Economic Profit Approach in Firm’s Performance Measurement: Evidence from Turkish Stock Market Oral ERDOGAN, Niyazi BERK, Erol KATIRCIOGLU* September 8, 1999 Presented in METU International Conference on Economics III “Submitted to Russian and East European Finance and Trade” (Suggested Reference: Russian and East European Finance and Trade Vol. 36, No. 5 (Sep. - Oct., 2000), pp. 54-74)

Abstract This paper focuses on the economic profit measurement which modifies Stewart’s (1990) economic value added (EVA) approach. Considering the cost of capital in a different perspective, we recommend an economic profit model and compare it with the EVA in Turkish stock market. Economic profit (EP) is the difference between the return on equity capital and the opportunity cost of investing. Similar concepts used in finance literature are residual income, economic value added, and shareholder value added (SVA).1 Despite the abundance of discussion and the extensive use of these techniques by the corporations, not enough empirical research has been conducted in this area. In the empirical part of this study, we calculate both the EVAs and EPs on 123 firms trading in the Turkish stock market. We subscribe to the market value added (MVA) approach and demonstrate that the model offered is more suitable than the EVA approach especially in an emerging market such as Turkey.

(JEL Classification: G1, G3, L2)

1 Introduction The concept of “economic profit” is included in the wealth maximization and the theory of the firm, and to a larger extent within the theory of finance.

According to Fisher (1930), consumers optimize their savings and consumption preferences in financial markets through their relationships with firms which provide services and income. As stated by Fama and Miller (1972), finance theory considers allocation of resources by companies to various investment areas in different periods. Thus, in the efficient transfer of funds and savings, the existence of capital markets become important. According to neoclassical theory, the ownership of the companies by the households generates services and income to these households which in turn increase their wealth positions. Such increase in wealth brings up the issue of shareholder wealth maximization. However, rather than the normative approach of neoclassical economics, a method to measure the investment performances of companies needed to be developed in line with the positive economics of 1 Electronic copy available at: http://ssrn.com/abstract=880020

Friedman (1958).

In the literature, the free cash flow (FCF) and corporate valuation of Modigliani and Miller (1958, 1961) constituted main references to economic profit. Direct relationship between the company decisions and investments has been stressed by Cleary (1999). This view supports the results of the study on the investment and cash flow sensitivity of Kaplan and Zingales (1997). Once the theoretical base has been considered, practical solutions measuring the wealth level of the family members (hence the shareholders) have become one of the financial research fields from 1970s to present. Rappaport (1979, 1998), Stewart (1990), and Grant (1996) have contributed by relating the free cash flow approach to the concept of value added. They introduced the concept of value added by using “economic profit” and “cost of capital” approaches.

Firms create or destroy wealth depending on their economic behavior and their membership in a particular industry.

A firm aims to maximize its profit level, while its shareholders

maximize their capital gains and dividends from their investment. The creditors, on the other hand, expect to receive their capital back including interest. For that reason, the valuation of companies is associated with the maximization of the shareholders wealth. However, in inefficient capital markets, inefficient market values may exist due to the manipulation, insider trading, asymmetric information, and/or insufficient liquidity and market capitalization. If the demanded funds can not be transferred in time, incorrect valuation may occur. Besides, excess fund transfer to one firm could lower the market values of other firms. In emerging markets, the ratio of floating shares to outstanding shares becomes a limiting factor in valuation due to small market size. Due to limited demand for more shares, the market value of a firm with the lower ratio could be higher than that of the firm with the higher ratio.

2 Electronic copy available at: http://ssrn.com/abstract=880020

The investors usually take into account the return on equity, market capitalization, market-tobook value ratio, earnings per share, and price-to-earnings ratio to invest in the shares. However, the most convenient answer to the question of “has the capital of creditor or the shareholder been increased or decreased by the company?” might be given by the market value added (MVA) criterion (Fisher 1995). Stewart (1990) defines the MVA as the difference between the total market value of firm and the book value of total assets employed in the business. MVA greater than zero means how much value company has added to its shareholders' investment. MVA less than zero means how much value company has subtracted from its shareholders' investment.

Compared to MVA, Stewart’s (1990) EVA is the company’s after-tax net operating profit in a given year minus the cost of capital.2 Gapenski (1996) states that both MVA and EVA are applicable to investor-owned organizations; however, EVA also is an appropriate measure for non-profit organizations. This view is acceptable since the managerial effectiveness in a given year can be examined by EVA.

The consideration of cash flows from investments, the investment periods and the term differential in the calculation of the impact of inflation is specified by Levonian (1994), DeVilliers (1997), Rappaport (1998), and Fama and French (1999). Pastor and Stambaugh (1999) imply that uncertainty about factor premiums is generally the largest source of overall uncertainty about a firm’s cost of equity, even though uncertainty about betas is almost as important. Thus, especially in those markets with a high risk-free rate, a weighted average cost of capital (WACC) may be used in economic profit measurement similar to ModiglianiMiller’s (1958) and Gordon’s (1959) model.

3

By supporting the use of EVA from several points, Graham (1996) clarifies that especially consultants use economic value added (EVA) for strategic planning work, financial advice and compensation studies. We concur with him that EVA may be questionable for companies with few assets, those where major assets are leased, and growth companies.

Kwon, Shin and Bacon (1997) demonstrate that the investors’ perception of stock returns in the Korean market, an emerging market, are quite different from those of US and Japanese investors suggesting that the dividend yield, foreign exchange rate, oil price and money supply are strongly correlated with stock returns. Erb and Harvey (1995) explain the different risk structure and portfolio composition between the emerging and developed markets. They conclude that dividend yield is more useful in emerging markets in measuring investment returns. Moreover, Schnabel (1985) needs to extend the CAPM by including a dividend in the new security market line, as a means of meeting the cash demands and reducing the financial distress costs. Mookerjee (1992) shows that dividend payments are rational when the unique institutional environment in India is taken into account as an emerging market, which is contrary to recent evidence from the US and Japan. Finally, Fama and French (1999) confirms that a by-product of calculating the real cost of capital and the real return on cost is information about the history of corporate earnings, investment and financing decisions. They accept the internal rate of return on value is a starting point for cost of capital estimates for individual firms.

