Chapter 5. Relationship of Economic Value Added and Conventional Performance Measures with Market Value Added

Transcription

1 Chapter 5 Relationship of Economic Value Added and Conventional Performance Measures with Market Value Added 5.1 Introduction Although Shareholder Value Creation has become the widely accepted corporate mission, much debate is taking place at its measurement level. As much as companies intensify to fulfill their vision of creating value for their shareholders, the quite obvious question arises i.e. which measurement metric is best among all. Lay investors and even most companies tend to focus too much on size and income based metrics such as share price (market value or market capitalization), earnings, growth in earnings, EPS, ROCE and ROE. But such metrics don t consider the cost of equity capital and are influenced by accrual accounting based conventions. Moreover, these traditional accounting measures do not take into account how much capital has been poured into the business to generate the additional income, so it is relatively easy to improve such measures simply by investing more. Thus, there are so many reasons (refer shortcomings of traditional financial performance measures discussed in Chapter 1) due to which these traditional measures have been regularly criticized as misleading, manipulative and incompetent in disclosing an organization s value creating performance. In turn, proponents of value based measures have responded with specific metrics and methodologies that claim to provide a better and reliable measurement of shareholder value creation (Armitage, 1995). For instance, US-based Consultancy firm Stern-Stewart & Company claimed that earnings, earning per share and earnings growth are misleading measures of corporate performance and the best practical periodic performance measure is Economic Value Added (EVA). Stewart (1991) argued that EVA comes closer than any other measure to capture the true economic profitability of an enterprise and is the performance measure, which is most directly linked to the shareholder value over time. To further support his claim, Stewart (1994) provided empirical evidence that EVA stands well out from the crowd as the single best measure of wealth creation on a contemporaneous basis and is almost 50% better than its closest accounting based competitors (including EPS, ROE and ROI) in explaining changes in shareholder wealth.

2 The proponents also claimed that mathematically, the EVA of a company is the net present value (NPV) of all its future EVAs. Thus, a company that continues to improve economic value added, year after year, will sooner than later, find favor with investors. Thus, over the long term, it is an improvement in EVA and not in accounting results that derives wealth creation. That is one reason why companies world over need to focus on improving their fundamental economic performance as measured by EVA. The literature for the relationship between EVA and Market Value involves a considerable debate regarding the superiority of EVA in comparison to the traditional performance measures like ROI, EPS, ROCE etc. This chapter of the dissertation is devoted to identify the result of this metric war (between traditional and value based measures of performance) empirically. It attempts to investigate Does EVA dominate Earnings in Indian corporate sector? Thus, the present study will be of immense use to financial analysts, corporate officials, researchers and policy makers who may be interested in EVA as replacement (or compliments) to earnings as key measure of corporate performance. Worth to be mentioned here that in a letter to the editor of Management Accounting, Stewart criticized studies that evaluated EVA s effectiveness in estimating value added by measuring how it explains stock returns, calling them meaningless and unimportant for the purposes of validating EVA. Stewart argued that using EVA as a proxy for MVA is what carries more importance. Thus, the present study considers MVA as a proxy for shareholder wealth created or eroded by the sample companies. 5.2 Sample Description and Database Initially, sample size of the study remained same i.e. 104 Companies (as described in section 3.2 of Chapter 3). However, while examining data, four companies namely Satyam Computer Services Ltd., Tata Steel Ltd., Sun Pharmaceutical Industries Ltd., Sesa Goa Ltd. and were identified as outliers and had to be deleted. Thus, a final sample of 100 companies was selected and studied for the subsequent analysis. Secondary data has been used for a period of 12 years i.e. from 1996 to All the relevant financial information has been sourced from the CMIE s corporate database Prowess and the data regarding share prices has been obtained from the Capitacharts of Capital Market 147

