Shareholder Activism Reached Unprecedented Levels Accounting Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

In recent years, shareholder activism reached unprecedented levels and led to increased pressure on firms to maximize shareholder value (Bacidore, Boquist, Milbourn, & Thakor, 1997). However, despite their best efforts, many companies failed to create shareholder wealth. Consulting companies such as Stern Stewart & Company (EVA), Boston Consulting Group's HOLT Value Associates (cash- flow return on investment: CFROI), KPMG Peat Marwick (economic value management EVM), and Marakon Associates (discounted economic profits: EP) have developed value-based performance measures in order to overcome the shortcomings of traditional accounting performance measures (Biddle, Bowen, & Wallace, 1997). The most popular value-based performance measure is Stern Stewart's Economic Value Added (EVA).

Corporate managers and the business press have shown great interest in the use of EVA as a measure of performance. This new metric is a profit measure based on the true economic income. The principal feature of EVA measure is that, unlike traditional accounting measures, it reduces income by a charge for the cost of capital that includes the cost of the equity capital provided by owners. This charge has long been included in certain traditional measures of income that mainstream economists have used for more than a century (McIntyre, 1999). A company's value is created only when the return on the invested capital is higher than its cost of capital. The opposite is also true. When the return on the invested capital is less than the cost of capital, a firms' value is undermined. EVA has gained widespread acceptance and credibility not only as an operational performance metric, but also as a way in which management's decisions contribute value to an organization. It has even been predicted by Zarowin (1995) that EVA may replace Earnings per Share (EPS) in The Wall Street Journal's regular stock and earnings reports. Specifically, EVA has been used as a management compensation tool and for management decision-making in capital budgeting. Financial investment firms are relying upon EVA to determine the best companies in which to invest. Stewart (1991) argued that such traditional measures as earnings, earnings per share (EPS), return on equity (ROE), and return on investment (ROI) are misleading measures of corporate performance. EVA, in contrast, is what drives stock prices and is conspicuous as the single best measure of wealth creation (Stewart, 1994).

Despite wide interest in EVA, there is a dearth of empirical evidence on the efficacy of this measure versus other measures of firm performance in the banking industry. The relevance of this alternative performance measurement to traditional accounting measurement has not been fully explored. Academic research has not established EVA's correlation to market value and stock returns. The evidence from previous studies is mixed and has not resolved the debate over performance measures.

The objective of this study is to investigate whether EVA is a better predictor than currently mandated performance measures, earnings per share (EPS), return on equity (ROE), return on investment (ROI) profit after tax (PAT), and return on net worth (RONW) in explaining the market value of banking firms in India. EVA has been compared to traditional accounting performance measures. The remainder of this article is organized as follows, calculation of EVA for banks, summary of empirical studies, the adoption of EVA technique in the banking industry, the hypotheses, and the methodology, followed by empirical results, conclusions, and limitations and suggestions for future research.

Economic Value Added (EVA) for Banks: EVA expresses the surplus value created by a company in a given period, i.e. the firm's profit net of the cost of all capital. This measure is computed as the product of the difference between the return on investment and its composite financing cost (i.e. excess return) and the capital invested.

EVA = Capital Invested*(Return on Capital Invested- Cost of Capital)

= (Capital Invested*Return on Capital Invested) - (Capital Invested* Cost

of Capital)

= NOPAT - (Capital Invested*Cost of Capital)

As noted in Velez-Pareja (2000), when EVA is used to assess company performance in a given period, capital invested and NOPAT should not be calculated for the same period. As investors expect to receive returns on the investment made in the beginning (and not on the cumulative amount at the end of the period), shareholders compare returns (i.e. NOPAT) earned over the period with the capital invested in the beginning (and not at the end) of the period. For this reason, capital invested is measured with a lag of one year and EVA is calculated as follow:

EVAt =NOPATt - (Capitalt-1 * Cost of Capital)


EVAt = EVA of period t

NOPATt = NOPAT of period t

Capital Investedt-1 = Capital Invested measure at the period of t-1

NOPAT and capital invested cannot be calculated on an accounting basis, but need to be calculated on an economic basis. Advocates of EVA have identified more than 160 accounting adjustments, but it is unrealistic even to think of making all these adjustments for any single company. In the empirical investigation, we calculate a "disclosed EVA", which EVA is obtained after making some standard adjustments to publicly available accounting data. The calculation of EVA requires, in fact, to express NOPAT and capital invested on an economic basis, for this reason, advocates of EVA suggest some adjustments in order to:

Avoid mixing of operating and financing decisions;

Provide a long term perspective;

Avoid mixing of flow and stock;

Convert GAAP accrual items to a cash-flow basis or, in other cases, convert GAAP cash-flow items to additions to capital.

