases invest less in research and development, and in long-term investment instruments. Almeida et al. (2016) reported a negative relationship between share repurchases and capital expenditure, employment expenditure, and research and development, respectively, when considering US data from the 1988–2010 period. Focusing on the 2006–2015 period in the US, VanDalsem (2019) also found a negative relationship between share repurchases and both employment expenditure and research and development. The negative relationship was, however, less pronounced for family-owned companies, where shareholders might be more focused on long-term value creation (VanDalsem 2019). Iyer et al. (2017), using data from the US from 1971 to 2014, found that share repurchases and capital expenditure were negatively related. Similarly, a negative relationship between share repurchases and capital expenditure was noted using Japanese data from 2000 to 2019 (Apergis et al. 2021) and data from the United Kingdom during the period 1999–2019 (Chasiotis and Georgantopoulos 2022). Wang et al. (2021) employed global data, excluding data from the US, and similarly reported a negative relationship between share repurchases and capital expenditure, and research and development. Using US data from 1996 to 2005, Swift (2022) found a negative relationship between share repurchases and research and development.

The three separate dependent variables (capital expenditure, employment expenditure, and research and development) in the Almeida et al. (2016) study were measured as the change from four quarters before the share repurchase to four quarters after the repurchase, scaled by assets—which was considered appropriate for the US environment, given that US companies report on a quarterly basis and because the US had a relatively low inflation rate during the period of study. Wesson and Botha (2019) replicated the study of Almeida et al. (2016) using South African data from 2000 to 2009. Contrary to global results, Wesson and Botha (2019) found no significant relationship between share repurchases and capital expenditure nor between share repurchases and employment expenditure. However, a positive relationship was found between share repurchases and research and development in South Africa.

As South African financial data are only available annually, Wesson and Botha (2019) adapted the method of Almeida et al. (2016) and measured the three dependent variables (the internal investment variables) as the change from one year prior to the share repurchase to one year after the repurchase, scaled by assets. However, the inflation rate in South Africa during the 2000–2009 period (calculated at 5.4% using data from the Organisation for Economic Co-operation and Development (OECD)) was substantially higher than in the US (calculated at 2.6%) (OECD 2023). As a result of the higher inflation in the South African environment, the measurement employed by Wesson and Botha (2019) might have positively biased the coefficients reported in their study. In the South African context, therefore, an alternative method of measuring the dependent variables for internal investment should be employed when investigating the relationship between share repurchases and internal investment.

Share repurchases could also affect the sustainability of companies (Vaupel et al. 2022). Vaupel et al. (2022) studied the interaction between share repurchases and companies’ performance relating to corporate sustainability, including social sustainability. Social sustainability performance is a measure that considers a company’s value-creation process in respect to stakeholders other than shareholders—including employees and customers (Vaupel et al. 2022). Vaupel et al. (2022) found that increased share repurchases were associated with lower long-run social sustainability performance in the years after the repurchase. Share repurchases could be utilised as a financial tool to enhance financial metrics and increase executive remuneration (benefitting shareholders and executives), while reducing the free cash flow available for investing in employees (Lazonick 2014).

After considering prior research on the relationship between share repurchases and internal investment, it was evident that no South African research on this topic was published after 2009. Furthermore, the results of Wesson and Botha (2019) pertaining to the 2000–2009 period needed to be reconsidered using a more appropriate measurement base for the three internal investment variables. Based on prior global research, the following hypothesis was developed:

Hypothesis 1 (H1).

There is a negative relationship between share repurchases and internal investment variables (capital expenditure, employment expenditure, and research and development) in the South African context.

The relationship between share repurchases and internal investment could be influenced by the growth and investment opportunities (hereafter, growth opportunities) available to a company. One of the most commonly used and information-rich proxies for growth opportunities is the market-to-book ratio (Adam and Goyal 2008). Companies with high growth opportunities and wanting to increase investment spending while having limited resources, are more likely to reduce share repurchases to channel cash towards internal investment (Apergis et al. 2021). It is thus expected that companies with high growth opportunities would exhibit a stronger negative relationship between share repurchases and internal investment than low-growth companies, and the following hypothesis was articulated:

Hypothesis 2 (H2).