3 Publishers of India Ltd. E-Views and Statistical Package for Social Sciences (SPSS) have been used for the analysis of data. 5.3 Hypothesis of the Study Relative information content comparisons are appropriate when one desires a ranking of performance measures by information content or when making mutually exclusive choices among performance measures i.e. when only one measure can be chosen. In contrast, Incremental information content comparisons assess whether one measure provides value-relevant inferences beyond those provided by another measure, evaluating the benefit of supplemental disclosures in financial reporting (Biddle et al., 1997). As the literature contains both i.e. studies favorable to EVA as well as those which disagree with EVA to be the best predictor of MVA, the present study takes a neutral position. It tests the hypothesis that Value Based Measures as well as Traditional Financial Performance Measures have equal relative and incremental information content i.e. equal association with MVA, a surrogate of shareholder value creation. 5.4 Choice of Variables Nine independent financial variables are chosen for the purpose of the study, of which five represent Accrual Accounting based traditional performance measures, two are Value Based Performance Measures and the remaining two are Economic Variables. Accrual accounting based performance measures includes Return on Capital Employed (ROCE), Return on Net Worth (RONW), Profit after Tax (PAT), Earning per Share (EPS) and Return on Total Assets (ROTA) whereas Value based measures include Economic Value Added (EVA) and EVA in percentage terms (EVA%). Employees Productivity (Ep) and Capital Productivity (Cp) are the performance measures representing Economic Indicators. For testing the hypothesis, Market Value Added (MVA) has been taken as the dependent variable. A brief description of all these variables is given in Table Dependent Variable Market Value Added (MVA): MVA being an absolute measure assesses that how much capital a company has added to or subtracted from its shareholders investment. It is the cumulative amount by which a company is perceived to have enhanced or diminished shareholder wealth. It is based upon the logic that if the total market value of a company 148

4 is more than the capital invested in it, the company has managed to create shareholder value. However, if the market value of a company comes less than its invested capital, company has destroyed the shareholder value. MVA thus, measures the value added by the management over and above the capital invested in the company by its shareholders and lenders. For the purpose of the study, MVA is obtained by subtracting the economic capital of a corporation (book value after adjusting for economic anomalies) from its total market value i.e. what investors can take out of the company. Mathematically, MVA = Market Value of the firm Economic Capital Hence, the way in which shareholder wealth is maximized is by increasing the difference between the company s market value and its economic capital. Market value of a firm as represented by market value of its equity is arrived at by multiplying the stock price by the number of outstanding shares of the firm. Taking share price at the end of the financial year for the calculation of the market capitalization can be biased. Hence, in the present study, 364-days average market cap has been taken as proxy for the market value of equity. Market value of the firm has been taken as the sum of book value of debt and 364-days average market capitalization. MVA is the perfect summary assessment of corporate performance that shows how successful a company has been in allocating and managing resources to maximize the value of the enterprise and the wealth of its shareholders (Stewart, 1994). In the present study, MVA being the surrogate of shareholder wealth addition has been taken as the dependent variable Explanatory Variables i. Return on capital employed (ROCE) ROCE measures the profit which a firm earns on investing a unit of capital and tells whether the company s borrowing policy was wise economically and whether the capital had been employed fruitfully (Maheshwari, 2004). If the long-term return of a business enterprise is not satisfactory in any case, then the deficiency needs to be corrected and the activity can be abandoned for a more favourable one (Kishore, 2002). Obviously, it is quite impractical to assess profits or profit growth properly without relating them to the amount of funds (capital) that were employed in making profits. 149

5 ROCE is one of the most important profitability ratios which assess how much the capital invested has earned during the period. It is determined by dividing net profit or income by the capital employed or investment made to achieve that profit. The ROCE is determined by the formula: Adjusted Net Profit ROCE = 100 Capital Employed Higher a company s ROCE, stronger will be its financial position. ROCE has two components profit as a percentage of sales (Profit margin) and sales as a percentage of capital employed (Investment turnover). Alternatively, it can be defined as: PAT Sales Sales 100 Average Capital Employed A firm can improve its ROCE by increasing one or both of its components viz., profit margin i.e. (PAT / Sales) and the investment turnover i.e. productivity of its capital employed (Sales / ACE). It also indicates that a company with higher operating profit margin may have a lower ROCE if its asset efficiency is poor. Thus ROCE analysis provides a strong incentive for optimal utilization of the assets of the company and is used as a measure of success of a business. For the purpose of the study, it is expected that ROCE will not only find a significant reflection in the market value addition of a company but will also be a significant predictor of the same. ii. Return on Net Worth (RONW) This ratio measures the relationship between net profits and proprietor s funds and thus, reveals how well the firm has used the resources of owners. So, this ratio is of great interest to the present as well as prospective shareholders and also of great concern to management, which has the responsibility of maximizing the owners welfare. Further, RONW is also capable to reveal the relative performance and strength of the company in attracting future investments. It is calculated by the formula: RONW = Net profit after interest and taxes Net worth 150