Within such a framework, in the environments where the risk-free rate of return is high, the Gordon model of weighted average cost of capital (WACC) is included in EP formula.

4

Section 2 of the paper presents an extended model of economic profit. In Section 3, data and the methodology of data analysis used are presented. Section 4 covers the empirical results of the study followed by the conclusion section.

2 The Model Simple introduction and definition of the preferences of the investors in capital markets with the indifference curves provides general equilibrium in stock markets. Thus, in stock markets consumption preferences of the shareholders and potential investors as well as the performance of companies will be determined. Having advanced finance theory to a further stage, CAPM represents the most common positive proposal for the perfect capital markets. It reaches to the theory of capital markets, to the theory of “fair play” or” martingales” and thus to the theory of speculation of Bachelier (1900). With such a model, the asset pricing is available on the basis of the expected return from the capital markets:

E( Pi , t + 1 Φt ) = (1 + E( Ri , t + 1 Φt ) pit ,

[1]

where E refers to the expected value, P refers to the price of the security, R refers to the rate of return for one period, Φ refers to the information set with regard to the security for a specified period, i refers to the security, and t refers to time period. On the basis of this statement, the investor’s utility maximization in terms of the capital markets theory and on the one hand and the company profit maximization in terms of the theory of firm on the other may well intersect.

This study is heavily dependent on the views by Stewart’s (1990) EVA and Rappaport’s (1998) SVA. To adjust for inflation, suggestions by de-Villiers (1997), Rappaport (1998), and

5

Fama and French (1999), and to estimate cost of capital, classical Sharp (1964, 1991)3 and Gordon (1959) models have been used. We will first present the economic profit as used in EVA. The financial expenses are also included in the net profit account indicating the company profit with “the principal of considering the capital cost” in the economic profit equation.

EP = NOPAT - (C x COC) = (ROI - WACC) x C ,

[2] [3]

where; NOPAT = Net Operating Profit + Financial Expenses C : Total Capital Employed C0C4: Cost of Capital

ROI: Return on Investment WACC: Weighted Average Cost of Capital. Although the company has earned financial expenses, they still represent the expenses as the cost of borrowing. Once the cost of capital concept is considered while leaving the accounting technique aside, the NOPAT has been identified as a new profit measure. Thus, NOPAT is determined by adding the interest expenses to the figure of net profit after the deduction of taxes and interest expenses. In a way, NOPAT shows how much revenue is provided by the companies for the capital employed. In other words, it is used to measure investment performance of the company.5

To represent the cost of equity, the expected rate of return as suggested by original or static CAPM as tested by Attanasio (1991) is as follows:

Ei = λ

fm

+

2V m

τm

β im − z im ,

[4]

6

where τm is the wealth-weighted risk tolerance of the investors, λfm is a weighted average of the values of λs for the investors, Vm is the variance of market portfolio, βim is the sensitivity of stock’s return to market portfolio’s return, and zim is the value of non negative constraint. In this model no riskless borrowing is allowed, since negative positions are precluded. In its traditional form, 2/τm is substituted with the risk premium per unit of variance and λfm is interpreted as the expected rate of return on any "zero beta portfolio" or as riskless asset (if available).6 Hence, the following traditional CAPM of Sharpe (1964, 1991) is obtained: E(Ri)= RF + (E(Rm)- RF) βim ,

[5]

where E(Ri) refers to the expected rate of return, RF refers to the risk free rate of return, and Rm refers to the rate of return of market portfolio. The economic profit formulation below is equivalent to the economic value added of Stewart (1990) by substituting as cost of equity, the expected rate of return which is obtained by the capital asset pricing model as the cost of equity. EVA = NOPAT - C [wdRd+we (RF + (E(Rm)- RF) βim)],

[6]

= [ROI - (wdRd+we(RF + (E(Rm)- RF) βim))] C ,

[7]

where wd and we refers to the weights of debt and equity capital to total capital employed respectively, Rd refers to the cost of debt, and Re refers to the cost of equity.

Should the rate of returns as the basis for the opportunity costs become considerably high and volatile in the emerging markets,7 the cost of capital will be high. In such a case, since the cost of capital in the EVA formula would be high also, and the view that companies decreasing wealth will be dominant. Moreover, in the inflationary economies such as Turkish capital

7

market, the unstable pattern of the market risk premium (Rm-RF) and the insufficient use of a long-run model by using relatively short term rate of returns restrict the use of CAPM in economic profit measurement. Thus, the use of traditional dividend approach of Gordon (1959) will be more appropriate in the calculation of cost of capital. On the other hand, the restrictive impacts of the offering rate and/or floating rate in the stock market must be considered in an efficient pricing. Under these conditions, the economic profit within the context of the cost of capital model of Gordon (1959) is stated below.8 EP = NOPAT - C ((d/P)+g)

[8]

where among the notations additionally used in the equation (8); d refers to the dividend of company that may be expected to be paid next period, g refers to the expected growth rate, and P refers to the price of stock at the beginning of period. Thus, the cost of capital is explained in terms of the earnings. As the dividend yield represents the cost of capital, the borrowing capacity of the companies will be related to the profitability, payout ratio, and the taxation. Once the borrowing instruments are considered in terms of either amount or taxation potential, the following economic profit equation is derived:

EP = NOPAT - C [wdRd+we (d/p+g)]

[9]

In this equation of economic profit, while the cost of equity will be dependent on the expected dividend, the borrowing cost will be described by the loan costs in the markets.9

The relation of economic profit to the market value added can be evaluated as Stewart (1990) and Grant (1996):

Market Value Added = Net present Value of Future Economic Profits

8

[10]

Equation (10) says that the investors would aim to increase the market value of the company’s assets over their book value if any wealth increase were expected with in the future. In a contrary case, the company’s assets value (equivalent to the book value of the debts and equity) will be realized over or below the market value. Since the specified explanations have attempted to reveal wealth increase/decrease of different companies, the non-dimensional measurements are considered in the study. We can then easily conclude that the empirical analysis is concordant for the MVA-to-Capital and EP-to-Capital as Grant (1996) and Rappaport (1997) have already used. (MVA/C)t =a+b(EP/C)t+1 +ε

[11]

According to the linear regression formula (11); the change in the MVA-to-C ratio is classically represented by b unit against one unit change of EP-to-C ratio of a company in the market, while t denotes the time, and є indicates the error term.

