Corporate Ownership & Control / Volume 18, Issue 1, Autumn 2020

EVA-BASED FINANCIAL PERFORMANCE

MEASUREMENT: AN EVIDENTIAL STUDY
OF SELECTED EMERGING COUNTRY
COMPANIES
Kushagra Goel *, Sunny Oswal **
* Faculty of Management, NMIMS University, Mumbai, India
** Corresponding author, NMIMS University, Mumbai, India
Contact details: NMIMS University, Vile Parle (W), Mumbai 401107, India

Abstract

How to cite this paper: Goel, K., & This paper aims at examining the claims that economic value
Oswal, S. (2020). EVA-based financial added (EVA) is a superior performance indicator than
performance measurement: An evidential
study of selected emerging country the traditional performance indicators like ROCE, NOPAT, EPS,
companies. Corporate Ownership & OCF, and RONW. This study investigates the relative explanatory
Control, 18(1), 179-195. power of EVA measure of non-financial Indian companies with
http://doi.org/10.22495/cocv18i1art14
respect to two measures, market value added and stock returns
Copyright © 2020 The Authors used as a proxy for shareholder value. The analysis is performed
for a sample of 46 Indian companies for the period of 2009-2019.
This work is licensed under a Creative The panel data regression models are employed to test the
Commons Attribution 4.0 International
License (CC BY 4.0). relative and incremental information content of EVA and other
https://creativecommons.org/licenses/by/ audited accounting-based measures. Relative information content
4.0/ tests reveal that NOPAT and OCF appear to be more value-
ISSN Online: 1810-3057 relevant than EVA in explaining the market value of Indian
ISSN Print: 1727-9232 companies. It was also found that ROA is more closely associated
with stock market returns than EVA. Additionally, incremental
Received: 25.05.2020
Accepted: 16.10.2020
information content tests suggest that EVA underperforms in
comparison with NOPAT and OCF in analysing market value
JEL Classification: G32, F65, M41 added. It was also found that EVA does not add any incremental
DOI: 10.22495/cocv18i1art14 information content to that provided by ROA and ROE accounting
measures in explaining stock returns. Overall, the findings do not
support the purported superiority of EVA to established
accounting variables in association with market value or stock
market returns of the firm. It is concluded that non-financial
variables such as research and development, customer
satisfaction, internal business process efficiency, innovation,
employee satisfaction, CSR, product quality apart from financial
variables drive market value and should be considered by
investors in developing their investment strategies.

Keywords: EVA, Panel Data, Relative Explanatory Power, Market

Value Added

Authors’ individual contribution: Conceptualization – K.G.;

Methodology – S.O.; Formal Analysis – K.G.; Resources – S.O.;
Writing – Original Draft – S.O.; Writing – Review & Editing – K.G.;
Funding Acquisition – S.O. and K.G.

Declaration of conflicting interests: The Authors declare that there is

no conflict of interest.

1. INTRODUCTION an organisation means shareholder’s wealth.

A shareholder’s wealth in an organisation is
The objective of wealth maximisation is the product of the number of shares with the share
a well-accepted concept in business. Wealth for price. Share market price is the performance

179
 Corporate Ownership & Control / Volume 18, Issue 1, Autumn 2020

indicator of the progress of a company. Wealth documented similar results. Bao and Bao (1998)
maximisation seeks to serve the interests of its proved that abnormal economic earnings (AEE) do
suppliers, employees, management, customers, and not appear to be associated with a stock value or
society at large. stock price in a significant manner. The study by
With the aim of analyzing the financial and O’Byrne in 1996 reported EVA to have explained
economic health of a company, many performance greater than double the variance. O’Byrne also found
matrices in finance are employed. The same could be that changes in EVA explained in a significant
the measures of accounting or value-based measures manner more variation in changes in market price.
that contribute to the wealth of shareholders. Lehn and Makhija (1996) found contrasting results
The returns on investment by the investors indicate that economic value added has a “slight edge as
the wealth created for them. The return received by a performance measure” in comparison to multiple
an investor can be like in the format of appreciation measures of accounting. Bacidore, Boquist, Milbourn,
of capital or dividends or sometimes all. Therefore, and Thakor (1997) used a new measure REVA –
a company having better accounting performance which is defined as refined economic value added.
measures will drive the stock returns, resulting in Ismail (2006) reported the earnings to be strongly
an increase in shareholder’s wealth. Usually, the associated along with equity returns compared to
measures that are more closely correlated with EVA and RI. The findings of Ismail’s (2006) study
the equity returns are considered better indicators. failed to support the findings of EVA advocates
Chen and Dodd (2001), Mcgrattan, Rogerson, about its superior indications over accounting
and Wright (1997), and some other researchers measures for explaining variations in stock returns
reported that a single accounting measure could not using variables in level and changes.
explain the variability in shareholders’ wealth. The Some studies provide support to Stern Stewart
traditional audited accounting metrics like earnings hypothesis that EVA adds to the wealth of
for each share (EPS), equity return, and assets return shareholders and drives the stock returns
would not take into account the cost of capital for (Lefkowitz, 1999; Milunovich & Tsuei, 1996;
their calculations. Other measures like NOPAT, OCF, Worthington & West, 2004; O’Byrne, 1996; Uyemura,
and ROCE also do not consider cost of capital so Kantor, & Pettit, 1996; De Villiers & Auret, 1998;
cannot be used to predict the firm value and thereby Turvey et al., 2000; Lehn & Makhija, 1997; Forker &
may not be a good measure for corporate Powell, 2004) greater in comparison to traditional
performance. However, capital charges on the capital measures of accounting. Few studies indicate that
employed are incorporated in determining value- EVA is not associated to the equity returns and does
based measures. not add to shareholder’s wealth (Kyriazis &
In the early 1990s, several scholars suggested Anastassis, 2007; Peixoto, 2002; de Wet, 2005;
value-based performance measures which are Maditinos, Sevic, & Theriou, 2009; Biddle et al., 1997;
unaudited, one of these measures is EVA (economic Ismail, 2006).
value added). EVA was advocated earlier and The primary objective of the study is to examine
pioneered by Stern Stewart & Co., which was the idea of proponents, the basic objective of this
US-based and, therefore, is considered as a trade- study is to examine the claim of proponents of EVA
marked variant of RI (residual income). It may be of its richness and superiority over the traditional
an intelligent argument that EVA should be measures of performance in the Indian context.
employed rather than cash flow or earnings from the Relative and incremental information content tests
operations for construing as the best performance are conducted using panel data regression models.
measure that is periodic. Other measures based on A sample of 46 companies is selected and tests are
value are cash value-based addition, ROI (cash flow), performed for the period 2009-2019. The study
discounted profits (economic), shareholders value contributes to the existing literature by providing
added were also developed by consulting companies evidence from the Indian market on examining
to measure the financial performance of the firm. the superiority of value-based performance measure
Stewart (2010) claimed that EVA’s most over conventional measures.
accurate representation of the company’s true This study addresses the core research
profit. Many companies adopted EVA, as an internal question: “Which measure of performance out of EVA
control measure in the later years such as Sprint, and traditional is a better measure for Indian
Allied Holdings, Whirlpool, Coca-Cola, Toys “R” Us, markets?”. The findings reflect that economic
and Georgia-Pacific. The unique selling point of EVA value-added in the Indian market does not have
in comparison to residual income is that it takes into relative content which is better than traditional
account economic profit and economic capital which measures. The study also contends that a single
have certain accounting adjustments. The number of financial metric does not drive the market price,
adjustments has been debated in the academic and there are other non-financial factors that drive
literature and Stern Stewart & Co. has suggested 164 shareholder value and could be collectively
such adjustments. considered as measures of performance.
A number of researchers have conducted The paper is divided into six sections, viz.,
empirical studies to investigate whether EVA, Introduction, Review of literature, Methodology,
a value-based performance measure, is more reliable Analysis, Discussion, and Conclusion. The context
measure of financial performance for a company and relevance are discussed in the first section
than the accounting-based measures. The results are followed by the review of literature and
mixed and controversial. methodology. The working is shown in the analysis
Biddle, Bowen, and Wallace (1997) analysed section followed by the discussion of results and,
the earnings dominating residual income, which in finally, the conclusion.
effect dominates the economic addition of value for
explaining equity returns. Chen and Dodd (2001)