6 Where, Net worth = Equity Capital + Reserves and surplus + Preference share capital Accumulated losses It is believed that the stock prices and hence the market capitalization reacts favorably to an improvement in RONW (Misra and Kanwal, 2004). This expected positive relationship leads to improvement in the market value added of a company. Thus, RONW is selected as one of the independent variable having positive relationship with MVA of a company. This variable is also a relative measure and has been expressed in percentage terms. iii. Profit after Taxes (PAT) It is the net profit earned by the company after deducting all expenses like interest, depreciation and tax. It is defined as: PAT = EBDIT Depreciation Interest Taxes Where, EBDIT is earnings before depreciation, interest and taxes. PAT is expressed in absolute (Rupee) terms. It has been selected as an independent variable as normally it is expected to have a positive correlation with the MVA i.e. increasing Profitability in a well functioning capital market is likely to give a boost to the share prices, market capitalization and market value added. A company, whose profitability is not sufficient to cover up its overall cost of capital, face adverse EVA situation, the result of which is the decline in its stock prices and therefore, its Market Value also falls. For the purpose of the study, PAT figures are taken from the corporate database, Prowess. iv. Earning per share (EPS) EPS is an absolute measure of profitability that identifies how much each share has earned for the shareholders. Investors, in general, look upon earnings per share as the best yardstick to analyze their investment decisions. It is calculated by the formula: Net profit after tax Preference dividend EPS = Total number of Outstanding Equity shares Traditionally it is believed that EPS has a positive relationship with share prices and hence MVA (Misra and Kanwal, 2004). It is also considered as one of the major factors 151

7 affecting the dividend policy of the firm and market prices of the company (Kishore, 2002). Thus, the study expects a positive association between EPS and MVA of a company and EPS has been taken as an independent variable affecting market value addition of a company. v. Return on Total Assets (ROTA) The ROTA measures a company's profits before interest and taxes (PBIT) against its total assets. It is considered as an indicator of how effectively a company is using its assets to generate earnings before the contractual obligations must be paid. This ratio takes out the impact of interest and tax to depict clearly how well the operational managers have done with the assets of the company. ROTA is calculated as: Profit before Interest and taxes ROTA = 100 Total Assets Greater a company's earnings in proportion to its assets, more efficient that company is considered in using its assets and contributing towards firm value and shareholders wealth. vi. Economic Value Added (EVA) EVA is conceptually a superior measure of performance because it charges management for using capital at an appropriate risk-adjusted rate, and it eliminates financial and accounting distortions to the extent it is practical to do so (Stewart, 1994). Economic Value Added is calculated as the difference between NOPAT and the stakeholders expectations, which is the capital charge for both debt and equity i.e. overall cost of capital. Operationally defined, EVA = NOPAT Capital charge = NOPAT WACC Economic Capital Where, NOPAT is Net Operating Profits after adjusting for non-operating items, nonrecurring events and other economic adjustments to compute economic profits from accounting profits. The detailed explanation of these adjustments is given in Chapter 4. NOPAT = (PAT + non-recurring expenses + revenue expenditure on R & D + interest expense + goodwill written off + provision for taxes) - non-recurring income - R & D amortization cash operating taxes. 152

8 WACC = Weighted average cost of capital. = Cost of equity proportion of equity in total capital + Cost of debt proportion of debt in total capital (1 tax rate) + Cost of preference capital proportion of preference capital in total capital. Economic Capital = Net Fixed Assets + Investments + Current Assets (NIBCLs + Miscellaneous Expenditure not written-off + Intangible Assets) + (Cumulative Non-Recurring Losses + Capitalized expenditure on R & D + Gross Goodwill) Revaluation Reserve Cumulative Non-Recurring Gains Grant (2003) emphasized that EVA is the internal performance measure that is most highly correlated with MVA and provides the most reliable guide to- whether and by how much, management actions have contributed to shareholder wealth. Grant (2003) also stated that companies having positive EVA momentum should on balance see their stock prices go up over times as the increasing profits, net of the overall capital costs lead to a rise in the company s Market Value Added. Moreover, the relationship between MVA and EVA has also been supported empirically by a number of prominent researchers (section 2.1 in chapter 2). The purpose of selecting EVA as an important explanatory variable is to identify in Indian context that to what extent EVA is associated with MVA and whether EVA dominates other traditional performance measures in explaining the changes in MVA of sampled companies. vii. EVA as a percentage of capital employed (EVACE) It indicates that how much value has been added by the company at given level of capital employed and is determined by the formula: EVA EVA as a % of capital employed = 100 Invested capital The logic behind considering this relative measure of value creation as independent variable is to identify that what explains MVA more; EVA in absolute terms or EVA as a relative measure. This ratio can assist the policy makers to infer whether the market s response to the relative measures of financial performance is better than that to the absolute measures of financial performance (Misra and Kanwal, 2004). 153