3 Data and Methodology The sample of companies for the study covers the non-bank firms for which detailed financial data are available for the 1993-1998 period and whose equity shares were traded on the Istanbul Stock Exchange between the period of 1991-1998. The combined balance sheets items for these firms is given in Table 1. When the aggregate market values of 123 companies in the study are examined on a yearly basis, it is observed that the market capitalization share of the sample companies in the stock market has never dropped below 50% as shown in Graph 1. At the end of 1998, it still represented a share over 52%. Thus we assume that the results of the analysis will be sufficient in reflecting the general performance of the market.

9

In order to calculate the β coefficients, the rate of returns for 48 months preceding the end of each accounting period are used. The β coefficients estimated in the study are presented in Table 2. As the rate of returns have been calculated ex-post by using the market values, no capital increase correction has been required while the correction is made for the payment of dividends. Average public borrowing rates with 1-year maturity is considered instead of 3-monthtreasury bills to represent the risk free rate (chart 2), because the longer the maturity, the more the stable pattern of interest rates. As is shown on chart 2, the interest rates of public borrowing have already appeared with certain variation on a yearly basis. Therefore, when costs at the beginning of the term are specified, the interest rates of the last month of the preceding year have been taken into the consideration.10 Moreover, the borrowing rate used in the calculation of economic profit is dependent on the foreign exchange fund representing the values that have been permitted to use by the development banks or the values that are specified with the inflation adjustment according to Fisher (1930) for the direct foreign currency loans.11 It is assumed that all companies have carried out all of their borrowings at the stated interest rates.

The analysis of variables used with the regression analysis has been performed under the models given in Table 3. NOPAT has been included in all of the models identically. On the other hand, the capital amount of investments is considered to be equivalent to the total assets. The basic difference of the models is in terms of the cost of capital. Model 1 and 2 are comparable to the economic value added of Stewart (1990). In model 3, 4 and 5 rather than using cost of capital, net present value of dividends for the following 4 years discounted for inflation is used. Thus, we include the inflation and growth on an ex-post basis and then 10

dividend to price ratios are estimated. Model 4 and 5 represent the models by considering WACC. Model 5 with market value weighted cost of capital differs from Model 4.

The findings related to the cross-sectional regression models for each year of 1993-1998 period are given on a yearly basis in Table 4. In the linear regression models, the determination coefficients a and b specify the relation between MVA/C and EP/C.12

4 Empirical Results When the contents of Table 4 are analyzed, the most interesting finding seems to be the lack of any significant relation for the first two models, while the model 3, 4, and 5 are significant at 95% confidence level. Except for the year 1994, the coefficients of linear models 1 and 2 are insignificant at 95% in terms of t statistics. On the other hand, model 4 has a higher explanatory power than model 3 on the basis of their respective adjusted R2 and F-statistics. For the year 1993, the value of R2 has been calculated 3.2% for Model 3, and 6.8% for Model 4. Again, for the year 1998, while the former is 9%, the latter is 17%. Once all years are considered, it is demonstrated that the measurement of cost of capital using dividends is fairly consistent, thus the cost of borrowing is required to be distinguished and considered separately. Since leverage is specified more realistically in model 5, it is available to reflect the priority of the market participants and more significant findings were obtained accordingly.

Looking at model 5; it seems that the (b) coefficient has changed each year.13 The rate has increased significantly to 2.5 in 1994 from 0.81 in 1993; decreased down to 1.62 in 1995, increased again to 2.04 in 1996 and with a rapid increase reached to 8.85 in 1997. Since the rate has a positive sign, it is assumed to be an indicator to show that an additional market value 11

is created during the current period against the expected economic profit. At the end of 1993, the shareholders had to produce market added value of 0.81 unit for the company which is to achieve and realize an economic profit of 1 unit. Because this rate is at a level of 8.85, a relatively high value in 1998, the investors had a favorable outlook on economic profitability. On the other hand, economic expectations were not as favorable in 1994 (see footnote 13). We imply that the change of the magnitude of the b coefficient year by year is as a result of the omitted variables from the regression. Since we do not investigate the effect of macroeconomic factors on the economic profits during the sample period, we should be cautious in concluding that the b coefficients are rather high or low.

The findings on Table 5 is calculated for the 35 largest firms out of a total 123 firms. 35 largest represent 75% of sample’s asset values. Results based on larger companies are more significant in terms of R squares. In terms of model 5, the results for larger companies are more significant compared to total sample especially for b coefficients for 1993 and 1995. These results confirm the studies of Grant (1996) and Graham (1996). To test an additional argument of this study, the opinion that the criterion of offering rate to public having restrictive impact on the correct measurement of capital costs, Model 5 is used to study the companies for which offering rate to public was over 30%. Such companies in question were of 68 out of 123. When the adjusted coefficient of determination of the regression, R2, is analyzed, the model did not turn out to have more explanatory power when compared to the findings of 123 companies. The general evaluation of the above results show that, due to the volatility of risk free rate of returns and market risk, a reliable capital cost based on CAPM and short term data cannot be used in emerging markets. The interest rate policy of government and the shortage of private

12

sector bonds make it difficult to calculate a consistent capital cost for the companies. According to the findings obtained, more significant models are made available by considering the dividend pay-out as the equity capital cost for the companies.

5 Conclusion This paper focuses on the economic profit measurement which modifies Stewart’s (1990) economic value added (EVA) approach. Considering the cost of capital in a different perspective, we recommend an economic profit model and compare it with the EVA in Turkish stock market.