180
 Corporate Ownership & Control / Volume 18, Issue 1, Autumn 2020

2. LITERATURE REVIEW To conduct the incremental information content

tests, a pairwise combination of one value-based
2.1. Seminal works measure and one traditional measure was taken.
Lin and Zhilin (2008) conducted a study using
Chen and Dodd (2001) analysed and examined an integrated EVA performance measurement model
the fitness and relevance of matrices that measure to test the superiority of this model to traditional
profitability. The study suggests that EVA has measures. The data was taken from China-listed
incremental information value than RI and OI companies for the period 2000-2001 and was
measures, but the increase in R2 is marginal from analysed using BP neural network. The neural
a practical point of view. So, the study concluded network was designed as an input layer, output
that the benefit of the EVA system is not so huge layer, and a hidden layer. The input layer consisted
that it justifies the multiple costs in calculating of traditional and IEPM models with the variables.
adjustments to the audited statements. A less costly The output layer was developed based on market
measure is RI. A measurement paradigm other than value per share. The hidden layer was created
EVA could be taken to align organizational metrics to show the input processing and internal structure
with stock value. Overall, their findings failed of the problem. Mean square error was used
to support the assertion. to measure network performance. The IEPM model
Erasmus (2008) analysed the fitness of has less mean square error than the traditional
the measures based on value like RI, EVA, CVA (cash model. It was proved that IEPM is a more effective
value added), and CFROI (cash flow return on tool in terms of prediction ability to measure
investment) to explain adjusted share returns. the company’s performance.
The sample data is 316 firms on Johannesburg Worthington and West (2004) used pairwise
Securities Exchange during the 15 year period from regression to extend their own earlier research work
1991 to 2005. The initial part of the research is (Worthington & West, 2001) by employing multiple
based on information that is relative to content tests alternative formulations that helped for pooling
on all performance measures. The first variable, panel data which are random effects, fixed effects,
which is dependent, is the market adjusted equity and common effects. The analysis here was
return which is the gap between the returns, conducted for the Australian market. It was found
compounded on a share and the ALSI index. that the fixed effect model is the most appropriate
The measures used for relative tests are CFO, EBEI, pooling technique for the models. The first phase of
RI, EVA, CVA, as well as inflation-adjusted measures the study used a valuation model that takes net cash
EVA-real, CVA-real, CFROI. One independent variable flow, earnings before extraordinary items, residual
at time t and the same variable lagged by one time income, and economic value-added as explanatory
period divided by the share’s market value are variables and stock return as a dependent variable.
included in individual regressions. The data is Multicollinearity was not significant. The pairwise
removed from heteroscedasticity by the division. regressions showed that EVA best explains stock
The results showed that EBEI outperformed the returns and the largest relative information content
other measures with the highest adjusted R2 value. than the other accounting measures.
The second part of the study is based on the The second phase of the study used
incremental information content tests to examine a component model with multiple adjustments that
whether the addition of the value-based measures used various independent variables and EVA as
contributes significantly to the explanatory power. the dependent variable. It was found that accounting
The components of the CFROI and lagged by one adjustments contribute the most to variation in EVA
time period are taken and the results showed that and hence explains stock returns. The reasons cited
adjusted R2 for CFROI components is lower than that for the divergence in results between this paper and
of EVA and CVA. It was proved that although earlier US studies could be the differences in
the incremental information content of CFROI, CVA, GAAP between the US and Australia, different
and EVA components is statistically significant, research design (fixed effects formulation, different
however, the level of significance is low. explanatory variables).
Maditinos et al. (2009) analysed how much is Gupta and Sikarwar (2016) studied the
the explanatory power of EVA and SVA (shareholder relationship between four performance measures
value added) those are value-based measures of EPS, ROE, ROA, EVA/CE with stock returns. They
performance in comparison with three traditional used a sample size of 50 companies, mostly Indian
accounting performance measures. Data from listed from Nifty 50 for the period 2008-2011 for
companies of the Athens Stock Exchange was the analysis. Panel regression models are used
collected over the period of 1992-2001 to execute to perform the incremental and relative information
the pooled OLS regression models. This study is content tests. Hausman test was performed to find
conducted for companies in an emerging market. the suitability of fixed and random effects models.
The relative information content tests were The findings indicated that the fixed effects model
conducted using five regression models with level is more appropriate. The study finds evidence in
and change variables for EPS, ROI, ROE, EVA as support of EVA as a better measure than traditional
independent variables. Change in SVA was not taken accounting measures.
since it represents a change of shareholders’ value Fountaine, Jordan, and Phillips (2008) conducted
added from one period to another. Annual the study to test if EVA can be used as a portfolio
compounded stock returns was used as separation criteria. The study used 1000 listed
the dependent variable. The results of relative US companies ranked by revenue as the buy list.
information content tests showed that the EPS EVA/CE (eva2cap) was calculated to select
equation (with the highest R2) provides more the top 100 and bottom 100 performers to create
information than EVA in explaining stock returns. separate portfolios. Portfolios were also created

181
 Corporate Ownership & Control / Volume 18, Issue 1, Autumn 2020

using 100 best and 100 worst performers in terms not support the argument of the superior
of the average of eva2cap for the past 3 years – informational content of EVA with MVA as
eva3yravg and percentage change in eva2cap – the dependent variable. The test to verify the relative
evadelta. It was found that EVA has economic information content showed that NOPAT and OCF
significance as a selection criteria for portfolio are better measures than EVA and are more closely
separation since the null hypothesis of no associated with the value of Indian firms.
significant returns between the two portfolios was The results of the incremental information
rejected. The study also concluded that these content test suggest that EVA has a marginal
criteria can be used in bear markets 2000-2002 and contribution to information content beyond that
bull markets. provided by five popular traditional accounting
De Wet and Hall (2005) analysed the results of measures such as EPS, NOPAT, OCF, ROCE, and ROE.
89 industrial companies listed on JSE Securities It was also concluded that non-financial variables
Exchange, South Africa, for the period 1995-2004 like the quality of the product, customer
using market value divided by invested capital satisfaction, and employee retention need to be
added as a proxy for shareholder value. The definitely considered to explain the market price
independent variables used for regression analysis variation of a company.
were EVA/invested capital, cash flow from Ismail (2006) conducted research to test the
operations/invested capital, ROE, ROA, EPS, and DPS. assertion of EVA in comparison to other accounting
Six regressions were taken with one independent measures. This paper used a sample of 2252
variable at a time and standardised MVA as firm-year observations from the UK market. They
the dependent variable. The results suggested that applied panel data regressions to conduct relative
operating cash flows has a stronger association with information content. The dependent variable used is
MVA than standardised EVA. The study has also the annualized compounded annual rate of return to
reaffirmed the significance of cash flow management. the shareholders to study information content of
The study also found that there is little explanatory profit measures. The independent variables used are
power in EPS, DPS in explaining MVA, questioning net income, net operating profit after tax, operating
the use of earnings or dividends in share valuations cash flow, residual income, and EVA. Accounting
of South African listed companies. The findings did adjustments were carried for NOPAT and capital to
not support the claims of EVA proponents that calculate EVA. The correlation matrix showed that
EVA is a better measure than other accounting EVA has the lowest correlation with stock return.
benchmarks and has a significant impact on The panel data regression was carried out by
the company’s MVA. conducting five separate regressions for each
Kyriazis and Anastassis (2007) tested the performance measure (EVA, RI, NI, NOPAT, and OCF)
incremental and relative content explanatory power using variables in levels and in changes. The results
in the EVA model for stock returns in the context of showed that EVA does not outperform standard
a small European developing market, the Athens accounting measures NI and NOPAT in explaining
Stock Exchange. The Greek market has differentiated the stock return. So, they concluded that the relative
standards in accounting for the US and other major information content tests refute the claim of EVA
European countries. The period of study 1996-2003 proponents that EVA has been so far the best
was selected since the Greek stock market was in financial metric.
the transitory phase to more efficient and developed Gupta and Venkata Vijay Kumar (2013) studied
after the year 2000. 107 non-financial Greek firms, the concept of value-based accounting with respect
that were publicly traded, were used in this study to Indian companies. Data of companies from BSE
giving 847 annual observations. 100 for the year 2009 was taken as a sample. OLS
Bhasin (2017) studied the disclosures on regression analysis is conducted for the research
the EVA from the annual reports that were a result paper. Three regressions are carried out with
of the 500 sample corporations from India. accounting profit, economic value added, market
The study was majorly based on the secondary data value added as the dependent variable for each
sources for five years from 2007-2011. The study regression. Cost variables – the cost of raw material,
concluded that EVACE, ROCE, ROE, and EPS of interest expenses, depreciation, employee expenses,
sample companies differ significantly using ANOVA and market value of debt and equity as independent
single factor test. Chi-square test is used to indicate variables. The second part of the study used
that the difference between observed and expected the ANOVA test and concluded that negative EVA
values of EVACE is not statistically significant and decreases MVA of the organization. It was found
the same is attributed to sampling fluctuations. Karl that most of the organizations have positive MVA (10)
Pearson’s Correlation Matrix shows the correlation and 51 companies have negative EVA. The study
between EVAE, ROE, ROCE, and EPS for Indian concluded that continuance of negative EVA in
companies. the future would lead to negative MVA.
Kumar and Sharma (2011) examined and Masyiyan (2019) conducted a study and tested
analysed the incremental information and relative the superiority of EVA in comparison to other
content in economic value added and other accounting measures. The paper uses regression
conventional accounting measures of performance amongst variables related to depreciation, cost, and
to explain market value added. They used a sample expenses.
containing 97 non-financial companies over Shah, Haldar, and Nageswara Rao (2014)
the study duration of 2000-2008 and obtained 873 discussed the applicability of EVA as a powerful tool
firms-year observations. Ordinary least square to introduce financial flexibility in the organization.
regression was applied to test the claims about the The findings of this study reveal that EVA can be
superiority of EVA as a corporate financial used as a tool to enable the company to differentiate
performance measurement tool. Their results did between value-creating and value-destructing

182
 Corporate Ownership & Control / Volume 18, Issue 1, Autumn 2020

activities and facilitating the managers in achieving traditional performance measures for a firm’s
strategic goals. This research emphasized the performance is more superior.
importance of EVA as a performance measure that EVA brings in two aspects of flexibility –
measures the company’s progress, guides investors, operational and financial. At the operational level,
and leads to better decision making, thereby reflects it can be used as a comparative performance
the market’s assessment of the company’s value. assessment tool for activities, divisions, or
Ali (2018) talks about the analogy between businesses, it brings in the culture of ownership,
economic value added and market value added in and pushes managers to constantly evaluate
the agro context for Russian markets. The author investment decisions. At the financial level, EVA not
uses panel data for the study. The findings provide just focuses on cutting costs but also on cutting
a holistic overview of the areas where each measure excess capital from activities that have return on
can be used. capital lower than the cost of capital. It is concluded
Ronen, Pliskin, and Pass (2018) along with his that the adoption of EVA enhances flexibility in
colleagues analysed the shortcomings of traditional the organization in various dimensions, sends
accounting measures in their work “The evils of positive signals to the investors, and positions
traditional accounting”. The paper compared the the organisation to rapidly adapt to the changes in
financial performance as reflected by the traditional the external environment bringing inherent stability.
measures vis-a-vis modern measures for selected
companies. 3. RESEARCH METHODOLOGY

2.2. Gap identification 3.1. Sample selection

The research works mentioned above focus majorly The sample is taken for a duration that spans across
on the achievement of goal congruence of ten years, from FY 2010 to 2019. The sample for this
managerial and shareholder goals by tying study consists of 46 non- financial listed companies.
compensation to EVA measure. The work done so far Thus, a balanced panel set of 460 firm-year
has a limited discussion on the opportunity cost of observations was obtained. 8 sectors have been
equity in-spite of being the most distinguished taken for this – Automobile, Personal care, IT,
feature of EVA. This would results in better Cement, Paints, Pharmaceuticals, Steel, and Tyres.
investment decision-making by managers as Majorly, the data used in this study is extracted
maximising returns after deducting the cost of from Bloomberg. The companies that are considered
equity. Corporates have introduced EVA linked for this study are mentioned in Table 9.
flexible bonus systems which ensure managers and The following companies have been removed
employees act like owners. from the study for the below-given reasons.
Also, no study examines the modalities and
evidence about which measure between EVA and