Adjustments: In calculating EVA, seven adjustments have been carried out concerning the following items:

Loan loss provision and Loan loss reserve: Loan loss reserve is a reserve aiming to cover any future loan losses, for this reason, it should be equal to the net present value of all future loan losses. In any single period, this reserve is reduced by net charge-off (i.e. the current period losses due to credit risk) and replenished by loan loss provisions (i.e. the provision made in the current period to adjust the reserve both for pre-existing loans and for estimated future loan losses related to newly originated loans).

This convention is certainly commendable from a management perspective since it implies that all loan losses are pre-funded out of current earnings. However, loan losses provisions are commonly used to manage earnings: if a bank achieves high operating returns, bank managers tend to overestimate this provision, while they are inclined to underestimate it if operating earnings are poor. This accounting practice introduces an important distortion in analyzing bank performance since it smoothes earnings. Business is risky, and the volatility of profits is a manifestation of this risk, for purposes of economic performance evaluation, smoothening earnings is inappropriate.

Taxes: Most banks show significant and persistent differences between book tax provisions and cash tax payments. Since these differences are quasi-permanent, deferred taxes should be considered as capital and, similarly to loan loss provisions, taxes need to be considered as current period expenses for purposes of economic performance evaluation.

Restructuring charges: Over the last decade, many banks have carried out restructuring plans in order to improve their operating efficiency. To the extent that such restructuring charges represent disinvestments, these costs should be treated as a capital reduction rather than costs (and therefore reduce NOPAT). Since availability of data do not allow us to evaluate the extent of real disinvestments due to restructuring charges, these costs are omitted when adjusting NOPAT and capital invested.

Security accounting: In many countries (such us U.S., Italy, France and U.K), "available for sale securities" (AFSS) are marked to market through the capital accounts. From an economic perspective, however, one might claim that selling a security, with a coupon below (or above) the current market yield and using the proceeds to replace it with a current market yield security is a zero sum game. In evaluating the economic performance of a bank, it is therefore more accurate to remove from NOPAT the effect due to gains and losses on sale of AFSS: these gains and losses should be amortized against NOPAT over the remaining lives of the securities. However, since data on the remaining lives of the securities are not available and a reasonable assumption cannot be made, these costs are omitted adjusting NOPAT. Capital gains and losses generated marking to market AFSS (rather than past capital gains and losses amortized in the period t) are therefore considered as a part of NOPAT.

General risk reserve: This adjustment aims to correct the distortions derived by the "general risk reserve", a standard feature for Italian banks. This provision is a reserve aiming to cover a bank's future generic loan loss in any single period; this reserve is reduced by net charge-off (i.e. the current period losses) and replenished by general risk provisions (i.e. the provision made in the current period to adjust the reserve according to the bank's risks). Similarly to the loan loss reserve, this convention is certainly commendable from a management prospective, but it is used in an opportunistic manner. This accounting practice introduces an important distortion in analyzing banks' performance since it smoothes earnings.

Research and Development (R&D) costs and training costs: i.e. expenses designed to generate future growth. Current assets do not benefit from these expenses and it would be incorrect to reduce operating income by the amount of these expenses. However, GAAP requires companies to treat all outlays for R&D as operating expenses in the income statement. As a consequence, it is appropriate to correct this accounting distortion by considering operating income without these expenses.

Operating lease expenses: These are disguised financial expenses.

Before going on with the EVA calculation, it is necessary to precisely define how capital invested and cost of capital should be measured for commercial banks. Many studies (e.g. Velez-Pareja 2000) measure book value of capital using total assets and, therefore, measure the cost of invested capital as Weighted Average Cost of Capital (WACC). While this solution is certainly accurate for non- banking companies, this procedure would be misleading for commercial banks. Since financial intermediation is the core business for banks, debts cannot be simply considered as a financing source (as for other companies), since they really are productive inputs (as the workforce, IT assets, etc). This view is also confirmed when analyzing NOPAT's significance. In non-banking companies, interest costs are not considered in NOPAT because these are not operating costs, but financial expenses. However, a bank's operating costs mainly derive from interest expenses because financial

Intermediation is the bank's core business. As a consequence, if the capital charge is calculated applying WACC on total assets (as usually done for non-banking companies), EVA will be biased since it will be obtained counting twice the charge on debt:

Subtracting from NOPAT a capital charge on the overall capital (equity and debt) invested in the bank;

Calculating NOPAT, when interest expenses (i.e. the charge on debt capital) are subtracted from operating revenues.