Economic profit approach is mathematically equivalent to discounted cash flows. Therefore, using the performance measure in the accounting profit that is calculated at the end of the term by considering the weighted average cost of capital (WACC) is in line with the investment decisions. The market inefficiencies and the uncertainty related to future estimations restrict the practical use of the economic value added (EVA) as one of the economic profit measurements. From the investor’s point of view, in evaluating company performance, cost of capital and wealth maximization become the most important issues.

One of the main shortcomings in the calculation of EVA in inefficient markets is related to the CAPM approach. Even though the cost of equity calculation can be performed by including β, it is clear that the future expectations are not realized in practice with the ex-post analysis in the inefficient markets. Should the rate of returns as the basis for the opportunity costs become considerably high and volatile in the emerging markets,14 the cost of capital will be high. In such a case, since the cost of capital in the EVA formula would be high also, and the view that companies decrease wealth will be dominant. Moreover, in the inflationary economies such as 13

Turkish capital market, the unstable pattern of the market risk premium (Rm-RF) and the insufficient use of a long-run model by using relatively short term rate of returns restrict the use of CAPM in economic profit measurement.

In the calculation of economic profit in the study; EVA, MVA, and SVA measures are considered. Required descriptions in connection with the efficiency of use have been given, after suggesting static CAPM in the measurement. The rate of returns according to CAPM have been calculated in two ways: first by taking the tax deductibility of borrowing cost for the company into consideration, and secondly by considering the cost of equity in the calculation of weighted average cost of capital. Almost all of the company EVAs for the period of 1994-1998 have resulted in negative values. Thus, from these results, it may be concluded that the companies in sample have not created any economic value added in terms of macro economic significance.

The cash flow and the impact of inflation resulting from term differential and investment periods should be considered in the calculations. Besides, the performances of companies operating in different industries may be changed with respect of inflation. In this study, instead of an inflation adjustment for all variables, the economic profit calculation with the context of capital cost model of dividend discount has been favored. Hence, the current period expected profit values are specified after having adjusted the profit figures that are already realized by the companies through inflation. Within such a framework, in the environments where the risk-free rate of return is high, the Gordon model of weighted average cost of capital (WACC) is included in EP formula. Regressing the MVA to Capital ratio with the EP to Capital, we demonstrate that the model offered is more suitable than the EVA approach especially in an emerging market such as Turkey. Moreover, the findings based on larger companies in the

14

sample are more explanatory and significant.

For further research, the borrowing skills of the companies, and the incentive and investment credits granted may also be included in the analysis. However, such studies must be done in accordance with the efficiency of the market. On the other hand, as in the developed markets, it is required to have credit rating institutions focused on joint stock companies. As a matter of fact, had such ratings existed, it would have served to carry out more accurate specification of borrowing costs of the companies, while the investors would have trade with more accurate information accordingly.

When any economic profit criterion is used, adjustments for different types of markets become necessary. As demonstrated in this study, when an emerging market is considered, instead of the economic value added approach, the economic profit approach which contains dividends adjusted for inflation is recommended.

References Attanasio, Orazio P., 1991, "Risk, Time Varying Second Moments and Market Efficiency," Review of Economic Studies, No.58, 479-494. Beaekert, G., Harvey, C.R., 1997, “Emerging Equity Market Volatility,” Journal of Financial Economics, Vol. 43, No.1, (January), 29-77. Berk, N., 1999, Financial Management, (in Turkish) Fifth Edition, Türkmen Yayınevi, Istanbul. Chow, G.C., 1960, “Tests of Equality between Sets of Coefficients in Two Linear Regressions,” Econometrica, Vol. 28, No.3, 591-605. Cleary, 1999, “ De-Villiers, J., 1997, “The Distortions in Economic Value Added (EVA) Caused by Inflation”, Journal of Economics and Business, Vol.49, No.3, May-June, 285-300 Downie, Jack, 1958, The Competitive Process, Gerald Duckworth & Co. Ltd., London. Erb, C.B., Harvey, C.R., 1995, “Country Risk and Global Equity Selection,” Journal of Portfolio Management, Vol.21, No.2, 74-83. Erdogan, O., 1996, Comparable Approach to the Theory of Efficient Markets, A Modified Capital Asset Pricing Model, Capital Markets Board, No. 53, Ankara. Fama, E.F., French, K.R., 1992, “The Cross Section of Expected Stock Returns,” Journal of Finance, Vol. 47, 427-465. 15

Fama, E.F., French, K.R., 1999, “The Corporate Cost of Capital and the Return on Corporate Investment,” Journal of Finance, Vol. 54, No. 6, 1939-67. Fama, E.F., Miller, M.H., 1972, The Theory of Finance, New York. Fisher, I., 1930, The Theory of Investment, Augustus M. Kelley, Publishers (1965), New York. Fisher, Anne B., 1995, “Creating Stockholder Wealth”, Fortune, Chicago, Vol. 132, No. 12, Dec 11, 105. Friedman, M., 1956, “The Methodology of Positive Economics”, Essays in Positive Economics, University of Chicago Press, Chicago. Gapenski, L., 1996, “Using MVA and EVA to measure financial performance”, Healthcare Financial Management, Vol.50, No.3, March. 50. Gordon, M.J., 1959, “Dividends, Earnings, and Stock Prices,” Review of Economics and Statistics, Vol.41, No.2, May, 99-105. Graham, J.R., 1996, “Taking Stock of Economic Value Added”, Consultants News, Vol.26, No.11, November, 6-7. Grant, James L., 1996, “Foundations of EVA for Investment Managers,” Journal of Portfolio Management, Fall, 41-47. Grant, James L., 1997, Foundations of Economic Value Added, F. J. Fabozzi Associates, USA. Grinold, Richard C., 1993, "Is Beta Dead Again?" Financial Analysts Journal, July-August, 28-34. Gultekin, N.Bülent, 1983, "Stock Market Returns and Inflation: Evidence from Other Countries", the Journal of Finance, Vol. 38, No.1, March, 49-65. Jensen, M.C., Meckling, W., 1976, “Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure”, Journal of Financial Economics, Vol. 3, October, 305-60. Kaplan, S.N., Zingales, L., 1997, “Do Investment-Cash Flow Sensitivities Provide Useful Measures of Financing Constraints?” The Quarterly Journal of Economics, February, 169-213. Katırcıoğlu, E., 1987, “The Concept of Money Capital and the Theory of Firm,” The Proceedings of the XI. International Congress on Operational Research, Vol.2, Istanbul, 92-110. Kwon, Shin, Bacon, 1997, “The Effect of Macroeconomic Variables on Stock Market Returns in Developing Markets,” Multinational Business Review, Vol. 5, No. 2, 63-70. Lawrence, S., 1995, “Comment on "EVA: Reinventing the Wheel", Chartered-AccountantsJournal-of-New-Zealand, Vol. 74, N.4, May, 36-38. Levonian, Mark E., 1994, “The persistence of bank profits: What the stock market implies”, Economic Review - Federal Reserve Bank of San Francisco, No.2, 3. Lintner, J., 1965, “The Valuation of Risk Assets and The Selection of Risky Investment In Stock Portfolios and Capital Budgets”, Review of Economics and Statistics, 47, 1347. Markowitz, H., 1999, “The Early History of Portfolio Theory: 1600-1960,” Financial Analysts Journal, Vol.55, No.4, (Jul/Aug), 5-16. Modigliani, F., Miller, M.H., 1958, “The Cost of Capital, Corporation Finance, and the Theory of Investment", American Economic Review, 261-97. Modigliani, F., Miller, M.H., 1961, “Dividend Policy, Growth, and the Valuation of Shares", The Journal of Business, Vol.34, No.4, 411-33. Mookerjee, R., 1992, “An Empirical Investigation of Corporate Dividend Pay-Out Behaviour in an Emerging Market,” Applied Financial Economics, Vol. 2, No. 4, December, 243-47. Mossin, J., 1966, “Equilibrium in a Capital Asset Market,” Econometrica, Vol. 35, No.4, 16