Table 1. Companies dropped from the study (Part 1)

Sectors Companies Reason for deselection

Accounting cycle:
Eicher Motors Ltd (EIM IN) Ends 12/31 for FY2010- FY2014.
Automobile
Ends 03/31 for FY2015-FY2019.
Hero MotoCorp Ltd (HMCL IN) Data for FY2010 - FY2012 not available.
Accounting cycle:
Nestle India Ltd (NEST IN) Ends 12/31 for FY2010-FY2018.
FY2019 data not available.
Personal care Accounting cycle:
Procter & Gamble Hygiene & Health Care Ltd (PG IN)
Ends 06/30 for FY2010-FY2019.
Accounting cycle:
Gillette India Ltd (GILL IN)
Ends 06/30 for FY2010-FY2019.
Accounting cycle:
Mphasis Ltd (MPHL IN) Ends 10/31 for FY2010-FY2013.
Ends 03/31 for FY2015-FY2019.
Accounting cycle:
HCL Technologies Ltd (HCLT IN) Ends 06/30 for FY2010-FY2014.
IT
Ends 03/31 for FY2015-FY2019.
Accounting cycle:
Hexaware Technologies Ltd (HEXW IN) Ends 12/31 for FY2010-FY2018.
FY2019 data not available.
Larsen & Toubro Infotech Ltd (LTI IN) Data for FY2010-FY2011 not available.
Accounting cycle:
ACC Ltd (ACC IN) Ends 12/31 for FY2010-FY2018.
FY2019 data not available.
Accounting cycle:
Ambuja Cements Ltd (ACEM IN) Ends 12/31 for FY2010-FY2018.
FY2019 data not available.
Cement
Accounting cycle:
Shree Cement Ltd (SRCM IN) Ends 06/30 for FY2010-FY2014.
Ends 03/31 for FY2015-FY2019.
Accounting cycle:
HeidelbergCement India Ltd (HEIM IN) Ends 12/31 for FY2010-FY2014.
Ends 03/31 for FY2015-FY2019.

183
 Corporate Ownership & Control / Volume 18, Issue 1, Autumn 2020

Table 1. Companies dropped from the study (Part 2)

Sectors Companies Reason for deselection

Akzo Nobel India Ltd (AKZO IN) Data for FY2010-FY2011 not available.
Paints
Alembic Pharmaceuticals Ltd (ALPM IN) Data for FY2010 not available. Company listed in 2011.
Accounting cycle:
Abbott India Ltd (BOOT IN) Ends 12/31 for FY2010-FY2013.
Ends 03/31 for FY2014-FY2019.
Pharmaceuticals
Accounting cycle:
GlaxoSmithKline Pharmaceuticals Ltd (GLXO IN) Ends 12/31 for FY2010-FY2014.
Ends 03/31 for FY2015-FY2019.
Accounting cycle:
Goodyear India Ltd (GDYR IN) Ends 12/31 for FY2010-FY2015.
Ends 03/31 for FY2016-FY2019.
Tyres
Accounting cycle:
MRF Ltd (MRF IN) Ends 09/30 for FY2010-FY2014.
Ends 03/31 for FY2015-FY2019.

3.2. Variables definition 3.3. Objective of the study

The primary objective that has been considered for The foremost objective of this research is to
this study is to analyse and examine the content of examine the modalities and evidence that EVA is
relative information on traditional performance superior to traditional performance measures for
measures and EVA. With intent to achieve this, a firm’s performance. With this intent, the
the study is conducted in two phases. In the first traditional measures of performance and EVA
phase of the study, market value added (MVA) is information content are analysed.
used as the dependent variable. Economic value On the basis of this objective, the below
added (EVA), return on capital employed (ROCE), hypotheses are developed.
return on net worth (RONW), EPS, net operating First phase:
profits after taxes (NOPAT), and cash flow from H1a: The relative information content of EVA is
operations (OCF) are used as explanatory variables. superior to traditional performance measures
In the second phase of the study, the (NOPAT, ROE, ROCE, EPS, and OCF) in explaining
dependent variable is annual stock returns. the market value of Indian companies.
The independent variables are EPS, ROA, ROE, and H2a: EVA adds more information content
EVA_CE. The variable EVA has been scaled by beyond the NOPAT, ROE, ROCE, EPS, and OCF in
the capital employed to find out EVA_CE for analysis explaining the market value of firms.
to provide meaningful comparison. Second phase:
ROE/RONW is another measure of profitability H1b: EVA has more relevant information
that focusses on the return on the shareholders’
content than traditional accounting measures (ROE,
equity. EVA is defined as the surplus that is left after
ROA, and EPS) in explaining stock returns.
an appropriate charge is made for the capital that is
H2b: EVA has more incremental information
employed in the firm. It is calculated as:
content than traditional accounting measures in
explaining stock returns.
𝐸𝑉𝐴 = 𝑁𝑂𝑃𝐴𝑇 – (𝑊𝐴𝐶𝐶 ∗ 𝐶𝐸) (1)

where, 3.4. The model specifications

𝑁𝑂𝑃𝐴𝑇 = 𝐸𝐵𝐼𝑇 – 𝑇𝑎𝑥𝑒𝑠 𝑜𝑛 𝐸𝐵𝐼𝑇
The following analysis is conducted using R tool:
𝐶𝐸 = 𝑆ℎ𝑎𝑟𝑒ℎ𝑜𝑙𝑑𝑒𝑟𝑠’ 𝑓𝑢𝑛𝑑𝑠 + 𝐿𝑜𝑛𝑔 𝑇𝑒𝑟𝑚 𝐷𝑒𝑏𝑡
1) Pooled regression (OLS);
+ 𝑆ℎ𝑜𝑟𝑡 𝑇𝑒𝑟𝑚 𝐷𝑒𝑏𝑡 2) Least square dummy variables (LSDV);
+ 𝑃𝑟𝑒𝑓𝑒𝑟𝑟𝑒𝑑 𝐸𝑞𝑢𝑖𝑡𝑦 3) Panel data regressions (fixed effects model,
𝑊𝐴𝐶𝐶 = 𝑤𝑑 ∗ 𝐾𝑑 ∗ (1 − 𝑡) + 𝑤𝑝 ∗ 𝐾𝑝 + 𝑤𝑒 ∗ 𝐾𝑒 random effects model).
Hausman test for endogeneity is used to test
Ke is calculated by the capital asset pricing to decide whether fixed or random effects model
model which is used for pricing risky securities and would be more appropriate.
describes the relationship between systematic risk The data is tested for pitfalls of regression,
and expected return on assets. The formula used is multicollinearity, and auto-correlation in the error
given as: term.
The below regression models are deployed in
𝐾𝑒 = 𝑅𝑓 + 𝛽 ∗ (𝑅𝑚 – 𝑅𝑓 ) (2) the first phase of the study.
Relative information content of each variable
where, tested by univariate regression model:
𝑅𝑓 – 𝑅𝑖𝑠𝑘 𝑓𝑟𝑒𝑒 𝑟𝑎𝑡𝑒
Model 1a
𝑅𝑚 – 𝑅𝑓 – 𝑀𝑎𝑟𝑘𝑒𝑡 𝑟𝑖𝑠𝑘 𝑝𝑟𝑒𝑚𝑖𝑢𝑚
𝑀𝑉𝐴𝑖𝑡 = 𝛽0 + 𝛽1 𝑁𝑂𝑃𝐴𝑇𝑖𝑡 + 𝜀𝑖𝑡 (3)
Rm, Rf values are directly taken from Bloomberg.
β – Beta of the stock. Raw beta values are taken Model 2a
from Bloomberg which uses BSE SENSEX as an index,
2 years, weekly data of stock, and index for 𝑀𝑉𝐴𝑖𝑡 = 𝛽0 + 𝛽1 𝑅𝑂𝐶𝐸𝑖𝑡 + 𝜀𝑖𝑡 (4)
regression of past stock returns on index returns.
Kd and Kp values are directly taken from
Bloomberg.

184
 Corporate Ownership & Control / Volume 18, Issue 1, Autumn 2020

Model 3a Comparison of the coefficient of determination

(R2) of the regression analysis helps in determining
𝑀𝑉𝐴𝑖𝑡 = 𝛽0 + 𝛽1 𝑅𝑂𝑁𝑊𝑖𝑡 + 𝜀𝑖𝑡 (5) as to which variable better explains the variation in
the stock returns and therefore has the information
Model 4a content that is more relevant.
With the objective of testing the incremental
𝑀𝑉𝐴𝑖𝑡 = 𝛽0 + 𝛽1 𝐸𝑃𝑆𝑖𝑡 + 𝜀𝑖𝑡 (6) information content of the accounting measures and
EVA, the following multiple regression models are
Model 5a used:

𝑀𝑉𝐴𝑖𝑡 = 𝛽0 + 𝛽1 𝑂𝐶𝐹𝑖𝑡 + 𝜀𝑖𝑡 (7) Model 5b

Model 6a 𝑅𝑖𝑡 = 𝛽0 + 𝛽1 𝑅𝑂𝐴𝑖𝑡 + 𝛽2 𝑅𝑂𝐸𝑖𝑡 + 𝛽3 𝐸𝑃𝑆𝑖𝑡 + 𝜀𝑖𝑡 (15)