The negative relationship between share repurchases and internal investment is more pronounced for companies with high growth opportunities.

3. Methodology

A quantitative research methodology, associated with a positivist paradigm, was followed. Ordinary least squares regression analyses were carried out using secondary data relating to South African listed companies. The regressions employed fixed effects for year and company and robust standard errors to compensate for heteroscedasticity (clustered by company).

The target period (2002–2017) was chosen to match the share repurchase database created by Steenkamp et al. (2023) to ensure that complete and accurate data on South African share repurchase activity were employed. As explained in the literature review, financial databases do not contain accurate and complete data on South African share repurchases, and thus, the end date of the target period could not be extended beyond 2017. The relationship between share repurchases and internal investment in South Africa had previously only been studied using data from before 2010, and the present study thus expands on this. Moreover, it was decided that the 2002–2009 period should be reconsidered since Wesson and Botha (2019) had employed a possibly inappropriate measurement base for the dependent variables in their study. The population examined in this study was also aligned to the Steenkamp et al. (2023) database—that is, all companies with primary listings on the main board of the JSE during the target period were included, but companies in the financial services and basic resources industries were excluded. Companies were only included in the Steenkamp et al. (2023) database, and thus in the present study, if they had been listed for at least three years during the target period.

3.1. Methods to Test Hypothesis H1

The regression analyses to test hypothesis H1 employed the methodology of Almeida et al. (2016) and Wesson and Botha (2019), adjusted to measure the dependent variables more appropriately, taking into account the higher inflation rate in South Africa. Moreover, while Wesson and Botha (2019) employed only a limited number of control variables, the present study included additional control variables—as determined from prior studies on the relationship between share repurchases and internal investment (Almeida et al. 2016; Apergis et al. 2021; Chasiotis and Georgantopoulos 2022; Iyer et al. 2017; Wesson and Botha 2019).

To determine whether share repurchases were associated with internal investment, the dependent variable was identified as internal investment. In line with Almeida et al. (2016) and Wesson and Botha (2019), three separate dependent variables relating to internal investment were selected: capital expenditure, employment expenditure, and research and development. Each dependent variable was calculated as the difference in the value of the specific internal investment variable between the current year and the prior year, scaled by assets in the current year. This alternative measurement base was similar to the one applied by Almeida et al. (2016) and Wesson and Botha (2019), although it was adapted to take into account South Africa’s substantial inflation rate. While Wesson and Botha (2019) calculated the dependent variable as the change in the value of internal investment from one year prior to the share repurchase to one year after the share repurchase (i.e., a two-year interval), the measurement employed by the present study employed only a one-year interval to minimise the effect of inflation while still accommodating the dynamic nature of the dataset. The dataset was seen as dynamic as internal investment in a certain year is influenced by the prior year’s internal investment level.

Given that three separate dependent variables were identified, three regression models were specified (with identical independent and control variables). The model specifications for testing Hypothesis H1 were as follows:

D_CAPEXi,t = β0 + β1SRi,t + β2DIVi,t + β3SIZEi,t + β4CFi,t + β5LEVi,t +
β6TANGi,t + β7GRi,t + β8PRi,t + εi,t

(1)

D_EMPi,t = β0 + β1SRi,t + β2DIVi,t + β3SIZEi,t + β4CFi,t + β5LEVi,t +
β6TANGi,t + β7GRi,t + β8PRi,t + εi,t

(2)

D_RDi,t = β0 + β1SRi,t + β2DIVi,t + β3SIZEi,t + β4CFi,t + β5LEVi,t +
β6TANGi,t + β7GRi,t + β8PRi,t + εi,t

(3)

Table 1 shows the measurement of all variables. All data, except the data on share repurchases, were extracted from the IRESS financial database. Share repurchase data were obtained from the Steenkamp et al. (2023) database.