For these reasons, our advice is to focus on equity capital and measure the capital invested in the bank as the book value of shareholder equity. Regarding the cost of capital, the capital charge cannot be obtained applying the bank's WACC on the capital invested because the latter is given by the equity capital and not by the overall capital (debt and equity). Consequently, a commercial bank's cost of capital invested should be measured by the cost of equity. To support this view, Sironi (1999) identifies four differences (labelled as "the separation principle", "banks as providers of liquidity services", "capital ratios", "off-balance sheet pro") between a bank's cost of capital and that of a non financial company, and observes "with a capital structure exogenously determined by regulators, a marginal cost of debt close to that obtainable from the inter-bank market, and relatively similar to that of all other major banks, and an array of products that do not need any debt financing, banks should look at their cost of equity capital as a key variable". The cost of equity is estimated using the Capital Asset Pricing Model (CAPM) looking at investors' expected return.

A quick glance at results in empirical studies

Most of the studies dealing with value base management (VBM) have investigated the information contents of the innovative performance measures (especially EVA, the most popular) over the traditional measures (i.e. ROE, ROA, Net Income, etc). In other words, there is a growing number of studies investigating which performance measure is the most suitable to maximize value of concern.

The evidence surrounding this issue is mixed and these studies can be divided in two groups: those carried out by EVA promoters and those carried out by academics. As stated in Lehen and Makhija (1997, p. 90), "EVA is seen by its proponents as providing the most reliable year-to-year indicator of a market-based performance measure known as Market Value Added … Despite wide interest in EVA, little is known empirically about the efficacy of this measure versus other measures of performance… The evidence from these studies is mixed, however, and has not be resolved the debate over performance measures".

Regarding the practitioner literature, these studies usually observed the EVA superiority since EVA is found to better explain stock returns and firm values. As noted in Garvey and Milbourn (2000, p.211), "Stern Stewart, Boston Consulting Group, and LEK/Alcar make the claim that their proprietary performance measure correlates more closely with stock returns than do either traditional accounting measures or the measures of rival firms, allegedly making it a more desirable compensation tool".

O'Byrne (1996): analyses industrial companies and found the EVA superiority in a two-step analysis. In the first, the firm market value was regressed on EVA and then on earnings (namely, NOPAT): O'Byrne (1996) found an adjusted R2 of 0.31 and 0.33 for EVA and NOPAT respectively. In the second step of the analysis, a set of adjustments were proposed: firstly, EVA separate coefficients were allowed for positive and negative value of EVA; secondly, the natural log of capital was introduced as predictor in order to take into account differences in the way the market value firm of different sizes; thirdly, 57 dummies variables were introduced to consider potential industry effects. In this second stage, he found an R2 of 0.56 for EVA, which enable him to conclude that EVA is superior to earnings in explaining firm value.

Al Ehrbar (1998): reports that several empirical analyses have been carried out by Stern Stewart using the Performance 1000 database. According to the Stewart findings, EVA explains half of the volatility in companies' MVA, the highest correlation found.

Uyemura et al., (1996): a particularly interesting study for our purposes since its focus is on banking analyzed the largest 100 U.S. bank holding companies over a period of ten years (1986-95). By regressing changes in standardized MVA against changes in standardized EVA (defined as EVA divided by capital) and traditional performance measures, EVA was found to have the highest correlation with MVA.

Table -1: Uyemura et. el. (1996) results





Standard Error






















Net Income














Source: Uyemura et. el. (1996,p.99 ) results

Focusing on studies proposed by academics, the superiority of EVA is usually not verified. In detail:

Peterson and Peterson (1996): analyzed traditional and value-added measures of performance and compared them with stock returns. According to their findings, traditional measures are not empirically less related to stock returns than return on capital: as result, traditional measures should be not eliminated as a means of evaluating performance, though these have no theoretical appeal. From this point of view, Peterson and Peterson (1996) rule out the possibility of value added measures not being worthwhile: since value added measures focus on economic rather than accounting profit, these play an important role in evaluating performance because managers will aim towards value creation rather than mere manipulation of short-sighted accounting figures.

Biddle et al., (1997 and 1999): analyzed a sample of 6174 firm-years over the period 1984-93 by comparing adjusted R2 obtained regressing stock market adjusted returns against EVA, Residual Income (RI), accounting earnings (namely, Earning Before Extraordinary Item - EBEI) and Operating Cash Flow (CFO). According to their results, EBEI has the highest adjusted R2 and EVA has a smaller adjusted R2: these results do not support the hypothesis that EVA dominates traditional performance measure in its association with stock market returns. In addition, Biddle et al., (1997 and 1999) also assessed the relationship between performance measures and firm value by replicating O'Byrne's (1996) study with some adjustments. In order to level the playing field, Biddle et al., (1999) extended the adjustment proposed in the second stage of O'Byrne's(1996) analysis to the regressions run against NOPAT: in this case, the EVA superiority disappears. In fact, according to their results, accounting earnings have the highest adjusted R2 (0.53), EVA has an adjusted R2 of 0.50 and NOPAT has an adjusted R2 of 0.49. These results suggest that EVA does not dominate accounting earnings in explaining firm values.