(October), 76-83. Pastor, L., Stambaugh, R.F., 1999, “Costs of Equity Capital and Model Mispricing,” The Journal of Finance, Vol. 54, no.1, 67-121.February, Rappaport, A., 1979, “Strategic Analysis for More Profitable Acquisitions,” Harvard Business Review, July/August. Rappaport, A., 1998, Creating Shareholder Value, The Free Press, USA. Schnabel, J.A., 1985, “On Cash Demands, Dividend Yields, and the CAPM,” Journal of Business Research, June, 259-66. Sharpe, F. W., 1964, ”Capital Asset Prices: A Theory of Market Equilibrium under Conditions of Risk,” Journal of Finance, Vol.19, No.3, September, 425-442. Sharpe, William F., 1991, "Capital Asset Prices With and Without Negative Holdings," the Journal of Finance, Vol. XLVI, No.2, June, 489-509. Shanken and Smith, 1996, “Implications of Capital Markets Research for Corporate Finance”, Financial Management, Vol. 23, No.1, Spring, 98. Stewart III, G. Bennett, 1990, The Quest for Value, Harper Collins Publishers, USA. Young, D., 1999, “Some Reflections on Accounting Adjustments and Economic Value Added,” Journal of Financial Statement Analysis, Vol. 4, No. 2, Winter, 7-19.

17

ENDNOTES * Oral Erdogan,

Ph.D.,

is a staff member in the Istanbul Stock Exchange, and is a visiting lecturer in the Yildiz

Technical University and Marmara University ([email protected]); Niyazi Berk is a Professor of Accounting and Finance in Marmara University; Erol Katırcıoglu is a Professor of Economics in Marmara University. Comments from Noyan Arsan of the State University of West Georgia were very helpful. Vedat Akgiray, Nazim Engin, Burak Saltoglu, and conference participants at the 1999 METU International Conference on Economics III also provided useful comments. We also wish to thank the Editors and especially the referee whose comments and suggestions greatly improved the paper. We are responsible for any remaining errors. 1

See Shanken & Smith (1996) and Levonian (1994) for Economic Profit; See Jensen & Meckling (1976), and

Rappaport (1998) for Shareholder Value Added, SVA; See Stewart (1990), and Grant (1996, 1997), for Economic Value Added, EVA; see Young (1999) for the early studies on the residual income from 1920 to 1950s, and see Lawrence (1995) for the similarity of the EVA and the residual income. 2

Stewart (1990) uses the CAPM for measuring the cost of capital.

3

Although the Capital Asset Pricing Model (CAPM) of Sharpe (1964), Lintner (1965), and Mossin (1966) has

been supported with many researches so far, there have also been many opponents [See. Cheng & Grauer (1980), Fama and French (1992), Grinold (1993)]. Erdogan (1996) and Markowitz (1999) also dedicate it to Tobin (1958). 4

Cost of Capital (COC) in Equation 2 is tax adjusted through multiplying by (1-TC D/C); where TC is the

corporation tax rate, D is the book value of debt, and C is the equity . The value of T used in this study is 46 % for all companies. 5

For the theoretical explanations of profit concept in Accounting and Finance see Downie (1958); Stewart (1990,

68-117); Rappaport (1998), and Berk (1998). 6

For a comparison see Erdoğan (1996).

7

Bekaert and Harvey (1997) indicate that the volatility is an important factor for determining the cost of capital in

emerging markets. 8

In the empirical part of the study, as expected (average) net income is calculated ex-post from the historic

realized net incomes and adjusted to inflation rates, the growth rate is assumed to be included indirectly.

18

9

Moreover, the borrowing rate in the calculation of economic profit is subject to the foreign exchange fund

representing the values that have been permitted to use by the development banks. It is assumed for the simplicity that all companies have carried out all of their borrowings with the stated interest rates. 10

The related interest rates are 87.1%, 104.8%, 99%and 92% for the last month of 1993, 1994, 1996, and 1997

respectively. Since there was no treasury bill outstanding in December 1995, the monthly average of 1995 is calculated as the risk free rate of return. That is 105%. The values are available in the Web-site Central Bank of Republic of Turkey (www.tcmb.gov.tr). 11

We are grateful to Sınai Yatırım Bankası and Türkiye Sınai Kalkınma Bankası for providing the related data.