9 viii. Capital Productivity (Cp) Capital productivity captures revenue generated per unit of capital employed and indicates the productivity taken out of the fixed assets of a company. This ratio improves when a company manages to generate more revenue out of the same assets i.e. through better utilization of its capital resources. It is calculated as: (Net Sales + Change in Stock Raw Material Consumed Power and Fuel Cost) Cp = Net Fixed Assets Where, Net Sales is Gross Sales net of indirect taxes and Net Fixed Assets are Gross block net of accumulated depreciation. It is believed that shareholders value can improve only when capital productivity improves. If fixed assets are efficiently used, it would generate wealth for the stakeholders in a company and more particularly for equity shareholders (Misra and Kanwal, 2004). Thus, Cp is expected to be a significant predictor of market value added of a company. ix. Employees Productivity (Ep) Employees productivity is the ratio of (the real value of) output to the input of employees. This is defined as: (Net Sales + Change in Stock Raw Material Consumed Power and Fuel Cost) Ep = Salaries & Wages To calculate Ep, denominator i.e. employees input can be expressed in terms of hours worked, numbers of employees or expenses on salaries and wages. Based upon Review of literature (Banerjee and Jain, 1999; Bhatnagar et al., 2004; Misra and Kanwal, 2004 and Singh and Garg, 2004), this study uses salaries and wages expenditure as a surrogate of employees input. Thus, Ep denotes value addition per rupee of salaries and wages bill. Ep will improve if value addition improves for the same level of salaries and wages i.e. if the efficiency of workforce improves and/or if same value can be achieved with lower salaries and wages cost, which implies that the less number of employees can do the same job with equal efficiency (Misra and Kanwal, 2004). It is expected that the employees productivity significantly affect the profitability of a company and consequently establishes a significant relationship with shareholder value (MVA) of a company. 154

10 Table 5.1: Variables used in the Data Set and Other Formulas Applied Variables Name Notations Description 1. Dependent Variable Market Value Added MVA Market Capitalization-Economic Capital 2. Independent Variables Economic Value Added EVA NOPAT-( WACC *Economic Capital) Economic Value Added (%) EVA% (EVA/Economic Capital) * 100 Earning Per Share EPS (PAT-Dividend on Pref. Shares)/No. of outstanding Equity Shares Return on Capital ROCE (PAT nnrt/average Capital Employed)*100 Employed Return on Average Net RONW (PAT/Average Net Worth)*100 Worth Return on Total Assets ROTA (Profit before Interest and Taxes/Total Assets) x 100 Profit After Taxes PAT The profit earned by the company after accounting for all expenditures (operational, selling & distribution, administrative & other overheads and financial costs) is included under this datafield. Capital Productivity Cp (Net Sales + Change in Stock Raw material consumed Power and Fuel Costs) / Net Fixed Assets Employees Productivity Ep (Net Sales + Change in Stock Raw Material Consumed Power and Fuel Cost)/Salaries & Wages 3. Other Formulas Applied Market Capitalization Market 365 days Avg. Market Cap. + Average Borrowings Cap Economic Capital EC Net Fixed Assets + Investments + Current Assets (NIBCLs + Miscellaneous Expenditure not written-off + Intangible Assets) + (Cumulative Non-Recurring Losses + Capitalized expenditure on R & D + Gross Goodwill) Revaluation Reserve Cumulative Non-Recurring Gains Net Operating Profit After Taxes Weighted Average Cost of Capital NOPAT WACC (PAT + non-recurring expenses + revenue expenditure on R & D + interest expense + goodwill written off + provision for taxes) - non-recurring income - R & D amortization cash operating taxes Ke = cost of equity shareholders funds Kd = cost of debt Kp = cost of preference capital E = book value proportion of average shareholders funds D = book value proportion of average total borrowings P = book value proportion of average preference capital. 155