𝑀𝑉𝐴𝑖𝑡 = 𝛽0 + 𝛽1 𝐸𝑉𝐴𝑖𝑡 + 𝜀𝑖𝑡 (8) Model 6b

where, 𝑅𝑖𝑡 = 𝛽0 + 𝛽1 𝑅𝑂𝐴𝑖𝑡 + 𝛽2 𝑅𝑂𝐸𝑖𝑡 + 𝛽3 𝐸𝑃𝑆𝑖𝑡

MVAit – MVA amount for the company i in (16)
+ 𝛽4 𝐸𝑉𝐴𝑖𝑡 /𝐶𝐸𝑖𝑡 + 𝜀𝑖𝑡
period t;
EVAit – EVA amount of company i in period t;
NOPATit – net operating profits after taxes for The differential in the value of R2 from
the company i in period t; Regression models 5b and 6b indicate the
ROCEit – the ratio of earnings before taxes by incremental information content of EVA.
capital employed for the company i in period t;
RONWit – the ratio of net income after tax by 3.5. Theory on panel data
net worth for company i in period t;
EPSit – net income by the total number of shares Cross-sectional time-series panel data is a dataset
outstanding for a company i in period t. where the behaviour of the respective entities across
Further, for testing H2a with respect to time is observed. The entities can be counties/states,
incremental content for EVA, NOPAT, ROCE, RONW, countries, firms, or even individuals. Balanced panel
EPS, and OCF, multiple linear regression models are data has the even and same count of observations
used. The current study makes use of two separate (time) with respect to each unit, cross-sectional.
models of multiple regression, one including all Observations that exhibit the same cross-sectional
explanatory variables and another excluding of EVA: units across time periods or individual, constitute
panel data regression models. As such, the panel
Model 7a data assists in controlling non-observable and
non-measurable variables such as culture or
𝑀𝑉𝐴𝑖𝑡 = 𝛽0 + 𝛽1 𝑁𝑂𝑃𝐴𝑇𝑖𝑡 + 𝛽2 𝑅𝑂𝐶𝐸𝑖𝑡 + the differences amongst multiple firm’s business
(9) practices. that are factors that vary across entities
𝛽3 𝑅𝑂𝑁𝑊𝑖𝑡 + 𝛽4 𝐸𝑃𝑆𝑖𝑡 + 𝛽5 𝑂𝐶𝐹𝑖𝑡 + 𝜀𝑖𝑡
but not over time like federal regulations, national
policies, international agreements, etc. It factors in
Model 8a for the individual heterogeneity.
Mentioned below are the two tools that are
𝑀𝑉𝐴𝑖𝑡 = 𝛽0 + 𝛽1 𝑁𝑂𝑃𝐴𝑇𝑖𝑡 + 𝛽2 𝑅𝑂𝐶𝐸𝑖𝑡 + used for analysing panel data:
(10) 1) random effects (RE);
𝛽3 𝑅𝑂𝑁𝑊𝑖𝑡 + 𝛽4 𝐸𝑃𝑆𝑖𝑡 + 𝛽5 𝑂𝐶𝐹𝑖𝑡 + 𝛽6 𝐸𝑉𝐴𝑖𝑡 + 𝜀𝑖𝑡
2) fixed effects (FE).
Fixed effects – every individual entity exhibits
The below regression models are deployed in individual characteristics that are specific to itself
the second phase of the study. which are time-invariant. These account for
Univariate equations in regression used to test individual heterogeneity, factors vary across entities
the information content (relative) of variables: but not over time. They cannot be measured or
observed.
Model 1b For countries, it could be their trade practices,
political system, or regulatory environment. For
𝑅𝑖𝑡 = 𝛽0 + 𝛽1 𝑅𝑂𝐴𝑖𝑡 + 𝜀𝑖𝑡 (11) companies, it could be their leadership style, culture.
Culture – a soft concept that comprises
Model 2b long-standing, majorly implicit values that are
shared, assumptions, and beliefs that influence
𝑅𝑖𝑡 = 𝛽0 + 𝛽1 𝑅𝑂𝐸𝑖𝑡 + 𝜀𝑖𝑡 (12) attitudes, behavior, and meaning in a firm.
Decision-making style – varied decision-making
Model 3b styles like autocratic, free reign can impact
decision-making – both long term and short term,
𝑅𝑖𝑡 = 𝛽0 + 𝛽1 𝐸𝑃𝑆𝑖𝑡 + 𝜀𝑖𝑡 (13) implementation of decisions, and the overall
management of the firm.
Model 4b Leadership style – leadership style has a major
impact too – diffuse vs. clear, democratic vs.
𝑅𝑖𝑡 = 𝛽0 + 𝛽1 𝐸𝑉𝐴𝑖𝑡 /𝐶𝐸𝑖𝑡 + 𝜀𝑖𝑡 (14) dictatorial leadership.

185
 Corporate Ownership & Control / Volume 18, Issue 1, Autumn 2020

Ability to change – the firm’s willingness to Hausman test – RE model assumes no

adopt change, willingness to take the risk, meeting correlation between ui and independent variables.
goals, taking up risky projects. Violation of the above assumption may lead to
How people work together – defining of goals, biased coefficient estimates. RE is efficient than FE
organization structure, defining of roles, formal vs. as more degrees of freedom are there. But FE is
informal relationships amongst peers in the unbiased. There is a trade-off bias versus efficiency.
organization. Hausman test can be used to choose RE or FE model.
Beliefs about personal “success” – there are The null hypothesis (H0): Cross-section errors (ui)
organizations that treat individuals like “stars” and that are not correlated with any of the regressors (Xi).
thrive on teamwork or organizations where The preferred model is random effects.
employees grow thanks to connections with Individual characteristics (heterogeneity) are
superiors. exogenous – independent with respect to regressors.
Growth stage of a company – established, more RE model is reliable with this assumption.
mature companies have control more structural in The alternative hypothesis (HA): ui and
nature and properly framed and defined processes, regressors (Xi) are correlated. The fixed effects model
and procedures that are signs of certain is more appropriate.
performance vs. start-ups that have less structured A significant value of chi-square will reject
ways of conducting business. the null hypothesis of no correlation of cross-section
Pooled ordinary least squares regression – errors (ui) with the regressors meaning the RE model
ignores the fixed effects. The unobserved is inappropriate.
heterogeneity across entities is not taken care of in Durbin Watson test – OLS regression has one of
the model. The equation for the OLS model is the assumptions – the covariance between εi, the
represented by: random term for observation i, and εj, the random
term for observation j, is zero.
𝑌𝑖𝑡 = 𝛽0 + 𝛽1 𝑋1𝑖𝑡 + 𝛽2 𝑋2𝑖𝑡 + 𝜀𝑖𝑡 (17)
𝐶𝑜𝑣 (𝜀𝑖 , 𝜀𝑗 ) = 0 (22)
In the fixed effects model, the regression model
intercept is permitted to differ among individuals If this assumption doesn’t hold true,
to show the features that are unique for individual the problem of autocorrelation exists. Durbin
entities. Watson test is used to detect serial autocorrelation
Least squares dummy variable model (LSDV) – in error term only under the assumption that
LSDV is in effect FE model using dummy variables. the stated error term follows the first-order
Dummy variables are used to capture the fixed autoregressive (AR1) process.
effects. The equation for the LSDV model is Autoregressive process of order 1 (AR1) is
represented by: represented by:

𝑌𝑖𝑡 = 𝛽0 + 𝛽1 𝐷1 + 𝛽2 𝐷2 + 𝛽3 𝐷3 + 𝑌𝑡 = 𝛽1 + 𝛽2 𝑋𝑡 + 𝜀𝑡
(18) (23)
𝛽4 𝐷4 + 𝛽𝑚 𝐷𝑚 + 𝛽𝑘 𝑋𝑖𝑡 + 𝜀𝑖𝑡 𝜀𝑡 = 𝜌 𝜀𝑡−1 + 𝑢𝑡

Fixed effects within group differentiator – 4. ANALYSIS AND RESULTS

it removes the fixed effects, so it is a less efficient
model. As it was shown above that the intercepts 4.1. Descriptive statistics
have the fixed effects, the equation is represented by:
The table below provides a snapshot of the
𝑌𝑖𝑡 = 𝛽0𝑖 + 𝛽1 𝑋1𝑖𝑡 + 𝛽2 𝑋2𝑖𝑡 + 𝜀𝑖𝑡 (19) descriptive statistics in a summarized form for MVA
which is the dependent variable and six other
Random effects model – the fixed effect models
explanatory variables that are relevant for the first
assume that the difference between 1 and 2 phase of the study. The statistics show that all
individuals is fixed. The random effects model
the performance measures have a mean value that is
assumes that this difference is random. positive. The mean value of MVA is INR 353450
The composite error term wit has two million for Indian companies. The mean value of
parts/components: ui, that is the individual-specific EVA is positive 34 million that goes to show that
error term or cross-section, and εit, which is the
most of the firms in India are earning higher than
combination of cross-section error components and the cost of capital. Table 2 also shows that the value
time series. of the median for all the performance measures
Rho is a time-wise correlation in error term wit. is positive. Another thing to observe is the MVA
(353450) has the highest median and mean (153736)
𝜌 = 𝑐𝑜𝑟𝑟𝑒𝑙 (𝑤𝑖𝑡 , 𝑤𝑖𝑠 ) (20)
that is followed by EVA, OCF, NOPAT in million
terms. In percent terms, the mean ROE (23.72) is
Time-wise covariance in error term wit.
higher than the mean ROCE (19.36). EVA stands 4th
in terms of ranking based on mean and
𝑐𝑜𝑣 (𝑤𝑖𝑡 , 𝑤𝑖𝑠 ) = (𝑠𝑖𝑔𝑚𝑎_𝑢)2 (21)
median values.
The standard deviation is highest (685791) for
where,
MVA and lowest (30388) for EVA in million terms.
(sigma_u)2 = variance of individual-specific
In percent terms, ROCE (19.25) has a lower standard
random variable ui;
deviation than ROE (26.72). EVA (30388) has a lower
(sigma_e)2 = variance of residuals (overall error
standard deviation than OCF (48855), and NOPAT
term) εit.
(35933) amongst the regressors. The range is highest
(6869107) for MVA and lowest (395069) for OCF in

186
 Corporate Ownership & Control / Volume 18, Issue 1, Autumn 2020

million terms. In percent terms, ROE (438.56) has a value than NOPAT (337456) amongst the regressors.
higher range than ROCE (209.13). EVA (376670) has EVA (-190214) has the lowest minimum value
a lower value for range than OCF (395069), a higher amongst the regressors.