The investment credit rates at the beginning of 1994, 1995, 1996, 1997, and 1998 are 63.9%, 86.4 %, 81 %, 105 %, and 90 % respectively. 12

R square is the coefficient of multiple determination to identify what effect the independent variables have on

the dependent variable. t refers to tail significance for the coefficients; N shows the number of observations. 13

According to Chow’s (1960) test we apply for each successive year; we infer that there is a statistically

significant change, at 1% level, in the economic profit functions from 1993 to 1994 and from 1996 to 1997. For the other pair of regressions, we cannot reject the null hypothesis that both regressions propose the equal functions. For each pair: N=246; df=244. 14

Bekaert and Harvey (1997) indicate that the volatility is an important factor for determining the cost of capital

in emerging markets.

19

Graph 1: The Share of the Sample in Market Capitalization

100 90 80 70 60 50 40 30 20 10 0

72%

69%

65% 51%

52%

1997

1998

%

77%

1993

1994

1995

1996 Years

Note: Total market capitalization includes banks as well.

Graph 2: Public Internal Borrowing Rate for 1-Year Maturity

200

Interest R ates %

180 160 140 120 100 80 60 3 93 93 3 94 4 4 5 95 95 -95 -96 -96 -96 -96 -97 -97 -97 -98 t r n b y g n-9 y- g- v-9 r- n-9 p-9 n-9 r- ulc p n Ja Ma Au No Ma Ju Se Ja Ap J Oc Fe Ma Au De Ma Ju Se Ja Source: Central Bank of Republic of Turkey, CBRT, Ankara 1999.

20

Table 1: Main Items of Consolidated Balance Sheet and Capitalization of 123 Companies (Billions, Turkish Lira) Year

Total Assets Current Assets Fixed Assets Equity

Net Profit

Capitalization

1993

183,9

102,3

81,6

85,8

19,0

420,4

1994

407,3

218,7

188,5

188,2

40,3

605,2

1995

796,5

441,2

355,3

393,5

100,1

868,4

1996

1.468,8

816,9

651,9

669,0

165,9

2.115,4

1997

2.866,2

1.635,2

1.231,0

1.287,2

317,7

6.491,2

1998

4.937,1

2.668,3

2.247,6

2.328,2

405,0

5.487,3

Source: ISE Company Yearbooks 1994-1998.

Table 2: Characteristics of βs

Mean Standard Error Median Standard Dev. Kurtosis Skewness No. of Obs. Confidence Lev. (95%)

1,18 0,05 1,15 0,55 12,99 2,45 123 0,098

1,15 0,03 1,13 0,38 6,92 1,44 123 0,068

1,17 0,03 1,18 0,37 3,31 0,75 123 0,065

0,98 0,03 1,00 0,28 0,13 0,11 123 0,051

0,86 0,02 0,86 0,25 1,20 0,00 123 0,044

0,96 0,02 0,97 0,24 2,93 0,66 123 0,044

Note: Number of firms included 123. Responded market portfolio return is the change in the ISE 100. security i is the covariance of the security with the market, divided by the variance of the market. Cov [R i , R M ] . β = Var

β of

[R M ]

Table 3: The Models Model

Model

Cost of Capital

Model

(MVA/C)t=a+b(EP1/C)t+1

(RF + (E(Rm)- RF) βim)(1-TCD/(C-D))C

Model

(MVA/C)t=a+b(EP2/C)t+1

[we(RF + (E(Rm)- RF) βim)+wdRd]C

Model

(MVA/C)t=a+b(EP3/C)t+1

(d/P)C

Model

(MVA/C)t=a+b(EP4/C)t+1

(we(d/P)+wdRd)C

Model

(MVA/C)t=a+b(EP5/C)t+1

(we(d/P)+wdRd)C

*: In model 5, the leverage ratios are market value weighted.

21

Table 4: Regression Results for Linear Model between MVA/C and EP/C Model Year: 1993 Model 1 Model 2 Model 3 Model 4 Model 5

a

b

ta

tb

Adj.R2

F value

N

1,412820 1,313928 0,826676 0,880746 0,757522

0,277224 0,055458 0,604078 0,781327 0,810838

7,71* 7,85* 3,20* 4,39* 3,36*

1,15 0,22 2,26** 3,14* 3,21*

0,00 0,00 0,032 0,068 0,071

1,32 0,1 5,1** 9,9* 10,3*

123 123 123 123 123

Year: 1994 Model 1 Model 2 Model 3 Model 4 Model 5

0,652848 0,6407591 0,188026 0,748584 0,464131

-0,519959 -0,461604 1,533994 2,131048 2,504026

5,42* 5,35* 1,11 6,88* 4,61*

-2,28** -2,20** 3,13* 5,84* 7,18*

0,03 0,03 0,07 0,21 0,29

5,2** 4,8** 9,8* 34,1* 51,5*

123 123 123 123 123

Year: 1995 Model 1 Model 2 Model 3 Model 4 Model 5

0,061080 0,050443 -0,037207 0,197864 0,068403

-0,403781 -0,402085 0,657350 1,619388 1,617495

0,55 0,44 -0,25 2,19 0,77

-1,75 -1,74 1,80 4,67* 5,53*

0,02 0,02 0,02 0,15 0,20

3,1 3,0 3,3 21,8* 30,5*

123 123 123 123 123

Year: 1996 Model 1 Model 2 Model 3 Model 4 Model 5

0,293883 0,268463 0,066351 0,619042 0,318379

-0,44325 -0,426090 0,961082 1,912219 2,037988

1,97 1,66 0,28 5,07* 2,78*

-1,57 -1,48 1,88 5,04* 5,99*

0,01 0,01 0,02 0,17 0,22

2,5 2,2 3,5 25,4* 35,9*

123 123 123 123 123

Year: 1997 Model 1 Model 2 Model 3 Model 4 Model 5

0,988267 1,016540 -0,07995 3,326262 1,598727

-0,538672 -0,360459 6,817981 6,602580 8,854257

1,94 1,79 -0,15 5,98* 4,62*

-0,49 -0,32 3,53* 5,05* 6,15*

0,00 0,00 0,09 0,17 0,23

2,5 0,03 12,5* 25,5* 37,9*

123 123 123 123 123

Note: Dependent variable is MVA/C, Explanatory Variable is EVA/C or EP/C; a is the intercept, b is the slope coefficient; ta and tb denotes the t values. They are simply the ratios of the estimated coefficients to their standard errors. F test values show the statistically significance of the regression functions. N refers to the number of the observations. * : significant at the 0.01 level. **: significant at the 0.05 level.