11 5.5 Statistical Diagnostic Initially, the study used Partial Regression Plots as the detection method to identify observations that were outliers on the dependent variables. The analysis indicated four companies namely Satyam Computer Services Ltd., Tata Steel Ltd., Sun Pharmaceuticals Industries Ltd. and Sesa Goa Ltd. to be unrepresentative involving extreme values. Thus, these four companies had to be eliminated from the further analysis. Before proceeding to the further analysis, the existence of multicollinearity among independent variables had also been taken care of. For this purpose, Pearson s correlation matrix was at first formed that signaled high correlation among various independent variables i.e. ROCE, RONW, ROTA, and EVA%, causing the problem of multicollinearity. To overcome this problem, various combinations of the independent variables were created and tested. Finally on the basis of Best Model Fit Criteria, highly collinear variables i.e. RONW, ROTA, and EVA % were omitted from the model. Further, two variables i.e. Ep and EPS were also eliminated from the model due to their negligible and insignificant correlation with the dependent variable. Table 5.2: Correlation Co-efficient Matrix with Selected Variables ( ) Variables EVA ROCE PAT CP MVA EVA *.357*.162*.520* ROCE.526* *.265*.392* PAT.357*.213* * CP.162*.265* * MVA.520*.392*.724*.083* * Correlation is significant at the 0.01 level (2-tailed). Table 5.2 provides the Correlation Matrix for MVA and the four selected independent variables for the period Correlation is an extremely useful tool to estimate the strength of the relationship between the corresponding pair of variables in a correlation matrix. The analysis of the table reveals that all the selected variables are positively and significantly associated with MVA. The highest positive relationship exists between PAT and MVA at.724. That means, as PAT increases, there would be an increase in the shareholder value. The similar positive relationship can also be observed between EVA 156

12 and MVA as well as between ROCE and MVA. As far as correlation among independent variables is concerned, the maximum correlation can be observed between EVA and ROCE at.526 which is much lesser than the prescribed rule of thumb of 0.8 (Gujarati, 1995). Hence, it can be claimed that multicollinearity does no longer exist in the selected regression model. In addition, the study also undertakes to consider Average Variance Inflating Factor (VIF) to detect multicollinearity. Durbin-Watson statistics has been employed to check the assumption of independent errors (auto-correlation). The White Procedure is applied to ensure that coefficients are not heteroscedastic. 5.6 Model Development The next methodological requirement is to specify the regression model used to compare the relative information content of the competing measures of firm performance (Value Based Measures as well as Traditional Financial Performance Measures) on the basis of their association with MVA. The following model has been selected for the purpose of Panel Data Analysis i.e. MVA it = α + β 1 PAT it + β 2 ROCE it + β 3 Cp it + β 4 EVA it + e it. Equation 5.1 The dependent variable in the above equation is the Market Value Added (MVA) for firm i in period t. The explanatory variables in the model are Profits after Taxes (PAT), Return on Capital Employed (ROCE), Capital Productivity (Cp) and Economic Value Added (EVA). Following the literature on the relative information content of various firm performance measures, the hypothesis suggests positive coefficients for PAT, ROCE, Cp and EVA when specified as explanatory variables for MVA. It also suggests that the more closely these measures approximate market value addition, the higher will be the relative information content of these measures. This model is estimated using a pooled time-series, cross-sectional least squares regression. Test for Relative and Incremental Information Content To assess relative and incremental information content, the study employs a statistical test from Biddle et al. (1997) that allows a test of the null hypothesis of no difference in the ability of two competing sets of independent variables to explain variation in the dependent variable. Using this test, the study makes four univariate regressions (between MVA and each of the four independent variables) and six pairwise comparisons of regressions among the value based and accounting performance measures namely EVA, PAT, ROCE and Cp. The test is constructed as a comparison of R 2 s (Biddle et al., 1997). 157

13 5.7 Empirical Results and Discussions To select the most appropriate pooling technique, the study estimated coefficients, standard errors and t-statistics for the model across three alternative pooling techniques i.e. pooled results, fixed effects and random effects. In the first instance, a pooled regression was run the results of which are presented in table 5.3. These results reported an adjusted R 2 of.618 and F statistic was also found to be significant (p<.001). Table 5.3: Multivariate Results of Pooled Regression Analysis (Restricted Model) Variable Coefficient Std. Error t-statistic Prob. C EVA PAT ROCE CP R-squared Adjusted R-squared F-statistic Prob (F-statistic) However, the pooled regression does not anticipate the firm or time specific effects. To consider whether firm-specific or time-specific factors had any significant effect on the dependent variable-mva, there was a need to estimate both fixed and random effect models with firm, time or both effects. Thus, at first, Fixed Effects were observed for cross sections, the results of which are presented in Table 5.4. Table 5.4: Multivariate Regression Results with Cross-Sectional Fixed Effects and No Period Effects Variable Coefficient Std. Error t-statistic Prob. C EVA PAT ROCE CP Effects Specification Cross-section fixed (dummy variables) R-squared Adjusted R-squared F-statistic Prob (F-statistic)