Table 1. Measurement of variables employed in the study.

Variables Employed	Measurement Using South African Data
Dependent variables
Capital expenditure (D_CAPEX)	Difference in capital expenditure (fixed assets acquired from the cash flow statement) between the current year and the prior year, scaled by assets in the current year
Employment expenditure (D_EMP)	Difference in staff costs (excluding directors’ remuneration) between the current year and the prior year, scaled by assets in the current year
Research and development (D_RD)	Difference in research and development expenditure between the current year and the prior year, scaled by assets in the current year
Independent variable
Share repurchases (SR)	Net amount spent on share repurchases (excluding intragroup repurchases), scaled by assets
Control variables
Dividends (DIV)	Ordinary dividends, scaled by assets
Company size (SIZE)	Natural logarithm of assets
Cash flow (CF)	Cash flow from operations, scaled by assets
Leverage (LEV)	Debt-to-assets ratio
Tangibility (TANG)	Fixed assets, scaled by assets
Growth opportunities (GR)	Price-to-book ratio
Profitability (PR)	Return on assets percentage

3.2. Methods to Test Hypothesis H2

The regression analyses to test hypothesis H2 were based on the methodology employed by Apergis et al. (2021), using the same control variables as for hypothesis H1. To evaluate the impact of share repurchases on internal investment under both conditions of low- and high-growth opportunities, the median value for growth opportunities (proxied by the price-to-book ratio) in the dataset was calculated as 1.60695. All line items (company years) in which the company’s price-to-book ratio was larger than 1.60695 were labelled as high-growth (1), while the remainder were labelled as low-growth (0).

Regressions were executed in two separate but coinciding ways. First, a binary value (1 or 0) relating to high growth or low growth was employed as an interaction variable with share repurchases to identify whether the existence of growth opportunities affected the nature and strength of the relationship between share repurchases and internal investment. Second, to better understand the interaction between share repurchases and internal investment for high-growth and low-growth companies separately, the full dataset was divided between those company years (line items) that were deemed high-growth (where the price-to-book ratio exceeded 1.60695) and those that were deemed low-growth. The regressions executed for hypothesis H1 were then re-executed, but for the high-growth and low-growth subsets separately.

4. Results

The share repurchase dataset from Steenkamp et al. (2023) contained 2392 line items (company years) pertaining to the period 2002–2017. For 23 of these line items, some of the control variables were not available in the IRESS financial database, which led to them being removed from the dataset for the purposes of the present study. The control variables for the remaining 2369 line items were scrutinised for any anomalies, and in the process another 31 line items, which contained control variables that seemed incorrectly captured by IRESS, were also removed from the dataset. Examples of such irregularities were negative price-to-book ratios (25 line items), price-to-book ratios larger than 50 (2 line items), negative dividends (2 line items) and extremely high or low return on assets (2 line items). The remaining dataset thus contained 2338 line items, which were employed in the present study. By calculating the dependent variables as the difference between the current and prior year, each company’s earliest year included in the study was lost (233 missing values), leaving 2105 line items to be employed in the present study pertaining to the dependent variables. The descriptive statistics related to the dataset employed may be seen in Table 2.

Table 2. Descriptive statistics relating to the variables employed in the study.

	Number of Observations	Mean	Standard Deviation
Dependent variables
Capital expenditure	2105	0.005	0.049
Employment expenditure	2105	0.033	0.106
Research and development	2105	0.000	0.007
Independent variable
Share repurchases	2338	0.007	0.027
Control variables
Dividends	2338	0.042	0.328
Company size	2338	6.208	0.803
Cash flow	2338	0.152	0.105
Leverage	2338	0.516	0.236
Tangibility	2338	0.248	0.208
Growth opportunities	2338	2.287	2.693
Profitability	2338	0.133	0.168

From the mean values of the dependent variables, an increase in capital expenditure and employment expenditure was noted from one year to the next. This made sense given the relatively high inflation rate in South Africa and agrees with the results of Wesson and Botha (2019). The research and development figure remained relatively constant from one year to the next, but this was primarily caused by the fact that 1883 (of 2338) of the observations were zero. The high number of observations equal to zero could indicate that research and development expenditure was not well-reported in the annual financial statements of South African listed companies, and thus not well-captured by the IRESS financial database. Table 3 shows the pairwise correlations of the variables employed.