Table 2. Descriptive statistics

Variable MVA NOPAT ROCE ROE EPS OCF EVA

Mean 353450 21688 19.36 23.72 25.75 25653 34
Standard error 31975 1675 0.90 1.25 1.98 2278 1417
Median 153736 8539 15.49 21.96 16.24 8599 1305
Standard deviation 685791 35933 19.25 26.72 42.37 48855 30388
Sample variance 470309191631 1291172410 370.67 713.89 1795.49 2386830157 923436885
Kurtosis 26 15 13.61 25.31 20.18 15 14
Skewness 4 3 2.99 1.35 -0.75 4 0
Range 6869107 337456 209.13 438.56 587.45 395069 376670
Maximum 6590871 284387 153.44 243.31 216.76 329725 186457
Minimum -278236 -53069 -55.68 -195.25 -370.69 -65344 -190214
Sum 162587098 9976556 8904.0 10910.71 11845.1 11800154 15700
Count 460 460 460 460 460 460 460

4.2. Descriptive statistics explanatory and dependent highest (587.45) for EPS and lowest (100.27) for
variables ROA. EVA_CE (114.21) has a lower value for range
than the annual stock return (359.29) and
The first set of variables ROE (438.56).
MVA, NOPAT, OCF, and EVA are in INR millions, EPS Shapiro-Wilk test – the test of normality of data.
figures are in INR million per share. ROE and ROCE The null hypothesis: Data is normally
are in percentages. distributed.
The second set of variables This test is conducted for variables studied for
The statistics reflect that all performance the two regression equations.
measures having a mean value that is positive. 1) The first set of variables:
The annual stock return has a mean value of 20.18 p-value = 0 < .05 = α, and so, the null hypothesis
percent for Indian companies. Table 2 also reflects is not supported. It is inferred at a 95% confidence
that the median value for all the performance that the data for the variables under study is not
measures is positive. In percent terms, ROE is normally distributed.
the highest median and mean which is followed by 2) The second set of variables:
stock return, EVA_CE, and ROA. p-value = 0 < .05 = α, and so, the null hypothesis
Another observation is that the standard is not supported. It is inferred at a 95% confidence
deviation is highest (43.75) for annual stock return that the data for the variables under study is not
and lowest (9.46) for ROA. EVA_CE (15.62) stands normally distributed.
4th rank in terms of standard deviation. The range is

Table 3. Shapiro-Wilk test of dependent and explanatory variables

MVA NOPAT ROCE ROE EPS OCF EVA

W-stat. 0.5399 0.6016 0.7014 0.6891 0.7190 0.5895 0.6729
p-value 0 0 0 0 0 0 0
Alpha 0.05 0.05 0.05 0.05 0.05 0.05 0.05
Normal no no no no no no no

4.3. Correlation matrix but weak correlation with MVA. Largest correlation
value was observed in between NOPAT, OCF (0.885)
The pairwise correlations between independent and and ROCE, ROE (0.705). It is significant to make
dependent variables are depicted. a note that EVA scores better than traditional
1) The first set of variables. accounting measures such as NOPAT and OCF which
It is observed that the mentioned variables are does not support the claim of advocates of EVA
all positively correlated. NOPAT (0.804) has the which states that EVA is closely associated with
highest correlation with MVA followed by OCF (0.679) the MVA.
and EVA (0.677). ROCE, EPS, and ROE have a positive

Table 4. Correlation matrix of dependent and explanatory variables.

MVA NOPAT ROCE ROE EPS OCF EVA

MVA 1
NOPAT 0.804 1
ROCE 0.309 0.107 1
ROE 0.199 0.075 0.705 1
EPS 0.222 0.210 0.182 0.047 1
OCF 0.679 0.885 0.061 0.031 0.199 1
EVA 0.677 0.515 0.385 0.294 0.226 0.297 1
Notes: MVA – market value added; NOPAT – net operating profit after taxes; ROCE – return on capital employed; ROE – return
on equity; EPS – earning per share; OCF – operating cash flows; EVA – economic value added.

187
 Corporate Ownership & Control / Volume 18, Issue 1, Autumn 2020

2) The second set of variables. noted which rejects the claim of EVA advocates that
A positive correlation amongst all variables is EVA has high correlation with equity returns.
observed. EVA/CE has positive correlation with
the stock return (0.142) but the small correlation in 4.4. Relative information content test
comparison to ROA (0.195) and the ROE (0.224). EPS
(0.063) have a positive but a very weak correlation 4.4.1. Phase 1: The first set of variables
with stock return. Highest correlation value is
observed between ROA, EVA/CE (0.809) and ROE, Below is shown the regression output for Model 1a
EVA/CE (0.685). The under-performance of EVA on with NOPAT as independent variable and MVA as
traditional accounting measures (ROA and ROE) in dependent variable. The below 3 tests are conducted
terms of its correlation with stock return may be for Models 1a-6a for the first set of variables.

Oneway (individual) effect within the model

Balanced panel: n = 46, T = 10, N = 460

Residuals:
Min. 1st Qu. Median 3rd Qu. Max.
-1179919.2 -56341.7 -8076.4 54858.5 1716064.3

Coefficients:
Estimate Std. Error t-value Pr (> |t|)
NOPAT 12.05288 0.67736 17.794 <2.2E-16***
Signif. Codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Oneway (individual) effect random effect model (Swamy-Arora’s transformation)

Balanced panel: n = 46, T = 10, N = 460
Effects:
Var Std. Dev. Share
Idiosyncratic 6.581E+10 2.565E+05 0.395
Individual 1.007E+11 3.173E+05 0.605
Theta: 0.7523

Residuals:
Min. 1st Qu. Median 3rd Qu. Max.
-1438842 -62752 -28034 41297 1970391

Coefficients:
Estimate Std. Error z-value Pr (> |t|)
(Intercept) 7.6885E+04 5.0431E+04 1.5245 0.1274
NOPAT 12.752 6.2415E-01 20.4307 < 2E-16***
Signif. Codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’

Hausman test
Data: MVA ~ NOPAT
Chisq = 7.0565, df = 1, p-value = 0.007898

Alternative hypothesis: one model is inconsistent

The relative information content test is Relative information content test may be
reported for all independent variables. The measured by R2 is also presented. The results of
assessment is based on the six regressions separate the test exhibit that NOPAT better explains the
for all measures of performance like ROCE, NOPAT, variation in market value added of firms in India
ROE, EPS, OCF, and EVA. with R2 equal to 43.4%. Also, OCF shows
a significantly higher R2 (34.43%) that is followed
4.5. Fixed effects model by EPS and EVA. ROCE and ROE have very low R2
value and negative value for adjusted R2 indicating
both the accounting measures contribute negatively
Table 5 is extracted from the above regression
to the explanatory power of MVA. The coefficient
output based on fixed effects model of panel data.
values for ROCE and ROE are also found to be
This table shows the coefficient, R2, Adjusted R2 of
negative which indicates that an increase in ROE or
each variable along with F-statistics.
ROCE results in a decrease in market value added
It can be observed that all the regressions
of Indian companies. The coefficients for ROE and
except ROCE (3.376) and ROE (0.174) are significant
ROCE are also not statistically significant at the
according to F-statistics at 0.1 percent significant 0.1 percent level. ROCE, ROE, and EPS have negative
level. The p-value results also talk about the six adjusted R2 values.
variables that are explanatory, all measures of It can be concluded that EVA, a measure of
performance are exhibit statistically significance performance based on value, stands third in terms
except ROCE (0.067) and ROE (0.677) at 0.1 percent of adjusted R2 suggesting that earnings have
level of significance. The coefficients for all dominance in explaining the variation in market
independent variables are positive except ROCE and value added for Indian firms.
ROE indicating an increase in NOPAT, OCF, EPS, and
EVA are going to lead to higher market value added
of companies in India.

188
 Corporate Ownership & Control / Volume 18, Issue 1, Autumn 2020

Table 5. Test results of the relative information content of NOPAT, ROCE, ROE, EPS, OCF, and EVA using
fixed effects model

NOPAT ROCE ROE EPS OCF EVA

Coefficients 12.053*** -3360.6 ‘.’ -322.24 3538.97*** 8.177*** 8.928***
p-value < 2.22e-16 0.067 0.677 1.06E-08 < 2.22E-16 < 2.22E-16
F-statistic 316.625 3.376 0.174 34.112 216.869 113.493
R2 (percent) 43.40 0.81 0.04 7.63 34.43 21.56
Adj. R2 (percent) 37.09 -10.24 -11.09 -2.66 27.13 12.82
Note: ‘.’, *, **, *** Significant at 10, 5, 1, and 0.1 percent levels, respectively.

4.6. Random effects model ROE have very low R2 value and negative value for
adjusted R2 indicating both the accounting measures
Table 5 is extracted from the above regression contribute negatively to the explanatory power of
output based on the random effects model of panel MVA. The coefficient values for ROCE and ROE is
data. This table shows the intercept, coefficient, R2, also found to be negative which indicates that
Adjusted R2 of each variable along with Chisq. an increase in ROE or ROCE results in a decrease in
It can be observed that all the regressions market value added of Indian firms.
except ROCE (0.399) and ROE (0.000) are significant It can be concluded that EVA, which is
according to Chisq at 0.1 percent significant level. a performance measure that is value-based, stands
The p-value results also exhibit and suggest a high third in terms of adjusted R2 suggesting that
statistical significance for all six explanatory corporate earnings dominate in explaining the
variables in form of relative information content, variation in market value added for Indian firms.
except ROCE (0.527) and ROE (0.990) at a 0.1 percent
level of significance. The coefficients for all 4.7. Hausman test
independent variables are positive except ROCE and
ROE indicating an increase in NOPAT, OCF, EPS, and It is conducted for the six Regression models 1a-6a
EVA will eventually increase the market value added to find out whether fixed effects or random effects
of companies in India. model is more appropriate for each of the models.
The relative information content test as It is found that NOPAT, EVA, and ROE performance
measured by R2 is also presented. The results of measures are better represented by the fixed effects
the test recommend that NOPAT has the highest model, whereas ROCE, EPS, and OCF regressions can
ability to elucidate the variation in market value be interpreted using the random effects model. It is
added of companies in India with R2 equal to 47.68 observed from Table 5 and Table 6, that both
percent. Next, OCF has a significantly larger R2 the fixed and random effects models give similar
(36.14 percent) followed by EVA and EPS. ROCE and interpretations for the six performance measures.

