“Predictive power of economic-based performance indicators on shareholder value: Evidence from South African listed firms”

Financial statements are often number intensive, and determining the importance and relevance of these numbers from the perspective of investors and equity holders is paramount. However, empirical studies concerning the correlation between several accounting and economic-based indicators with shareholder returns have yielded contradictory results. Additionally, considering the relatively limited studies on economic-based indicators such as refined economic value-added and economic value-added momentum, this study evaluated the predictive power of refined economic value added, economic value-added momentum, and economic value added (economic-based indicators), along with traditional accounting-based indicators such as return on equity and earnings per share on the shareholders’ returns. The study employed fixed-effect instrumental variable regression and panel quantile regression techniques to examine 49 non-financial companies listed on the Johannesburg Stock Exchange from 2007 to 2021. Overall, the results showed that economic value added is a significant negative predictor of shareholder returns, while refined economic value-added is a positive determinant. In addition, the refined economic value-added coefficient remains positive, with the impact increasing across the conditional quantiles. This study concludes that refined economic value-added provides a superior and realistic determinant of share-holder value on the Johannesburg Stock Exchange compared to other measures.


INTRODUCTION
In today's competitive world, which is characterized by resource scarcity, maximizing shareholder wealth is the ultimate objective of corporate managers.Shareholder wealth measurement can either be in the form of dividends received or in the form of capital appreciation, or both.Hence investors require a forward-looking construct that accurately assesses a company's performance and is associated with the market price of a share.
Traditional accounting-based indicators (ABI), such as earnings per share (EPS), return on equity (ROE) and return on assets (ROA), are used by shareholders to assess investment performance.However, these constructs have been criticized because they do not consider the total cost of capital, which consists of the cost of debt and equity (Panigrahi et al., 2014).Therefore, to address the ABI's shortcomings, economic-based indicators (EBI) such as economic value added (EVA), refined economic-valued added (REVA), and economic value-added momentum (EVAM), which incorporate the total cost of capital have been advocated (Stewart,1994;Bacidore et al., 1997;Stewart, 2009).
Several studies concerning the association of ABI and EBI with shareholder returns have been undertaken on the global capital markets; however, these exercises yielded different results.Additionally, to the authors' knowledge, ABI and EVA relationship with shareholder returns has been extensively tested globally; from the Johannesburg Stock Exchange-listed companies' perspective, constructs such as REVA and EVAM are not tested.Therefore, there is an incentive to examine the explanatory power of REVA and EVAM on shareholder return alongside EVA, EPS, and ROE simultaneously for the first time on the largest African stock exchange.