Table 3. Pairwise correlations of variables employed in the study.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)
(1) Capital expenditure	1.00
(2) Employment expenditure	0.14	1.00
(3) Research and development	−0.03	−0.09	1.00
(4) Share repurchases	0.02	0.01	0.00	1.00
(5) Dividends	−0.03	−0.03	0.00	0.01	1.00
(6) Company size	0.02	−0.10	−0.00	−0.04	−0.03	1.00
(7) Cash flow	0.14	0.17	0.01	0.16	0.01	0.07	1.00
(8) Leverage	0.03	0.06	−0.00	−0.02	−0.10	0.15	−0.11	1.00
(9) Tangibility	0.08	−0.01	0.00	−0.06	−0.04	0.14	0.17	−0.13	1.00
(10) Growth opportunities	0.07	0.08	0.04	0.06	0.09	0.15	0.27	0.16	0.07	1.00
(11) Profitability	0.07	0.15	−0.00	0.08	0.10	0.03	0.66	−0.09	−0.02	0.34	1.00

Given the large correlation between cash flow and profitability, it was decided to exclude profitability from the regressions that follow to reduce the potential effect of multicollinearity. The two hypotheses are tested and addressed separately in the sections that follow.

4.1. The Relationship between Share Repurchases and Internal Investment

In this section, the results of the regression analyses are reported separately for the entire period under study, namely 2002–2017, and for the period that was previously reported on by Wesson and Botha (2019), namely 2002–2009. Table 4 reports the results pertaining to the entire period under study (2002–2017). The regressions pertaining to the relevant dependent variables—capital expenditure, employment expenditure, and research and development—are listed as separate columns.

Table 4. Regressions pertaining to the entire period (2002–2017).

	Capital Expenditure	Employment Expenditure	Research and Development
Share repurchases	−0.023	−0.238 **	0.001
Dividends	−0.001	0.003	−0.000
Company size	0.014 **	0.037 **	0.001
Cash flow	0.089 **	0.207 ***	0.001
Leverage	0.014 *	0.002	−0.001
Tangibility	0.015	0.015	0.001
Growth opportunities	0.001	0.001	0.000
Constant	−0.103 **	−0.253 ***	−0.005
Number of observations	2105	2105	2105
R-squared	0.041	0.186	0.010
Number of companies	217	217	217

*** p < 0.01, ** p < 0.05, * p < 0.10.

Earlier global studies (Almeida et al. 2016; Apergis et al. 2021; Chasiotis and Georgantopoulos 2022) reported a significant negative relationship between share repurchases and capital expenditure—which was not found to be true in the South African context when considering the entire 2002–2017 period. Possibly, the monetary value that companies in South Africa spent on share repurchases was not as substantial (and thus did not influence capital expenditure as much) as in developed countries. Wesson et al. (2018) found that share repurchases in South Africa were mostly used for smaller distributions and by undervalued companies. As such, the negative relationship between share repurchases and capital expenditure might only exist for value (low-growth) companies—this is explored in Section 4.2.