Table 6. Test results of Hausman test for fixed effects and random effects models for NOPAT, ROCE, ROE,
EPS, OCF, and EVA performance measures

NOPAT ROCE ROE EPS OCF EVA

Chisq 7.057 3.420 15.648 0.001 1.199 17.124
p-value 0.008 0.064 0.000 0.978 0.274 0.000
Decision Fixed Random Fixed Random Random Fixed

So, it can be concluded from the first phase of 4.7.1. Phase 2: The second set of variables
this study, using the first set of independent
variables, that EVA underperforms in comparison Below is shown the regression output for Model 1b
with NOPAT and OCF. This makes one reject the null with ROA as an independent variable and stock
hypothesis that traditional methods have less return as a dependent variable. The below 3 tests are
relevant information content than EVA. Therefore, it conducted for Models 1b-4b for the second set of
is evident that the variation in market value added is variables.
better explained by traditional measures than EVA,
a value-based measure.

Oneway (individual) effect within the model

Call:
plm (formula = Return ~ ROA, data = myframe, model = “within”, index = c (“Company”, “Year”))
Balanced panel: n = 46, T = 10, N = 460

Residuals:
Min. 1st Qu. Median 3rd Qu. Max.
-188.112 -24.130 -4.285 16.170 139.134

Coefficients:
Estimate Std. Error t-value Pr (> |t|)
ROA 2.41862 0.44019 5.4944 6.868E-08***
Signif. Codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

189
 Corporate Ownership & Control / Volume 18, Issue 1, Autumn 2020

Oneway (individual) effect random effect model (Swamy-Arora’s transformation)

Total sum of squares: 878540
Residual sum of squares: 845240
R2: 0.037903
Adj. R2: 0.035802
Chisq: 18.0434 on 1 DF
p-value: 2.1592E-05

Hausman test
Data: Return ~ ROA
Chisq = 15.49, df = 1, p-value = 8.296E-05.

Alternative hypothesis: one model is inconsistent

Durbin-Watson test for serial correlation in panel models

Data: Return ~ ROA
DW = 2.3224, p-value = 0.9998

Alternative hypothesis: serial correlation in idiosyncratic errors.

Every independent variable’s relative The relative information content test as

information content test is reported. The measured by R2 is also presented. The test results
assessment is based on the 4 separate regressions suggest that ROA has the greatest ability to explain
measures for each measure of performance, the variation in equity return of companies in India
i.e., ROA, ROE, EPS, EVA_CE. with R2 equal to 6.81 percent followed by
EVA_CE (6.36) and ROE (6.33). It is depicted that
4.8. Fixed effects model ROA, EVA_CE, and ROE have almost similar R2 values
indicating that these performance measures have
Table 7 is extracted from the above regression a similar impact on the explanatory power of
output based on the fixed effects model of panel stock returns.
data. This table shows the coefficient, R2, Adjusted R,2 EPS (0.16) with the lowest value for R2 and
and Durbin Watson p-value of each variable along the highest negative value for adjusted R2 (-10.96)
with F-statistics. indicates that EPS is statistically insignificant in
It can be observed that except of EPS (0.669), all explaining the variation in stock returns.
the regressions are statistically significant as per the The beta coefficient indicates that ROA
F-statistics at 0.1%. These p-value results also provides a better explanation of variation in equity
recommend that in forms of the relative information returns than EVA_CE. It can be concluded that an
content of all four explanatory variables, all accounting measure is better and has more relevant
measures of performance are statistically significant information content than value-based measure.
except for EPS (0.414) at a 0.1 percent level of
significance. The coefficients for all independent 4.9. Durbin Watson test
variables are positive, indicating an increase in ROA,
ROE, EPS, and EVA_CE will lead to an increase in The results show that the p-value greater than
stock returns of companies in India. One unit 5 percent, which means the null hypothesis is
increase in ROA will lead to 2.419 units increase in accepted. There is no first-order serial autocorrelation
stock returns which is the highest. A unit increase in in the error term of the four Models 1b-4b.
EVA_CE will lead to 1.662 units increase in stock
returns which is lower than ROA but higher than
ROE (0.516) and EPS (0.067).

Table 7. Test results of the relative information content of ROA, ROE, EPS, and EVA_CE using the fixed
effects model

ROA ROE EPS EVA_CE

Coefficient 2.419*** 0.516*** 0.067 1.662***
p-value 0.000 0.000 0.414 0.000
F-statistic 30.189 27.917 0.669 28.066
R2 (percent) 6.81 6.33 0.16 6.36
Adj. R2 (percent) -3.57 -4.10 -10.96 -4.07
Durbin Watson p-value 0.9998 1.00 0.9998 0.9995
Note: *, **, *** significant at 10, 5, 1 and 0.1 percent levels, respectively.

4.10. Random effects model also recommend that in forms of the relative
information content of all four explanatory
Table 7 is extracted from the above regression variables, all measures of performance are
output based on the random effects model of panel statistically significant except EPS (0.177) at a 0.1%
data. This table shows the intercept, coefficient, R2, level of significance. The coefficients for all
Adjusted R2 of each variable along with Chisq. independent variables are positive, indicating an
It can be observed, in effect, all the regressions increase in ROA, ROE, EPS, and EVA_CE will lead to
measures except EPS (1.824) are statistically an increase in stock returns of companies in India.
significant as per Chisq at 0.1%. The p-value results One unit increase in ROA will lead to 0.9 units

190
 Corporate Ownership & Control / Volume 18, Issue 1, Autumn 2020

increase in stock returns which is the highest. A unit (equations (9) and (10)) are conducted. One model
increase in EVA_CE will cause 0.399 units to increase has only the traditional accounting measures as
in stock returns which is lower than ROA but higher independent variables and the second model has
than ROE (0.367) and EPS (0.065). EVA added as another independent variable.
The test results suggest that ROE holds The dependent variable is the market value added.
the greatest explanation power for the stock returns VIF – test for multicollinearity
variation for companies in India with R2 equal to To detect the presence of multicollinearity
5.02 percent followed by ROA (3.79) and amongst the regressors, variance inflation factor
EVA_CE (2.03). The adjusted R2 value is highest for (VIF) values are analysed. A general rule is that VIF
ROE followed by ROA, EVA_CE. EVA_CE stands third of 8-10 indicates the existence of collinearity of that
in terms of R2 and Adjusted R2 values indicating that variable with other independent variables. VIF values
accounting performance measures have a greater of all independent variables are in range 1-7 and
impact on the explanatory power of stock returns. the largest value of 6.932 for NOPAT thus indicating
It can be concluded that the accounting measures a low degree of multicollinearity amongst the
have more relevant information content than explanatory variables.
value-based measure. Pooled ordinary least square regression
Hausman test Table 8 shows the incremental information
It is conducted for the four Regression content test of all six explanatory variables based on
models 1b-4b to find out whether fixed effects or the pooled ordinary least square regression.
random effects model is more apt for each of The model results on a whole suggest that the two
the models. It is found that ROA, ROE, and EVA_CE models are statistically significant as F-statistic
performance measures are better represented by values (211.92 and 235.05) are greater than
the fixed effects model, whereas EPS regression can the F-critical value at a 0.1 percent significance level.
be interpreted using the random effects model. Results related to coefficients reveal that NOPAT,
So, it can be concluded from the second phase ROCE, and EVA have a statistically significant
of this study, using the second set of independent relation at 0.1%. The coefficient of OCF indicates
variables, that EVA underperforms in comparison that it is statistically significant at a 5% significance
with ROA. This causes the rejection of the null level. OCF with MVA has an association that is
hypothesis, that EVA has more relevant information negative, however, NOPAT, ROCE, and EVA have
content than traditional measures. Hence, the a positive correlation with MVA. It is observed that
variation in equity returns is better explained R2 increases from 70.01 to 75.69 percent if EVA is
by traditional measures than EVA, a value-based included in the model. Further, the adjusted R2
measure. increased from 69.68 to 75.37 percent between
the first model with only accounting measures and
4.11. Incremental information content test the second model which includes EVA. We can
conclude that the increase of 5.68 percent in R2 is
This test is conducted to validate whether one a small value but it is having statistical significance
variable provides more information content than indicating that EVA contributes to describing
another. the variation in market value added of Indian firms.
However, the pooled OLS regression model executed
for panel data ignores the fixed effects, which means
4.11.1. Phase 1: The first set of variables
that it does not take care of unobserved
heterogeneity across entities.
In order to determine the incremental information
content test of EVA, two regression models 7a-8a

Table 8. Test results of the incremental information content of EVA, NOPAT, ROCE, ROE, EPS, OCF, based on
pooled OLS regression

Model 7a – Equation (9) Model 8a – Equation (10)

Intercept -122030*** 24317.75
NOPAT 16.82*** 10.04***
ROCE 8700*** 5563.96***
ROE -944.32 -1313.61
EPS 282.48 -181.26
OCF -1.66* 1.53*
EVA - 7.48***
F-statistic 211.92 235.05
p-value < 2.22E-16 < 2.22E-16
R2 (percent) 70.01 75.69
Adjusted R2 (percent) 69.68 75.37
Note: *, **, *** significant at 10, 5, 1, and 0.1 percent levels, respectively.

Lagrange multiplier test – (Breusch-Pagan) is Data: MVA ~ NOPAT + ROCE + ROE + EPS + OCF +
conducted to test whether there is a panel effect in + EVA;
the data. The null hypothesis is rejected indicating Chisq = 122.98, df = 1, p-value < 2.2E-16.
panel regressions are better than OLS regressions. Alternative hypothesis: significant effects.
Lagrange multiplier test – (Breusch-Pagan) for The results of the F-test show that individual
balanced panels: and time effects are present in the panel data.