Table 5: Results for Larger Firms Years 1993 1994 1995 1996 1997

A

0,424747 0,562102

0,145718 0,550572

1,574155

B 1,728295 2,925896 2,330958 2,100074 7,968062

ta 0,71 2,88* 0,90 2,24** 4,59*

tb 2,38** 4,84* 4,44* 3,20* 5,14*

R2 0,12 0,40 0,36 0,21 0,43

F 5,6* 23,5* 19,8* 10,3* 26,5*

N 35 35 35 35 35

Note: Model 5 is applied here. The total asset value of the sample is 75-76 % out of 123 firms. Dependent variable is MVA/C, Explanatory Variable is EP/C. * : significant at the 0.01 level. **: significant at the 0.05 level.

22

Table 6: Results for Companies with Higher Offering Rates Year (MVA) 1993 1994 1995 1996 1997

a 0,7360263 0,3191521

-0,016987

b 0,659272 1,924181 1,2383764

0,1508509

1,4651409

0,6976725

2,9020326

ta 2,99* 3,32* -0,17 1,45 5,42*

* : significant at the 0.01 level. **: significant at the 0.05 level.

23

tb 2,70* 5,98* 3,91* 4,42* 5,68*

R2 0,09 0,34 0,18 0,22 0,32

F 7,3* 35,7* 15,3* 19,5* 32,3*

N 68 68 68 68 68

Abstract This paper focuses on the economic profit measurement which modifies Stewart’s (1990) economic value added (EVA) approach. Considering the cost of capital in a different perspective, we recommend an economic profit model and compare it with the EVA in Turkish stock market. Economic profit (EP) is the difference between the return on equity capital and the opportunity cost of investing. Similar concepts used in finance literature are residual income, economic value added, and shareholder value added (SVA).1 Despite the abundance of discussion and the extensive use of these techniques by the corporations, not enough empirical research has been conducted in this area. In the empirical part of this study, we calculate both the EVAs and EPs on 123 firms trading in the Turkish stock market. We subscribe to the market value added (MVA) approach and demonstrate that the model offered is more suitable than the EVA approach especially in an emerging market such as Turkey.

(JEL Classification: G1, G3, L2)

1 Introduction The concept of “economic profit” is included in the wealth maximization and the theory of the firm, and to a larger extent within the theory of finance.

According to Fisher (1930), consumers optimize their savings and consumption preferences in financial markets through their relationships with firms which provide services and income. As stated by Fama and Miller (1972), finance theory considers allocation of resources by companies to various investment areas in different periods. Thus, in the efficient transfer of funds and savings, the existence of capital markets become important. According to neoclassical theory, the ownership of the companies by the households generates services and income to these households which in turn increase their wealth positions. Such increase in wealth brings up the issue of shareholder wealth maximization. However, rather than the normative approach of neoclassical economics, a method to measure the investment performances of companies needed to be developed in line with the positive economics of 1 Electronic copy available at: http://ssrn.com/abstract=880020

Friedman (1958).

In the literature, the free cash flow (FCF) and corporate valuation of Modigliani and Miller (1958, 1961) constituted main references to economic profit. Direct relationship between the company decisions and investments has been stressed by Cleary (1999). This view supports the results of the study on the investment and cash flow sensitivity of Kaplan and Zingales (1997). Once the theoretical base has been considered, practical solutions measuring the wealth level of the family members (hence the shareholders) have become one of the financial research fields from 1970s to present. Rappaport (1979, 1998), Stewart (1990), and Grant (1996) have contributed by relating the free cash flow approach to the concept of value added. They introduced the concept of value added by using “economic profit” and “cost of capital” approaches.

Firms create or destroy wealth depending on their economic behavior and their membership in a particular industry.

A firm aims to maximize its profit level, while its shareholders

maximize their capital gains and dividends from their investment. The creditors, on the other hand, expect to receive their capital back including interest. For that reason, the valuation of companies is associated with the maximization of the shareholders wealth. However, in inefficient capital markets, inefficient market values may exist due to the manipulation, insider trading, asymmetric information, and/or insufficient liquidity and market capitalization. If the demanded funds can not be transferred in time, incorrect valuation may occur. Besides, excess fund transfer to one firm could lower the market values of other firms. In emerging markets, the ratio of floating shares to outstanding shares becomes a limiting factor in valuation due to small market size. Due to limited demand for more shares, the market value of a firm with the lower ratio could be higher than that of the firm with the higher ratio.

2 Electronic copy available at: http://ssrn.com/abstract=880020

The investors usually take into account the return on equity, market capitalization, market-tobook value ratio, earnings per share, and price-to-earnings ratio to invest in the shares. However, the most convenient answer to the question of “has the capital of creditor or the shareholder been increased or decreased by the company?” might be given by the market value added (MVA) criterion (Fisher 1995). Stewart (1990) defines the MVA as the difference between the total market value of firm and the book value of total assets employed in the business. MVA greater than zero means how much value company has added to its shareholders' investment. MVA less than zero means how much value company has subtracted from its shareholders' investment.

Compared to MVA, Stewart’s (1990) EVA is the company’s after-tax net operating profit in a given year minus the cost of capital.2 Gapenski (1996) states that both MVA and EVA are applicable to investor-owned organizations; however, EVA also is an appropriate measure for non-profit organizations. This view is acceptable since the managerial effectiveness in a given year can be examined by EVA.