When investigating employment expenditure, it was found that this expenditure increased by a smaller degree, or even sometimes decreased, in the financial years when larger amounts were spent on share repurchases. Alternatively, the negative relationship could indicate that employment expenditure increased more in the years when less was spent on share repurchases. This finding agrees with the global evidence of Almeida et al. (2016) with respect to employment expenditure but disagrees with prior South African evidence (Wesson and Botha 2019), possibly because of the way in which Wesson and Botha (2019) measured their dependent variables. The results of the present study point out that South African share repurchases might be to the detriment of employees, who would otherwise receive increased remuneration. This is an important finding from a social justice perspective, as South Africa’s income inequality is substantial and share repurchases might be aggravating the inequality. Earlier studies in the US have also noted that repurchases might be to the detriment of employees, who are an important stakeholder group to be considered in terms of corporate sustainability (Lazonick 2014; Vaupel et al. 2022). An alternative explanation for the inverse relationship noted, is that companies without growth opportunities reduce employment expenditure and then simultaneously increase share repurchases to distribute free cash flow to shareholders. The impact of growth opportunities will be considered in Section 4.2.

No significant relationship was noted between share repurchases and research and development in the South African context, contradicting prior US evidence (Almeida et al. 2016). The data on research and development captured by the IRESS financial database seemed to be incomplete, with many companies and years showing blank entries—which might have caused the lack of any obvious relationship.

When considering the control variables, both company size and cash flow were positively related to capital expenditure and employment expenditure—meaning that larger companies with more cash flow were more likely to increase their spending on these types of internal investments. Companies with higher leverage showed a larger increase in capital expenditure (p < 0.10)—implying that companies might be using borrowed funds to increase their capital expenditure.

The relationship between share repurchases and internal investment during the 2002–2009 period in South Africa had previously been studied by Wesson and Botha (2019), but in that study, the measurement of the dependent variables might have been inappropriate for a developing country with substantial inflation. Table 5 reports the results of the present study pertaining to the 2002–2009 period only to allow for comparison with the results of Wesson and Botha (2019). Again, the regressions pertaining to the relevant dependent variables are listed in separate columns.

Table 5. Regressions pertaining to 2002–2009.

	Capital Expenditure	Employment Expenditure	Research and Development
Share repurchases	−0.048	−0.345 **	0.002
Dividends	0.001	0.004	0.000
Company size	0.022	0.023	0.002
Cash flow	0.074 **	0.160 **	0.004
Leverage	0.007	−0.032	−0.001
Tangibility	−0.016	−0.082	0.002
Growth opportunities	0.001 *	−0.002	0.001
Constant	−0.133 *	−0.114	−0.015
Number of observations	1015	1015	1015
R-squared	0.030	0.202	0.015
Number of companies	196	196	196

** p < 0.05, * p < 0.10.

The results in Table 5 are similar to those of Table 4, pointing to companies increasing share repurchases and reducing employment expenditure simultaneously. This agrees with Almeida et al. (2016) but disagrees with earlier South African results from Wesson and Botha (2019), who also examined the pre-2010 period. The difference between the results of the present study and those of Wesson and Botha (2019) is attributed to different measurement bases employed for the dependent variables.

Hypothesis 1 stated that a negative relationship existed between share repurchases and internal investment variables in the South African context. The results of the present study are consistent with this hypothesis with respect to employment expenditure but not regarding capital expenditure or research and development. The impact of growth opportunities on the relationship between share repurchases and internal investment is considered next.

4.2. The Impact of Growth Opportunities

Earlier studies indicated that the relationship between share repurchases and internal investment might be influenced by the growth opportunities available to companies. Descriptive statistics pertaining to high-growth and low-growth companies are shown in Table 6.

Table 6. Descriptive statistics pertaining to high-growth and low-growth companies.