LITERATURE REVIEW
The equity market literature in the last five decades is characterized by a search for an indicator used for company performance evaluation that is correlated to stock price, considering that stock price affects shareholders' return.For instance, Ball and Brown (1968) conducted the first test on the value relevance of ABI in the United States, and Stern Stewart Company reintroduced EVA in the 1980s after being advanced by General Motors in 1920.Building upon Ball and Brown's (1968) seminal work, Nichols and Wahlen (2004) studied the correlation between annual stock returns and annual earnings on the New York Stock Exchange (NYSE), American Express Company (AMEX), and National Association of Securities Dealers Automated Quotations (NASDAQ) firms.Based on 31,923 firm-year observations from 1988 to 2001, the descriptive statistical test indicates a significant correlation.
Additionally, studies using several accounting constructs, different methodologies, such as ordinary least square (OLS), pooled least squares (Pols), random effect model (RE), fixed effect model (FE) and a combination of univariable and multivariable regression models, were undertaken.
From an ABI perspective, several studies have yielded a positive association with stock prices.For example, based on a sample of the top 30 firms of the Financial Times Stock Exchange (FTSE)-100 index of the London Stock Exchange from 2005 to 2014 and using a panel regression analysis, Anwaar (2016) found that net profit margin and ROA have a significant positive correlation with stock returns, while EPS negatively impacts stock returns; however, ROE and quick liquidity ratio do not exert any impact on stock returns.From the EBI perspective, in particular, EVA, where the total cost of capital is at book value, several scholars have tested Stewart's (1994)  In the JSE context, Magwegwe (2003) found that the EVA metric is not correlated with the share price of JSE-listed companies.Similarly, Sauro and Tafirei (2016), based on multiple regression analysis and the OLS method, found that EVA is relatively weak compared to EPS, dividend per share (DPS), and marginal revenue in commercial banks.In contrast, still from the JSE standing point, Weldon (2013) found that the coefficient correlation R 2 of EVA was positive and statistically correlated with the share price (2000-2010).However, De Wet (2004) tested 89 listed companies from 1995 to 2004 and concluded that ABI has a better positive correlation with MVA than EVA spread.
From the REVA proposition, the total cost of capital is at market value; hence the capital charge does reflect the opportunity cost for investors because if investors sell a firm for its market value and apply the proceeds to assets of identical risk profiles, they could expect to earn a return equal to the weighted average cost of capital (WACC) on the overall market value of the firm, and not only on the book value of the investment of the firm shown in the balance sheet (Richter & Honold, 2000).The critical advantage of REVA relative to EVA, according to its proponents Bacidore et al. (1997), is that incremental shareholder value has been created whenever REVA is positive, that is, the operating income flowing to shareholders at the end of the period as a percentage of their investments' market value at the beginning of the period exceeds their opportunity cost of capital.However, this condition is not always the case for EVA; from the EVA proposition, shareholders could be getting an operating-income-based return below their opportunity cost of capital, even when EVA is positive.Based on this scenario, Pourali and Roze (2013) state that REVA emphasizes the value relevance of information instead of its reliability embodied in book value.
Unlike ABI and EVA, there are relatively few studies on REVA; nevertheless, the dearth of articles yielded mixed results.The first study on the REVA conducted by its proponents, Bacidore et al. (1997) In summary, no conclusive evidence supports EVA's explanatory power of shareholder returns relative to ABI.On the other hand, EVA's equity valuation at book value is not aligned with shareholders standing point because the shareholder return is a function of the stock price differential between two periods plus dividends.Therefore, with the equity valuation based on the market price, REVA is more relevant for equity holders; consequently, REVA should impact stock price.Regarding EVAM, despite being based on EVA, EVAM attributes provide valuable insights into a company's economic profit variation between two consecutive periods.

AIM AND HYPOTHESIS
Based on the literature review, the theoretical superiority and the value relevance of REVA and EVAM have not been tested on the JSE.Therefore, this study's purpose is to examine the positive explanatory power of EBI (REVA and EVAM EVA) and ABI (EPS and ROE) on shareholder returns on the JSE.The null hypothesis is given as follows: Hoi: EVA, REVA, EVAM, ROE, and EPS do not have a significant positive explanatory power on shareholder returns on the JSE-listed firms.

Sample and data source
The sampling technique in this study was purposive sampling.

Operational variables and measurements
The study identifies the dependent variable as shareholders' return (ShR); the independent variables are the ABIs (EPS and ROE), EBIs (EVA, REVA, and EVAM), while total assets and revenue are the control variable.According to Ahsan (2012), the ROE links the statement of comprehensive income (NOPAT) and the statement of financial position component (average shareholder equity); therefore, it is highly regarded by investors and can be expressed as a ratio of the two.EPS was calculated as the ratio of net income after preferred stock dividends.EVA and REVA, according to Stewart (1991) and Bacidore et al. (1997), respectively, are defined as the difference between a company's NOPAT and its total cost of capital (cost of debt and cost of equity) and can be expressed as: Adjusted NOPAT measures company profit from its ongoing operations adjusted from GAAP distortions to convert the accounting profit into economic profit.IC is the operating capital employed; it is defined as interest-bearing debt and equity at the beginning of the period, and  is the weighted average cost of capital.The difference between EVA and R EVA arises from the equity valuation approach as the EVA uses equity valuation based on historical or book value, R EVA is based on the market stock price.Therefore, the R EVA proposition is aligned with the shareholder's valcreation.Eva Dimension (2009) quantifies EVA growth or deterioration over time and can be expressed as follows (Stewart, 2009).
where P 1 is the stock price at year-end, P 0 is the share price at the beginning of the year, and D represents the dividend per share paid in the current year -this variable is expressed as a percentage.