191
 Corporate Ownership & Control / Volume 18, Issue 1, Autumn 2020

Table 9. Individual fixed effects for the companies which are the intercepts in the regression equation

Companies and its’ intercept

Company Apollo Tyres Ashok Leyland Godrej Consumer Products Hindustan Unilever
Value -15638.5 80677.4 317779.1 1656058.7
Company Asian Paints Aurobindo Pharma India Cements Infosys
Value 633259.9 69129.2 -21826.9 671611.6
Company Bajaj Auto Balkrishna Industries Jindal Stainless Jindal Steel & Power
Value 250811.8 50749.8 4153.6 -145140.3
Company Berger Paints Cadila Healthcare JK Cement JK Tyre & Industries
Value 185762.1 170532.7 3620.3 4666.3
Company Ceat Cipla JSW Steel Kansai Nerolac Paints
Value -26967.7 178283.0 -318985.5 173309.3
Company Colgate-Palmolive Dabur India Lupin Mahindra & Mahindra
Value 598842.7 383170.6 233761.5 175197.3
Dr. Reddy’s
Company Divi’s Laboratories Marico Maruti Suzuki
Laboratories
Value
176654.1 77201.1 290567.5 279309.7
Glenmark
Company Emami Mindtree Natco Pharma
Pharmaceuticals
Value
238660.9 82872.9 112289.1 94761.5
Oracle Financial Services
Company Ramco Cements Tata Steel BSL Tech Mahindra
Software
Value
18588.3 36048.8 57496.6 92829.5
Company SML ISUZU SAIL Torrent Pharmaceuticals TVS Motor
Value 33177.6 -151649.7 101851.0 128976.6
Company Sun Pharmaceuticals TCS UltraTech Cement Wipro
Value 511818.5 1668226.4 204189.2 225622.0
Company Tata Motors Tata Steel
Value -764518.8 -604496.4

Least square dummy variable regression independent variable is statistically insignificant

Table 8 shows the results of the least square based on LSDV regression. The coefficient of EPS
dummy variable regression. indicates that it is statistically significant at
The least square dummy variables regression 5 percent level of significance. It is observed that R2
model is executed for the Model 8a of equation (10) value (88.81) and adjusted R2 (86.97) percent is
which uses 45 dummy variables for the 46 companies greater than the R2 (75.69) and Adjusted R2 (75.37)
to capture fixed effects in the companies. for pooled OLS regression for Model 8a –
The overall model results suggest that the equation (10). We can conclude that this model is an
model is statistically significant as F-statistic 61.05 improvement on pooled OLS regression since it
is greater than the F-critical value at 0.1 percent captures the fixed effects of individual companies,
significance level. Results related to coefficients however, the addition of 45 dummy variables which
exhibit that NOPAT, ROCE, and OCF are statistically results in the estimation of a high number of
significant at a 0.1 percent level of significance. parameters reduces the degree of freedom, thereby
ROCE has a negative association, whereas NOPAT decreasing the power of regression.
and OCF are positively correlated with MVA. EVA

Table 10. Least square dummy variable regression for NOPAT, ROCE, ROE, EPS, OCF, and EVA as independent
variables

Model 8a – Equation (10)

Intercept 15640
NOPAT 7.758***
ROCE -5779***
ROE 232.40
EPS 1289*
OCF 3.101***
EVA 1.15
factor(Company)Ashok Leyland 96320
factor(Company)Asian Paints 648900
factor(Company)Aurobindo Pharma 84770
F-statistic 61.05
p-value < 2.2E-16
R2 (percent) 88.41
Adjusted R2(percent) 86.97
Note: *, **, *** significant at 10, 5, 1, and 0.1 percent levels, respectively.

5. RESULT DISCUSSION with all variables and another regression model

except EVA.
5.1. Fixed effects model Table 10 is extracted from the above regression
output based on the fixed effects model of panel
data. This table shows the coefficient, R2, Adjusted R2
The fixed effects model is conducted to find out
of each variable along with F-statistics.
the incremental information content of EVA, on two
The overall results of the model suggest that
regression models 7a-8a (equations (9) and (10)),
both the models are statistically significant as

192
 Corporate Ownership & Control / Volume 18, Issue 1, Autumn 2020

F-statistic values (74.82 and 62.55) are greater than degree of multicollinearity amongst the explanatory
the F-critical value at a 0.1 percent significance level. variables.
Results about coefficients reveal that NOPAT, ROCE, The overall model results suggest that both
and OCF are statistically significant at a 0.1 percent the models are statistically significant as F-statistic
level of significance. The coefficient of EPS indicates values (9.08 and 7.81) are greater than the F-critical
that it is statistically significant at 5 percent level of value at 0.1 percent and 1 percent significance level.
significance. ROCE has a negative association Results about coefficients for Model 5b –
whereas NOPAT and OCF are positively correlated equation (15) reveals that ROE is statistically
with MVA. It is observed that R2 increases slightly by significant at 1 percent level of significance. Results
0.14 percent points if EVA is included in the model. about coefficients for Model 6b – equation (16)
Further, the adjusted R2 shows similar values reveals that ROE is statistically significant at 0.1
between the two models. EVA is statistically percent level of significance and ROA at a 5 percent
insignificant as depicted by its coefficient implying significance level. The coefficient of EPS and EVA_CE
that its impact on market value added is indicate that both statistically insignificant. EVA_CE
insignificant. We can conclude that the slight has a negative association with Stock Return whereas
increase in R2 and EVA being statistically ROE and ROA are positively correlated with Stock
insignificant means that EVA does not add any Return. An increase in ROE by 1 unit (in percentage
incremental information content to that provided by terms) will increase stock returns of the company
accounting measures in explaining the MVA of by 0.361 units (in percentage terms). An increase in
Indian companies. EVA_CE by 1 unit (in percentage terms) will decrease
stock returns of the company by 0.488 units (in
5.2. Random effects model percentage terms).
It is observed that R2 increases from 5.64 to
Table 10 is extracted from the above regression 6.42 percent if EVA_CE is included in the model.
output based on the random effects model of panel Further, the adjusted R2 increased from 5.02 to
data. This table shows the intercept, coefficient, R2, 5.6 percent between the first model with only
adjusted R2 of each variable along with Chisq. accounting measures and the second model which
The overall model results suggest that both the includes EVA. We can conclude that the slight
models are statistically significant as Chisq values increase (0.78 percent) in R2 value and EVA_CE being
(487.41 and 624.19) are greater than the respective statistically insignificant means that EVA_CE does
critical values at 0.1 percent significance level. not add any incremental information content to that
Results about coefficients reveal that NOPAT, OCF, provided by accounting measures in explaining
and EVA are statistically significant at a 0.1 percent the variation in stock return of Indian companies.
level of significance. The coefficient of OCF indicates However, the pooled OLS regression model executed
that it is statistically significant at 1 percent level of for panel data ignores the fixed effects which means
significance for equation (9). NOPAT, OCF, and EVA that it does not take care of unobserved heterogeneity
are positively correlated with MVA. across entities.
It is observed that R2 increases from 51.78 to
57.95 percent if EVA is included in the model. 5.4. Lagrange multiplier test – (Breusch-Pagan)
Further, the adjusted R2 increased from 51.24 to
57.39 percent between the first model with only It is conducted to test whether there is a panel effect
accounting measures and the second model which in the data. The null hypothesis is accepted
includes EVA. We can conclude that the increase of indicating pooled OLS regressions are better than
6.17 percent in R2 is a low value but it is statistically panel regressions.
significant, indicating that EVA contributes to Lagrange multiplier test – (Breusch-Pagan) for
explaining the variation in market value added of balanced panels:
Indian companies. Data: Return ~ ROA + ROE + EPS + EVA_CE;
Chisq = 2.9942, df = 1, p-value = 0.08356.
5.3. Hausman test Alternative hypothesis: significant effects.
So, it can be concluded from the second phase
It is conducted for the Model 8a – equation (10), of this study, based on the interpretation from
which has six independent variables to find out the pooled OLS model, using the second set of
whether fixed effects or random effects model is independent variables, that EVA_CE underperforms
more appropriate. It is found that the null in comparison with ROE and ROA. This leads to
hypothesis is rejected, which means the fixed effects the rejection of the null hypothesis that EVA_CE
model is a preferred model. has more incremental information content than
So, it can be concluded from the first phase of traditional measures. Therefore, it is evident that
this study, based on the interpretation from the variation in stock returns is better explained
the fixed effects model, using the first set of by traditional measures than EVA, a value-based
independent variables, that EVA underperforms in measure.
comparison with NOPAT and OCF. This leads to
the rejection of the null hypothesis, that EVA has 6. CONCLUSION
more incremental information content than
traditional measures. Therefore, it is evident that the A review of research has provided mixed results on
variation in market value added is better explained the dominance of EVA as a financial measure of
by traditional measures than EVA, a value-based performance over the traditional earnings-based
measure. measures. It is suggested that sometimes traditional
Variables are in range 1-4 and the highest value measures outperform the value-based measures and
of 3.874 for ROA and EVA_CE indicating a low another claim in the literature is based on the