Lehen and Makhija (1997): assess which performance measure does the best job of predicting the turnover of Chief Executive Officer (CEO). Focusing on the degree of correlation between different performance measures and stock market returns, they found that correlation coefficients vary between 0.39 and 0.76. In detail, EVA and MVA are the most highly correlated measure with stock market returns: 0.59 and 0.58 (respectively). The other performance measures have smaller correlations: 0.455 for ROA, 0.455 ROE and ROS 0.388. It is interesting to note that, similarly from all other studies where MVA was used as response variable, the measure mostly correlated with MVA is EVA.

Garvey and Milbourn (2000): assessed the "declared" EVA superiority by focussing on the suitability of EVA and earning to the management compensation system. This paper adds a different strain to the academic literature, since Garvey and Milbourn (2000) initially criticise the investigation techniques used previously (i.e. the statistical correlation with stock returns and/or firm value). They suggest that a strong statistical correlation with stock returns does not establish (a priori) that a performance measure adds value to a compensation system. In order to define the criteria for judging the value alternative performance measures, they proposed:

A theoretical analysis developing a standard agency model with a principal and one agent: Garvey and Milbourn (2000) concluded that it is irrelevant to investigate whether EVA beat earning per se, while it would be more accurate investigate under what circumstances EVA beat earnings (and for what reasons);

An empirical investigation by testing the model in Paul (1992) to verify the theoretical model. In detail, Garvey and Milbourn (2000) analysed the marginal value of EVA adoption for each company (using the estimated correlations among earning, EVA and prices) and associate this to the firm's EVA adoption (using a multivariate regression approach). Garvey and Milbourn (2000, p. 241) found that the "accounting measures continue to explain changes in compensation even when stock returns are used as explanatory variable. This is consistent with the Paul (1992) model in that firms do not use exactly the same weights as the stock market in determining compensation … More surprisingly, we show that the apparently simplistic idea of comparing the relative ability of alternative measures to explain stock returns is both theoretically defensible and a reasonable representation of practice. Therefore, firms contemplating the adoption of EVA would be well advised to begin with an examination of EVA's R2 with its stock returns".

Acheampong Y.J., Wetzstein M.E. (2001) propose an innovative type of analysis using parametric methods for estimating efficiency, focussing on the food industry. They conclude that: "the analysis showed that there are no significant differences between traditional and value added measures of performance".

On the basis of the studies above summarised, it appears that:

All studies carried out by practitioners found that EVA dominates traditional measures in explaining stock returns and firm values;

Studies carried out by academics found that traditional measures are not empirically less related to stock returns than EVA and other value added measures;

Garvey and Milbourn (2000) proposed a theoretical and empirical approach substantially different from all other studies. Although this contribution is very interesting, it appears to have a different focus because it compares EVA and earning as a basis for compensation systems rather than performance measures. In addition, it concludes that companies contemplating the adoption of EVA should (first of all) assess the EVA's R2 with its stock returns;

Although these studies adopted quite similar investigation techniques, the variables adopted (as predictors, but especially as response variables) are heterogeneous. Some studies [such as O'Byrne (1996), Peterson and Peterson (1996) and Biddle et al., (1997 and 1999)] attempt to evaluate different performance measures, including accounting earnings and residual income measures such as EVA, by examining their degree of correlation with stock returns on the ground that the best measure is the most highly correlated with stock returns. Some other studies [Al Ehrbar (1998) and Uyemura et al, (1996)] and compared financial measures looking at the degree of correlation with the MVA, considered by EVA promoters the "ultimate measure of shareholder wealth creation".



The primary purpose of this study is to provide empirical evidence on the relative and incremental information content of EVA and traditional performance measures. Relative information content comparisons analyze if one measure provides greater information content than another. Incremental information content comparisons assess whether one measure provides more information content than another. The first hypothesis tests the assertion that EVA dominates traditional performance measures in explaining firm values. Stewart (1994) argues that EVA is better than its traditional performance measures in explaining changes in shareholder wealth.

Hypothesis 1: The relative information content of EVA is superior to mandated performance measures, earnings per share (EPS), return on equity (ROE), return on investment (ROI) profit after tax (PAT), and return on net worth (RONW) in explaining values of banking companies.