Model specification and estimation techniques
Panel data (longitudinal data) regression analysis is paramount to establish the dimension of the relationship over several periods between the study's independent and dependent variables ( where α i is the constant, γ t is the time-invariant company-specific effect, and ε i,t is the error term, β is the elasticity of the explanatory variables.
This study further used a quantile regression technique to examine the varied and distributional effects at different levels of shareholder returns.
The parameters to be estimated are the probability: P{δ i + Z it } > 0 = 1.(α, β, δ, γ), (αi, δi), i takes the value 1 to n, representing the company's i fixed effect.The distributional determinants of shareholders' return on the JSE are presented in Table 3 to determine how the impact of EVA, REVA, EVAM, total assets, and revenue differs across the conditional distribution of shareholder returns using the Machado and Silva (2019) approach.Table 3 shows that the EBIs retain their sign and significance; however, the magnitude of impact varies across levels.For example, the magnitude of EVA's negative impact depends on the quantile as the effect at the lower quantile (25 th quartile) of shareholder returns (0.72%) reduces to 0.69% in the 50 th quartile and 0.66% in the 75 th quartile.Conversely, the positive impact of REVA at the lower quantile (25 th quartile) of shareholder returns (0.036%) increases to 0.037% in the 50 th quartile and 0.039%  in the 75 th quartile.Similarly, the insignificant positive impact of EVAM at the lower quantile (25 th quartile) of shareholder returns (0.21%) increases to 0.40% in the 50 th quartile and 0.58% in the 75 th quartile.

RESULTS
The ABIs' coefficients are not statistically significant.The results indicate that regardless of the distribution of shareholder returns, EPS positively and insignificantly impacts shareholder returns.However, coefficients indicate that the size of impact differs across quantiles as the lower quantile (25 th quartile) (0.00171%) impact reduces to 0.00161% in the 50 th quartile and 0.00151% in the 75 th quartile.ROE's impact across the distribution of shareholder returns aligns with the fixed effect results presented in Table 2, but the impact the lower quantile (25 th quartile) (-0.00191%) increases to -0.000533% in the 50 th quartile and 0.000753% in the 75 th quartile.Lastly, the impact of the control variables varies across quantiles, but they are not statistically significant.
Overall, the study rejects the null hypothesis which states that REVA does not have a significant positive explana-tory power on shareholder returns and accepts the null hypotheses that state that EVA, EVAM, EPS and ROE do not have a significant positive explanatory power for shareholder returns on the JSE.

DISCUSSION
Based on the findings, the EVA negative and statistically significant correlation with shareholder returns from the directional standing point contrasts Weldon (2013)