14 To test the joint significance of firm effects in the fixed effects model, the Redundant Fixed Effects-Likelihood Ratio was obtained. Table 5.5: Results of the Cross-Sectional Redundant Fixed Effects-Likelihood Ratio Effects Test Statistic d.f. Prob. Cross-section F Cross-section Chi-square (99,1096) Table 5.5 presents the results of the Cross-Section Redundant Fixed Effects that provided F-statistic being significant at 1% level. It indicated that the Fixed Effects Regression Model (Least Squares Dummy Variable Regression Model) was valid i.e. the fixed effects were found to be efficient among cross-sections in the sample. Similarly Fixed Effects were examined for time period, the results of which are provided in tables 5.6 and 5.7. In this case also, fixed effects were found to be efficient for the study period. Table 5.6: Multivariate Regression Results with Period Fixed Effects and No Cross- Sectional Effects Variable Coefficient Std. Error t-statistic Prob. C EVA PAT ROCE CP Effects Specification Period fixed (dummy variables) R-squared Adjusted R-squared F-statistic Prob (F-statistic) Table 5.7: Results of the Period Redundant Fixed Effects-Likelihood Ratio Effects Test Statistic d.f. Prob. Period F Period Chi-square (11,1184)

15 Further, the study attempted to choose between fixed and random effects specification. This was accomplished applying the Hausman test in each case. As explained earlier, the Hausman test is a test of H 0 : that random effects would be consistent and efficient, versus H 1 : that random effects would be inconsistent. (Here, fixed effects would certainly be consistent.). As for the fixed effects, the random effects could also be along either the cross sectional or period dimensions. Thus, at first the random effects were selected for the firms (i.e. cross-sectional) and not over time. Here, the slope coefficients were quite different compared with both pooled and fixed effects regressions. Thus, there was a need to identify whether the random effects model passed the Hausman test for the random effects being uncorrelated with the explanatory variables. The results of the Hausman test based on the firm random effects indicated the p-value for the test being less than 1%. It indicated that the random effects model was not appropriate and fixed effects specification was to be preferred in cross-sectional analysis. Table 5.8 reports the top panel of the Hausman test results for firms. Table 5.8: Results of the Cross-Section Random Effects - Hausman Test Test Summary Chi-Sq. Statistic Chi-Sq. d.f. Prob. Cross-section random Following the similar procedure, next, the random effects were selected for the period dimension. Here, as presented in table 5.9, the results of the Hausman test indicated an insignificant p-value of It indicated the acceptance of null hypothesis that random effects were efficient and consistent as far as period dimension was concerned. Table 5.9: Results of the Period Random Effects - Hausman Test Test Summary Chi-Sq. Statistic Chi-Sq. d.f. Prob. Period random Thus, finally the model selected was fixed effects for firms (cross-sectional) dimension and random effects for period dimension. 160

16 Table 5.10: Results of the Multivariate Panel Data Regression Analysis (Final Model with Cross-Sectional Fixed Effects and Period Random Effects) Variable Coefficient Std. Error t-statistic Prob. C EVA PAT ROCE CP Cross-section fixed (dummy variables) Effects Specification S.D. Period random Idiosyncratic random R-squared Adjusted R-squared F-statistic Prob(F-statistic) R-squared Weighted Statistics Unweighted Statistics Rho Table 5.10 provides that the variability in the MVA accounted for by the four final predictors comes out to be 81.97% (R 2 ). A high and positive value of Adjusted R 2 at 80.27% verifies that the cross- validity of this model is very good. F-statistic is found to be large (48.368) and significant (at 1% level). The results also show that all the selected independent variables i.e. EVA, PAT, ROCE and Cp have positive slope coefficients (i.e. β values) showing their positive association with MVA. However, tested on the basis of t-statistic, just two independent variables i.e. EVA and PAT are identified as the significant predictors of MVA at 1% confidence level (p<.001). On the other hand ROCE and Capital Productivity i.e. Cp do not seem to have established a significant statistical association with MVA. As explained earlier, multicollinearity is also not a cause of concern in the model and has properly been accounted for. Thus, the above indicators claim the regression model to be statistically fit and valid. 161