193
 Corporate Ownership & Control / Volume 18, Issue 1, Autumn 2020

value-based measures being superior and provide firms aim to align the organisation metrics with
additional information in explaining the market market value, a measurement paradigm other than
value variations or shareholder value of Indian firms. EVA has to be considered. It is implied that there are
Addressing this question of superiority of EVA other parameters that influence the market value of
over accounting measures is the primary objective of a company and should be factored in for financial
this study. In the first phase of this study, the prime performance.
objective is to examine empirical evidence about The relative information content test with
the association of EVA in comparison with the second set of variables reveals that accounting
the traditional accounting performance measures measures outperform EVA in its association with
(NOPAT, RONW, ROCE, EPS, and OCF) with market stock returns. The results of the incremental
value added of companies in India. In the second information content test reveal that ROA and ROE
phase in this research, the purpose is to test outperform EVA, and EVA is a statistically
the claims of EVA advocates and observe the impact insignificant variable, does not add incremental
of ROA, ROE, and EPS relative to EVA_CE, information to that provided by ROA and ROE in
a value-based measure. explaining the stock returns of Indian companies.
To achieve this, a dataset of 46 Indian The findings that demonstrate the relatively
companies for the period of 10 years from 2010 to low explanatory power of all measures of
2019 was taken to conduct relative and incremental performance under analysis is largely consistent
information content of explanatory variables. with the findings of many international studies
The first set of independent variables are NOPAT, (Chen & Dodd, 1997, 2001; Biddle et al., 1998;
RONW, ROCE, EPS, OCF, and EVA, and MVA is the Maditinos et al., 2009). The results with the second
dependent variable. The second set of independent set of variables also prove the findings of many
variables are ROE, EPS, ROA, and EVA_CE, and Stock scholars that more non-financial and financial
Returns is the dependent variable. determinants must be deployed to assess the stock
Pooled ordinary least squares regression, least return performance of the companies.
squares dummy variables, and panel data The results are consistent with the prior study
regressions (fixed and random effects models) tests conducted by Chen and Dodd (2001), which
are performed. It is proved that panel data suggested that non-financial variables customer
regressions are better than pooled OLS and LSDV satisfaction, CSR, employees, product quality,
models since the power of regression is reduced due research, and development have an impact on
to ignoring fixed effects and using dummy variables the market value of the firms. As suggested by
to capture fixed effects respectively. Mishra and Suar (2010) in the paper “Does corporate
Using the dataset of 460 firm-year observations, social responsibility influence firm performance of
the incremental and relative information content tests Indian companies?”, financial performance measures
are conducted. The empirical results do not support along with non-financial performance measures,
the hypothesis that EVA is a superior performance are indicators employed to assess a firm holistic
indicator than conventional audited performance performance.
measures to explain the market value added of The study has a few limitations. Variables like
companies in India. EPS, being market dependent, need to be considered
The relative information content test with with equity returns volatility and that might change
the first set of variables reveals that NOPAT and the calculations. The data exhibits signs of
OCF outperform EVA in its association with MVA, distortion of measurement error but this is usual for
thereby not supporting the hypothesis that EVA has panel data. The companies that are dropped for
better explanatory power than other variables. reasons, mentioned previously, may lead to
The observations and findings of this research are selectivity bias. The equity returns needs to be
similar to many international studies that reject factored in for the gap between the returns,
the claim of EVA advocates about its superiority as compounded for a share, and on the ALSI index.
a value-based corporate performance measure. In conclusion, our findings do not support
The results of the incremental information content the claim of EVA advocates, that EVA is the best
test reveal that EVA is a statistically insignificant measure for performance measurement systems.
variable, does not add incremental information to It is evident that traditional measures are better
that provided by NOPAT and OCF in explaining performance indicators for financial performance
the MVA of companies in India. Also, it was found for selected companies in the Indian stock market
that one variable EVA model is able to capture only than EVA.
21.56% of the variations in MVA. It implies that if

REFERENCES
1. Ali, T. F. (2018). The influence of economic value added and market value added on corporate value. Russian
Journal of Agricultural and Socio-Economic Sciences, 74(2), 90-98. https://doi.org/10.18551/rjoas.2018-02.10
2. Bacidore, J. M., Boquist, J. A., Milbourn, T. T., & Thakor, A. V. (1997). The search for the best financial performance
measure. Financial Analysts Journal, 53(3), 11-20. https://doi.org/10.2469/faj.v53.n3.2081
3. Bao, B., & Bao, D. (1998). Usefulness of value added and abnormal economic economic earnings: An empirical
examination. Journal of Business Finance Accounting, 25(1&2), 251-264. https://doi.org/10.1111/1468-5957.00186
4. Bhasin, M. L. (2017). A study of economic value added disclosures in the annual reports: Is EVA a superior
measure of corporate performance? East Asian Journal of Business Economics, 5(1), 10-26.
https://doi.org/10.20498/eajbe.2017.5.1.10
5. Biddle, G. C., Bowen, R. M., & Wallace, J. S. (1997). Does EVA® beat earnings? Evidence on associations with stock
returns and firm values. Journal of Accounting and Economics, 24(3), 301-336. https://doi.org/10.1016/S0165-
4101(98)00010-X

194
 Corporate Ownership & Control / Volume 18, Issue 1, Autumn 2020

6. Chen, S., & Dodd, J. (1997). Economic value added (EVA™): An empirical examination of a new corporate
performance measure. Journal of Managerial Issues, 9(3), 318-333. Retrieved October 16, 2020, from
http://www.jstor.org/stable/40604150
7. Chen, S., & Dodd, J. L. (2001). Operating income, residual income and EVA: Which metric is more value relevant?
Journal of Managerial Issues, 13(1), 65-86. Retrieved from https://www.jstor.org/stable/40604334
8. de Wet, J. H. (2005). EVA versus traditional accounting measures of performance as drivers of shareholder value –
A comparative analysis. Meditari Accountancy Research, 13(2), 1-16. https://doi.org/10.1108/10222529200500009
9. Erasmus, P. (2008). Value based financial performance measures: An evaluation of relative and incremental
information content. Corporate Ownership & Control, 6(1), 66-77. https://doi.org/10.22495/cocv6i1p7
10. Fountaine, D., Jordan, J. D., & Phillips, G. M. (2008). Using economic value added as a portfolio separation criterion.
Quarterly Journal of Finance and Accounting, 47(2), 69-81. Retrieved from https://www.jstor.org/stable/40473456
11. Gupta, V. K., & Sikarwar, E. (2016). Value creation of EVA and traditional accounting measures: Indian evidence.
International Journal of Productivity and Performance Management, 65(4), 436-459. https://doi.org/10.1108/IJPPM-
01-2014-0008
12. Gupta, V. K., & Venkata Vijay Kumar, P. (2013). Value-based accounting: A performance analysis of Indian industry.
The IUP Journal of Accounting Research & Audit Practices, 12(1), 7-20. Retrieved from
https://search.proquest.com/openview/35c8689ac7ab3f8e981a0a5c8d42e753/1?cbl=54439&pq-origsite=gscholar
13. Ismail, A. (2006). Is economic value added more associated with stock return than accounting earnings? The UK
evidence. International Journal of Managerial Finance, 2(4), 343-353. https://doi.org/10.1108/17439130610705526
14. Kumar, S., & Sharma, A. K. (2011). Association of EVA and accounting earnings with market value: Evidence from
India. Asia-Pacific Journal of Business Administration, 3(2), 83-96. https://doi.org/10.1108/17574321111169795
15. Kyriazis, D., & Anastassis, C. (2007). The validity of the economic value added approach: An empirical
application. European Financial Management, 13(1), 71-100. https://doi.org/10.1111/j.1468-036X.2006.00286.x
16. Lehn, K., & Makhija, A. K. (1996). EVA & MVA as performance measures and signals for strategic change.
Strategy & Leadership, 24(3), 34-38. https://doi.org/10.1108/eb054556
17. Lin, C., & Zhilin, Q. (2008). Empirical study of integrated EVA performance measurement in China. Canadian
Social Science, 4(2), 41-48. Retrieved from https://core.ac.uk/download/pdf/236294399.pdf
18. Maditinos, I. D., Sevic, Z., & Theriou, N. G. (2009). Modelling traditional accounting and modern value-based
performance measures to explain stock market returns in the Athens Stock Exchange (ASE). Journal of
Modelling in Management, 4(3), 182-201. https://doi.org/10.1108/17465660911006431
19. Masyiyan, R. A. (2019). Analisis kinerja keuangan dengan metode economic value added (EVA), market value added
(MVA), dan financial value added (FVA). Kajian Akuntansi, 20(2), 200-210. https://doi.org/10.29313/ka.v20i2.4852
20. Mcgrattan, E. R., Rogerson, R., & Wright, R. (1997). An equilibrium model of the business cycle with household
production and fiscal policy. International Economic Review, 38(2), 267-290. https://doi.org/10.2307/2527375
21. Milunovich, S., & Tsuei, A. (1996). EVA in the computer industry. Journal of Applied Corporate Finance, 9(1),
104-116. https://doi.org/10.1111/j.1745-6622.1996.tb00108.x
22. Mishra, S., & Suar, D. (2010). Does corporate social responsibility influence firm performance of Indian
companies? Journal of Business Ethics, 95, 571-601. https://doi.org/10.1007/s10551-010-0441-1
23. O’Byrne, S. F. (1996). Eva and market value. Journal of Applied Corporate Finance, 9(1), 116-126.
https://doi.org/10.1111/j.1745-6622.1996.tb00109.x
24. Ronen, B., Pliskin, J. S., & Pass, S. (Eds.). (2018). The “evils” of traditional cost accounting. In The hospital and
clinic improvement handbook: Using lean and the theory of constraints for better healthcare delivery
(Chapter 15). https://doi.org/10.1093/med/9780190843458.003.0015
25. Shah, R., Haldar, A., & Nageswara Rao, S. V. D. (2014a). Economic value added as a performance measure –
Relative comparison with MVA. The Empirical Economics Letters, 13(1). Retrieved from
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2607437
26. Shah, R., Haldar, A., & Nageswara Rao, S. V. D. (2014b). Economic value added: A financial flexibility tool.
Corporate Ownership and Control, 12(1-8), 703-708. https://doi.org/10.22495/cocv12i1c8p2
27. Sharma, M. K., Sushil, & Jain, P. K. (2010). Revisiting flexibility in organizations: Exploring its impact on
performance. Global Journal of Flexible Systems Management, 11(3), 51-68. https://doi.org/10.1007/BF03396587
28. Stewart, G. (2010). The quest for value: A guide for senior managers. New York, NY: HarperBusiness.
29. Uyemura, D. G., Kantor, C. C., & Pettit, J. M. (1996). Evar for banks: Value creation, risk management, and
profitability measurement. Journal of Applied Corporate Finance, 9(2), 94-109. https://doi.org/10.1111/j.1745-
6622.1996.tb00118.x
30. Worthington, A. C., & West, T. (2001). Economic value‐added: A review of the theoretical and empirical
literature. Asian Review of Accounting, 9(1), 67-86. https://doi.org/10.1108/eb060736
31. Worthington, A., & West, T. (2004). Australian evidence concerning the information content of economic
value-added. Australian Journal of Management, 29(2), 201-224. https://doi.org/10.1177/031289620402900204

195