CONCLUSIONS
The purpose of this study is to examine the positive explanatory power of economic-based indicators (economic value added, refined economic-valued added, and economic value-added momentum) and accounting-based indicators (earnings per share and, return on equity), on shareholder returns on the Johannesburg Stock Exchange.The study concludes that economic value-added and refined econom-ic-valued added can be used to predict shareholder returns; however, the prediction signal is different.In other words, economic value-added negative correlation with shareholder returns means that an increase in economic value-added leads to a decrease in shareholder returns.In contrast, an increase in refined economic-valued added increases shareholders' returns.A positive economic value-added momentum implies that a company is consistently creating value by generating returns above its cost of capital.However, economic value-added momentum does not significantly impact shareholders' returns, casting doubt on its reliability.
Concerning the accounting-based indicators, it can be concluded that earnings per share cannot be used to predict shareholder returns on the Johannesburg Stock Exchange, despite its relative leverage arising from its compulsory disclosure in the financial report.Consequently, one may assume that the equity market negatively perceives the potential higher debt/equity ratio (De Wet 2013), which leads to higher earnings per share.Moreover, the return on equity "emphasizes" the negative market sentiment towards higher return on equity, which may arise from a company's higher level of debt.This study contributes to the literature by presenting the investors and analysts on the Johannesburg Stock Exchange with a measure of refined economic value added that overcomes the limitations of accounting metrics and economic value added in determining shareholder returns.This study presents the first evidence of refined economic value-added superiority in predicting shareholder returns relative to other metrics on the Johannesburg Stock Exchange.
Using a panel regression analysis, Aveh and Vitor (2017) examine the relationship between EPS, ROE, book value market, dividend per share, dividend yield, leverage capitalization and stock price.The results indicate that ABI, namely EPS, ROE, book value and the value-based measures (market capitalisation), are correlated to stock prices.Additionally, based on multiple regression analysis, Menaje (2012) found a strong and positive correlation between EPS and share price, i.e., the model explained 73% of the stock price variation in the Philippine Stock Exchange.
, in the United States(1982)(1983)(1984)(1985)(1986)(1987)(1988)(1989)(1990)(1991)(1992), reveals that REVA is superior to EVA.Baseriet al. (2013) also showed that REVA, EVA, and OCF significantly and positively correlated with future OP in 56 companies listed on the Tehran Stock Exchange (1996-2008).In the same exchange, Pourali and Roze (2013), based on 67 companies (2006-2010), indicate that ROA and REVA were the most significantly and positively correlated with market value added (MVA) compared to ROE and EPS, EVA.Furthermore, Quintiliani (2018) conducted a similar test with 75 Alternative Investment Market (AIM) companies in Italy (2010-2015) and indicated that REVA correlates more with MVA than EVA.In addition, Nugroho (2018), using 104 manufacturing companies listed on the Indonesian Stock Exchange (2014-2016), reveals that EVA, MVA, and REVA substantially correlate with stock prices and return.In contrast, other empirical studies failed to attest to REVA supremacy over other constructs.For instance, Arabsalehi and Mahmoodi (2012) report that ABI, namely ROE and ROA, are better predictors of company stock prices than the four value-based metrics (EVA, REVA, MVA, and SVA).No difference was found between the EVAM for the lodging and restaurant industries.Nakhaei et al. (2012) measured the correlation between EBI and the MVA in Bursa Malaysia (2001-2010), and no evidence supporting that EVA, REVA, and EVAM positively correlated with MVA was reported.
Erasmus (2008, 2015)isted of the (i) adjustment to the accounting profit; (ii) computation of the invested capital (IC); (iii) determination of the total WACC(Jakub et al., 2015).Consequently, the EBI input variables, namely, NOPAT, IC, cost of debt, cost of equity, and dependent variable input factors such as share price and dividend per share, were obtained from Integrated Real-Time Equity System (IRESS) (former INET-BFA) data source.The IRESS database was also used by De Wet and Du Toit (2007), Du Toit (2015), andErasmus (2008Erasmus ( , 2010) )in studies that required financial information of JSE-listed companies.
Stewart's (1991)(2004)t al. (2004), there is no theory that guides the selection of the most relevant items to be adjusted.Therefore, the adjustment exercise was undertaken in line withStewart's (1991)recom-mended EVA 1 is the EVA in the current period, EVA 0 is the EVA in the prior period (t n -1), and Sales 0 is the revenue for the preceding period (t n -1).

Table 2 .
Determinants of shareholder returns
Magwegwe (2003)nd Evuebie (2017), who found a significant and positive correlation between EVA and stock price/return in Nigeria, andMagwegwe (2003)who found that the EVA metric is not correlated with the share price of JSE-listed companies.The findings, from the relative explana-tion power, are consistent with Gupta and Sikarwar (2016) who found that EVA is superior to ABI (EPS, ROA and ROE), but contradict Sauro and Tafirei (2016), who found that EVA is relatively weak compared to EPS, and DPS, in commercial banks.Pourali and Roze (2013) and Sharma and Kumar (2011) for the Indian market, and Menaje (2012) for the Philippine Stock Exchange, who found that EPS was strongly correlated with stock prices.The ROE outcome supports previous studies on the JSE, such as De Wet and Du Toit (2007), who found that ROE was not a significant determinant of shareholder value.Similarly, Vermeulen (2016) found that the ROE and EPS did not significantly affect stock returns.