Variable	Observations	Mean	Standard Deviation
PANEL A: High-growth companies
Capital expenditure	1077	0.01	0.039
Employment expenditure	1077	0.044	0.102
Research and development	1077	0	0.01
Share repurchases	1169	0.009	0.027
Dividends	1169	0.049	0.091
Company size	1169	6.378	0.811
Cash flow	1169	0.192	0.104
Leverage	1169	0.543	0.24
Tangibility	1169	0.252	0.205
Growth opportunities	1169	3.698	3.214
PANEL B: Low-growth companies
Capital expenditure	1028	−0.001	0.057
Employment expenditure	1028	0.023	0.108
Research and development	1028	0	0.002
Share repurchases	1169	0.005	0.027
Dividends	1169	0.034	0.455
Company size	1169	6.038	0.758
Cash flow	1169	0.111	0.089
Leverage	1169	0.49	0.229
Tangibility	1169	0.243	0.211
Growth opportunities	1169	0.876	0.445

On average, the high-growth companies included in the present study were more likely than low-growth companies to increase their capital expenditure and employment expenditure, which makes theoretical sense as these companies would seek to utilise the growth opportunities available to them. While high-growth companies showed an increase in capital expenditure from one year to the next, low-growth companies reduced capital expenditure year-on-year. In terms of employment, both high-growth and low-growth companies showed an increase in employment expenditure year-on-year, but the increase was larger for high-growth companies. This makes sense as companies would face severe backlash if they were to reduce their payments to employees, but low-growth companies might offer employees smaller increases or reduce the number of employees. The average increase in the amount spent on research and development was zero for both high-growth and low-growth companies. This finding may have been a result of the incomplete data on research and development expenditure available in the South African context, as previously explained.

The average amount spent on share repurchases is higher for companies with higher growth opportunities, which (in line with the free cash flow hypothesis) may be explained by the fact that the high-growth companies also have higher cash flow than low-growth companies. Wesson and Botha (2019) also found that companies who engage in share repurchases, on average, have positive market-to-book ratios—indicating the availability of growth opportunities. Table 7 reports on the interaction of growth opportunities with share repurchases when considering the relationship with capital expenditure and employment expenditure during the period 2002–2017. The regression focusing on research and development did not reveal any significant relationships and is thus not reported. When considering each of the two variables, capital expenditure and employment expenditure, the first column reports the regression results for the entire dataset when growth opportunities are employed as interaction variable with share repurchases. The second and third columns report separately on high-growth and low-growth companies.

Table 7. Regressions with interaction variable for growth and share repurchases.

	Capital Expenditure			Employment Expenditure
	All Companies, with Interaction	High-Growth Companies	Low-Growth Companies	All Companies, with Interaction	High-Growth Companies	Low-Growth Companies
Share repurchases	0.095 **	−0.092 *	0.102 **	0.102	−0.301 ***	0.218
Dividends	−0.001	−0.018	−0.001	0.003	−0.146 ***	0.009 **
Company size	0.014 **	0.006	0.011	0.039 ***	−0.015	0.078 **
Cash flow	0.084 *	0.060 **	0.103	0.219 ***	0.074	0.291 **
Leverage	0.013 *	−0.003	0.042 **	0.002	−0.041 *	0.078 **
Tangibility	0.017	−0.011	0.034	0.016	−0.126	0.069
Growth opportunities	0.001	0.001	0.004	0.002 **	0.003 *	0.014
High (versus low) growth	0.007 **			−0.008
High growth × share repurchases	−0.174 ***			−0.490 ***
Constant	−0.102 **	−0.029	−0.106	−0.266 ***	0.125	−0.541 ***
Number of observations	2105	1077	1028	2105	1077	1028
R-squared	0.044	0.051	0.051	0.190	0.218	0.227
Number of companies	217	172	175	217	172	175

*** p < 0.01, ** p < 0.05, * p < 0.10.

As expected, Table 7 shows that being a high-growth company has a positive effect on capital expenditure when share repurchases are zero (coefficient of 0.007, with a p-value smaller than 0.05). However, when share repurchases are factored in, the growth opportunities available to companies affect the relationship between share repurchases and capital expenditure (the interaction term of high growth and share repurchases has a negative coefficient and is significant at the p < 0.01 level). When considering high-growth and low-growth companies separately, the effect of this interaction becomes clear: for high-growth companies, there was a negative relationship between share repurchases and capital expenditure, while a positive relationship was noted for low-growth companies. In the entire dataset (Table 4), no conclusive (statistically significant) relationship was observed between share repurchases and capital expenditure. The opposite (and significant) relationships which were noted for high-growth and low-growth companies thus clouded the nature of the relationship between share repurchases and capital expenditure when the entire dataset was considered.