17 Table 5.11: Summary of the Results of the Association of Independent Variables with Market Value Added Models Estimated Coefficient EVA PAT ROCE Cp Standard Error I t- statistics Estimated Coefficient Standard Error II t- statistics Estimated Coefficient Standard Error III t- statistics Estimated Coefficient Standard Error IV V VI VII VIII IX X XI (.0568) (.021) (.838) t- statistics.6120 (.5406) (.132) (.666) (.478) (.522) F Adj. R Note: The dependent variable is Market Value Added (MVA) and explanatory variables represent one value based measure i.e. Economic Value Added (EVA), and three accounting based measures namely Profit after Taxes (PAT), Return on Capital Employed (ROCE) and Capital Productivity (Cp). The first four models (I to IV) present the results of the univariate association of each independent variable with the dependent variable- MVA. In the next six models (V to X), independent variables are specified in pair-wise combinations and finally considered jointly in the last model (XI). 162

18 Relative and Incremental Information Content Tests Table 5.11 presents the estimated co-efficient, standard errors, t-statistics, F-statistic and adjusted R 2 for each model. The dependent variable is Market Value Added (MVA) and explanatory variables represent one value based measure i.e. Economic Value Added (EVA), and three accounting based measures namely Profit after Taxes (PAT), Return on Capital Employed (ROCE) and Capital Productivity (Cp). The first four models (I to IV) present the results of the univariate association of each independent variable with the dependent variable- MVA. In the next six models (V to X), independent variables are specified in pair-wise combinations and finally considered jointly in the last model (XI). The detailed regression results for each of univariate and bivariate (pair-wise) models are given in table 1 through table 10 of Appendix B. Table 5.11 clearly shows that PAT is the most significant predictor of MVA when it is considered univariately as well as when paired with EVA. Similarly, ROCE is also found to be significant by itself and when compared with PAT. The pair-wise regressions that best explain the variations in MVA are EVA/PAT (80.31%), PAT/ROCE (80.07%), EVA/ROCE (68.81%), PAT/Cp (60.26%), ROCE/Cp (59.37%) and EVA/Cp (34.52%). Here, EVA comes twice among the best three pair-wise regressions which evidence EVA to be a highly significant explanatory variable. However, profit after taxes (PAT) can clearly be observed as the best predictor of MVA and is thus, recognized as the most legitimate and reliable measure of shareholder value creation. Further PAT is followed by ROCE, which depicts a slightly less explanatory power of 59.40% in comparison to 60.25% for PAT. These results show that traditional measures of performance have emerged as the more dominating determinants of MVA during the study period. Table 5.12 presents the summary results of regressions based on the Relative and Incremental Information Content tests. Panel A of the table summarizes the significant differences in the relative information content between accounting and value based measures. The results of single coefficient regressions clearly show that R 2 (PAT)> R 2 (ROCE)> R 2 (EVA)> R 2 (Cp) where R 2 depicts the percentage variation in shareholder wealth (MVA), as explained by each particular explanatory variable. 163

19 Table 5.12: Results of Relative and Incremental Information Content Tests Panel A: Results of Relative Information Content Test PAT > ROCE > EVA > Cp 60.25% 59.40% 34.51% 13.62% Panel B: Results of Incremental Information Content Test PAT/EVA PAT/ROCE EVA/ROCE PAT/Cp ROCE/Cp EVA/Cp ( ) ( ) ( ) ( ) ( ) ( ) 45.73% 20.60% 9.4% 46.64% 45.75% 20.89% EVA/PAT ROCE/PAT ROCE/EVA Cp/PAT Cp/ROCE Cp/EVA ( ) ( ) ( ) ( ) ( ) ( ) 20.02% 19.75% 34.26%.01% Note: The comparison of adjusted R 2 of the first four models in table 5.11 (where each explanatory variable is specified univariately) provides the results of relative information content test. For the incremental information content test, the adjusted R 2 of earlier univariate regressions have been subtracted from the adjusted R 2 of each pair-wise (bivariate) regressions in models V to X in table