The relative information content investigates which variables (EVA, ROA, ROI, EPS, and RONW) have greater association with firm value. The conventional way of assessing the relative information content is to compare the coefficients of determination (R2) of various simple regressions that analyze the relation between various performance measures and equity market value.

The second hypothesis refers to the incremental information content.

Hypothesis 2: EVA adds information content to that provided by EVA, ROA, ROI, EPS, and RONW. To assess incremental information content one analyzes the increase in the R2 that arises from the inclusion of more variables in the model.

Data Collection and Sample

This study used secondary data, which were collected from CMIE's (Centre for Monitoring Indian Economy) PROWESS database. The data required for this research included accounting information to derive EVA as well as traditional variables. These traditional accounting measures were selected based on the findings of previous studies that tested the relationship between EVA and accounting performance measures.

The sample for this study consisted of 38 publicly traded banks (public and private sector). Banks that did not have data for the entire period of 1996 to 2006 were eliminated from the analysis. The total number of banks with the Centre for Monitoring Indian Economy are 134 was reduced to 38, which are listed on National Stock Exchange, Bombay, whereas the number of banks with the data of entire period (1996-2006) was reduced to 20.

Variables of Study

To transform accounting income into economic income (NOPAT), some of the adjustments were adapted from Johnson (2001), Calaberese (1999) and Stewart (1991). Except for estimated income taxes, other recommended adjustments such as the LIFO reserves, the bad debt reserves, and capitalized R&D were immaterial or unavailable. For the purpose of this study, only estimated income taxes, which are measured as income taxes payable in PROWESS database, were adjusted from operating profit to derive NOPAT. To derive invested capital, interest-bearing debt and deferred income taxes are added to shareholder's equity. However, PROWESS database does not provide the information of other accumulated goodwill amortization.

To calculate a bank's weighted-average cost of capital, the following variables were operationalied. The 364 days Indian Treasury bill's rate obtained from Reserve Bank of India database was used as a proxy of a risk-free rate. There is a continuing dispute over what that premium should be and whether it should be derived from the arithmetic or geometric mean of this spread over time. Most service companies use a risk premium {βi*[E(rm) -rf]} of approximately 7% (Johnson, 2001). Following Stewart's (1991) methodology, 6% of the market-risk premium is used for the calculation of the cost of equity.

To avoid the ambiguity for the market risk premium we used calculated beta (b) as:

βi = COVim/ 2m


βi = The beta of security in question.

COVim = Stands for covariance between the return of security and return of

market portfolio.

 2m = Stands for variance of market portfolio.

Than substituting the values in the equation suggested in Capital Asset Pricing Model (CAPM)

Eri = rf+ βi [E (rm)-rf]


E (rm) = Expected return on market portfolio.

Eri = Expected or minimum required rate of return on security i.

rf = Risk free rate of return (364 days Treasury bill rates).

βi = Systematic risk of security i, beta.

To test the accuracy of EVA calculation method described above, EVA data from the Stern Stewart (BT-500) with all necessary adjustments are compared with EVA calculated in Business Today's study. The Pearson correlation coefficients between these two sets of data for the entire period of 1995 to 2001 showed high correlation with average correlation coefficients of .76 at the significant at one per cant. Therefore, we can safely conclude that the estimated EVA in this study has significant, positive association with Stern Stewart's estimated EVA.

The EVA Regression Models

To test the predictive power of EVA relative to return on assets (ROA), return on investment (ROI), earning per share (EPS), and return on net worth (RONW), five separate simple linear regression models were developed to examine the relationship between a company's market value and traditional performance measures. The EVA model would express a banking company's market value as a linear function of five independent variables (EVA, ROA, ROI, EPS, and RONW).

To test Hypothesis 1 regarding the relative information content of EVC, RAC, RIC, EPC, and RWC that were all standardized by capital, the following five linear regression models were estimated.

MVc = a0 + a1 EVC +e

MVc = a0 + a1 RAC +e

MVc = a0 + a1 RIC + e

MVc = a0 + a1 EPC + e

MVc = a0 + a1 RWC +e


a0 = Constant.

MVc = Market value of equity/capital.

EVC = Economic Value Added/ capital.

RAC = Return on assets/ capital.

RIC = Return on investment / capital.

EPC = Earning per share/ capital.

RWC = Return on net worth/ capital.

e = the error term of regression.

To test Hypothesis 2 regarding the information content of EVc, RAc, RIc, EPc and RWc the following multiple linear regression models is presented as follows:

MVc = a0 +a1 EVc +a2 RAc + a3 RIc + a4 EPc + a5 RWc + e