Finding that high-growth companies exhibit a stronger negative relationship between share repurchases and capital expenditure is consistent with Apergis et al. (2021). This might indicate that companies reduce their share repurchases to allow more funds to flow to capital expenditure. However, the negative relationship could also be indicative of high-growth companies choosing to execute share repurchases to the detriment of capital expenditure, thereby not utilising the available growth opportunities. This negative relationship emphasises the importance of monitoring share repurchase activity and its effect on internal investment.

When considering employment expenditure, a similar picture emerged. The interaction between high growth and share repurchases is negative and significant at the p < 0.01 level. As was found for capital expenditure, a negative relationship existed between share repurchases and employment expenditure for high-growth companies. For low-growth companies, however, an insignificant relationship was noted. The negative relationship between share repurchases and internal investment was therefore more pronounced for companies with high growth opportunities and supports H2.

5. Conclusions and Recommendations

The aim of this study was to investigate the relationship between share repurchases and internal investment (defined as capital expenditure, employment expenditure, and research and development), to determine whether the additional regulation of share repurchases is needed in the South African context. It is important to monitor the impact of share repurchases on internal investment, given the low economic growth, high unemployment and high income equality in South Africa, which could be aggravated by reduced internal investment.

The results showed that South African listed companies that spent larger amounts on share repurchases also spent less on employment expenditure (or had smaller increases in employment expenditure). Moreover, a negative relationship existed between share repurchases and both capital expenditure and employment expenditure when focusing only on companies with high growth opportunities. The practical implication is that South African listed companies might be increasing share repurchases to the detriment of employees and customers. An alternative explanation for the noted negative relationship is that high-growth companies could be spending less on share repurchases to allow more funds to flow to internal investment (in an attempt to utilise the growth opportunities available). The relationship between share repurchases and internal investment remains a complex matter as these two variables are often determined simultaneously.

To manage the risk of share repurchases diminishing internal investment to a level that negatively impacts economic growth, employment, and income inequality in the country, it is recommended that the South African corporate governance regulations pertaining to share repurchases be improved. Specifically, timeous announcements of all share repurchase activity would be advantageous to stakeholders. Before investing in share repurchases, South African companies should first evaluate the impact that such repurchases could have on their internal investment and long-term growth, as well as on the well-being of other stakeholders, such as employees. Increased disclosure in annual financial statements or integrated reports regarding share repurchases, and the reasons why companies engaged in them could be helpful to stakeholders. Shareholder activists should question and investigate the impact that share repurchases have on the long-term growth of companies and the well-being of employees.

The measurement bases and research methodologies applied when doing research in developed economies are not always appropriate for replication in the South African or emerging economy context. Wesson and Botha (2019) employed a measurement base for internal investment variables developed by Almeida et al. (2016) for the US environment, which was not suitable in the high-inflation South Africa context. This emphasises the importance of employing appropriate contextual measurement bases and proxies when replicating a study from a developed economy in an emerging economy.

The results of this study are limited to the target period and population studied and cannot be extrapolated to all South African listed companies, other countries, or other time periods. Although all control variables identified by prior studies were included in the regression models specified in the present study, the existence of exogeneous variables was not specifically considered. However, the inclusion of fixed effects for year and company would compensate for any exogeneous variables that are fixed in terms of time or company. A further complicating factor is that share repurchase and internal investment decisions might be determined simultaneously, which could cause endogeneity. This remains a limitation of this research.

Future research could consider the relationship between share repurchases and internal investment during and after the COVID-19 pandemic. Furthermore, the relationship between share repurchases and internal investment could be studied in other developing economies. Future studies could also evaluate the effect of share repurchases on the environmental and social sustainability of companies.