20 The results in Panel B provide the results of incremental information content tests for the pair-wise comparisons of all the four explanatory variables. For this purpose, the adjusted R 2 of earlier univariate regressions have been subtracted from the adjusted R 2 of each pair-wise (bivariate) regressions to identify the incremental information provided by each explanatory variable in relation to other variables. For instance, in panel B, PAT/EVA (45.73%) is equal to the information content of the pairwise comparison of PAT and EVA (80.27%) minus the information content of EVA (34.54%) from table Looking at the pairwise combinations, it can be observed that over the PAT measure alone, explanatory power has increased by 46.64% and 45.73%. Similarly the explanatory power has improved by 45.75% and 34.26% respectively over the ROCE measure alone. Combining both of these measures i.e. PAT and ROCE, the incremental information content of PAT (20.60%) is slightly more than the incremental information content of ROCE (19.75%). As far as the comparison between value based and accounting based measures is concerned, the results clearly depict that explanatory power improves by 20.89%, 20.02% and 9.4% respectively over the EVA measure alone. Although it is lesser than the incremental information provided by the traditional measures PAT and ROCE, yet it provides the most logical pairing of information variables in explaining MVA i.e. Models V (that best explains MVA) and XI (all variables considered jointly). Thus individually, EVA explains as much as 34.54% of the variation in MVA and in combinations, it also evidences increment information content (although lesser than that of PAT and ROCE). Thus, the results provide the sufficient evidence that traditional measures of firm performance (both absolute and relative measures i.e. PAT and ROCE respectively) are highly associated with its shareholder value creation as measured in terms of MVA. Finally, the present study denies the hypothesis of equal relative and incremental information content and identifies that Earnings dominate EVA (the Value Based Measure) in explaining the variations in firm value and hence shareholder wealth. 5.8 Potential Factors Contributing to the Failure of EVA to Dominate Earnings The present study finds no clear evidence to support Stern & Stewart s claim that EVA is superior to the traditional performance competitors in its association with MVA. On the contrary, the evidence suggests that the Indian market seems more focused on Profits 165

21 than value based measure EVA. The study empirically finds that although EVA and PAT both depict highly positive and significant association with MVA yet PAT s explanatory power is greater than the explanatory power of EVA. Further, the results also provide the sufficient evidence that traditional measures of firm performance (both absolute and relative measures i.e. PAT and ROCE respectively) are highly associated with its shareholder value creation as measured in terms of MVA. That means Indian market is more responsive to accounting based metrics whether these are expressed in absolute terms or in relative terms. So, accrual accounting based numbers can undoubtedly be continued for evaluating corporate financial performance. As the key findings of the study evidence the Earnings superiority to EVA in relative information content test (in their association with MVA), the study identifies the potential factors contributing to the failure of EVA to dominate Earnings in explaining the variations in shareholder value creation. Kramer and Pushner (1997) explained that with the market being fed almost constant news on earnings, it is not surprising that it is not much responsive to EVA in the short-run. Another reason might be that accounting adjustments and estimates of the capital charge given by the proponents may contain measurement error relative to what the market uses for valuing firms. Biddle et al. (1997) observes that in attempting to estimate economic profits, adjustments made by Stern & Stewart may remove accruals that market participants use to infer firm s future prospects. Thus, while computing EVA, the true measure of company s economic profitability is determined but its association with market returns is lost. Moreover, another reason for the comparatively weak value- relevance of EVA might be the prevalent notion of earnings myopia. Biddle et al. (1997) viewed that some adopters of EVA feel that they must still base their external performance on earnings because this is the measure on which financial analysts continue to focus. As a result, market fails to recognize the reporting benefits of EVA. However, the present study does not question the effectiveness of EVA because inspite of non-availability of detailed financial data required for EVA related computations and non-mandatory disclosure of EVA Statements in annual reports of Indian Companies, market seems to be quite responsive to EVA performance of a company. Thus, the findings advocate adoption of EVA for management compensation, external communication and security analysis and also 166

22 suggest disclosure of EVA in financial reporting, to align management objectives with shareholders interests and facilitate value-based performance monitoring (Holler, 2008). 5.9 Conclusion Analyzing a pooled, time series, cross-sectional data of 100 Indian companies for a period of twelve years i.e. from 1996 to 2007, this study has attempted to examine whether the value based measures of firms performance are more highly associated with firm s MVA than other long established traditional measures. The results indicate that the variability in the MVA accounted for by the four final predictors comes out to be 80.27% (adjusted R 2 ). However the study found no clear evidence to support Stern & Stewart s claim that EVA is superior to the traditional performance competitors in its association with MVA. The empirical evidence suggests that Indian market seems to be more focused on Profits than value based measure EVA. Relative tests show the dominance of PAT and ROCE over EVA; and incremental tests find that solely accounting based measures provide considerable and significant additional information, whereas EVA provides comparatively lower incremental information. Thus, Indian market being less responsive to EVA than PAT needs more ongoing investigation. 167