The consideration of cash flows from investments, the investment periods and the term differential in the calculation of the impact of inflation is specified by Levonian (1994), DeVilliers (1997), Rappaport (1998), and Fama and French (1999). Pastor and Stambaugh (1999) imply that uncertainty about factor premiums is generally the largest source of overall uncertainty about a firm’s cost of equity, even though uncertainty about betas is almost as important. Thus, especially in those markets with a high risk-free rate, a weighted average cost of capital (WACC) may be used in economic profit measurement similar to ModiglianiMiller’s (1958) and Gordon’s (1959) model.

3

By supporting the use of EVA from several points, Graham (1996) clarifies that especially consultants use economic value added (EVA) for strategic planning work, financial advice and compensation studies. We concur with him that EVA may be questionable for companies with few assets, those where major assets are leased, and growth companies.

Kwon, Shin and Bacon (1997) demonstrate that the investors’ perception of stock returns in the Korean market, an emerging market, are quite different from those of US and Japanese investors suggesting that the dividend yield, foreign exchange rate, oil price and money supply are strongly correlated with stock returns. Erb and Harvey (1995) explain the different risk structure and portfolio composition between the emerging and developed markets. They conclude that dividend yield is more useful in emerging markets in measuring investment returns. Moreover, Schnabel (1985) needs to extend the CAPM by including a dividend in the new security market line, as a means of meeting the cash demands and reducing the financial distress costs. Mookerjee (1992) shows that dividend payments are rational when the unique institutional environment in India is taken into account as an emerging market, which is contrary to recent evidence from the US and Japan. Finally, Fama and French (1999) confirms that a by-product of calculating the real cost of capital and the real return on cost is information about the history of corporate earnings, investment and financing decisions. They accept the internal rate of return on value is a starting point for cost of capital estimates for individual firms.

Within such a framework, in the environments where the risk-free rate of return is high, the Gordon model of weighted average cost of capital (WACC) is included in EP formula.

4

Section 2 of the paper presents an extended model of economic profit. In Section 3, data and the methodology of data analysis used are presented. Section 4 covers the empirical results of the study followed by the conclusion section.

2 The Model Simple introduction and definition of the preferences of the investors in capital markets with the indifference curves provides general equilibrium in stock markets. Thus, in stock markets consumption preferences of the shareholders and potential investors as well as the performance of companies will be determined. Having advanced finance theory to a further stage, CAPM represents the most common positive proposal for the perfect capital markets. It reaches to the theory of capital markets, to the theory of “fair play” or” martingales” and thus to the theory of speculation of Bachelier (1900). With such a model, the asset pricing is available on the basis of the expected return from the capital markets:

E( Pi , t + 1 Φt ) = (1 + E( Ri , t + 1 Φt ) pit ,

[1]

where E refers to the expected value, P refers to the price of the security, R refers to the rate of return for one period, Φ refers to the information set with regard to the security for a specified period, i refers to the security, and t refers to time period. On the basis of this statement, the investor’s utility maximization in terms of the capital markets theory and on the one hand and the company profit maximization in terms of the theory of firm on the other may well intersect.

This study is heavily dependent on the views by Stewart’s (1990) EVA and Rappaport’s (1998) SVA. To adjust for inflation, suggestions by de-Villiers (1997), Rappaport (1998), and

5

Fama and French (1999), and to estimate cost of capital, classical Sharp (1964, 1991)3 and Gordon (1959) models have been used. We will first present the economic profit as used in EVA. The financial expenses are also included in the net profit account indicating the company profit with “the principal of considering the capital cost” in the economic profit equation.

EP = NOPAT - (C x COC) = (ROI - WACC) x C ,

[2] [3]

where; NOPAT = Net Operating Profit + Financial Expenses C : Total Capital Employed C0C4: Cost of Capital

ROI: Return on Investment WACC: Weighted Average Cost of Capital. Although the company has earned financial expenses, they still represent the expenses as the cost of borrowing. Once the cost of capital concept is considered while leaving the accounting technique aside, the NOPAT has been identified as a new profit measure. Thus, NOPAT is determined by adding the interest expenses to the figure of net profit after the deduction of taxes and interest expenses. In a way, NOPAT shows how much revenue is provided by the companies for the capital employed. In other words, it is used to measure investment performance of the company.5

To represent the cost of equity, the expected rate of return as suggested by original or static CAPM as tested by Attanasio (1991) is as follows:

Ei = λ

fm

+

2V m

τm

β im − z im ,

[4]

6

where τm is the wealth-weighted risk tolerance of the investors, λfm is a weighted average of the values of λs for the investors, Vm is the variance of market portfolio, βim is the sensitivity of stock’s return to market portfolio’s return, and zim is the value of non negative constraint. In this model no riskless borrowing is allowed, since negative positions are precluded. In its traditional form, 2/τm is substituted with the risk premium per unit of variance and λfm is interpreted as the expected rate of return on any "zero beta portfolio" or as riskless asset (if available).6 Hence, the following traditional CAPM of Sharpe (1964, 1991) is obtained: E(Ri)= RF + (E(Rm)- RF) βim ,

[5]

where E(Ri) refers to the expected rate of return, RF refers to the risk free rate of return, and Rm refers to the rate of return of market portfolio. The economic profit formulation below is equivalent to the economic value added of Stewart (1990) by substituting as cost of equity, the expected rate of return which is obtained by the capital asset pricing model as the cost of equity. EVA = NOPAT - C [wdRd+we (RF + (E(Rm)- RF) βim)],

[6]

= [ROI - (wdRd+we(RF + (E(Rm)- RF) βim))] C ,

[7]

where wd and we refers to the weights of debt and equity capital to total capital employed respectively, Rd refers to the cost of debt, and Re refers to the cost of equity.

Should the rate of returns as the basis for the opportunity costs become considerably high and volatile in the emerging markets,7 the cost of capital will be high. In such a case, since the cost of capital in the EVA formula would be high also, and the view that companies decreasing wealth will be dominant. Moreover, in the inflationary economies such as Turkish capital

7

market, the unstable pattern of the market risk premium (Rm-RF) and the insufficient use of a long-run model by using relatively short term rate of returns restrict the use of CAPM in economic profit measurement. Thus, the use of traditional dividend approach of Gordon (1959) will be more appropriate in the calculation of cost of capital. On the other hand, the restrictive impacts of the offering rate and/or floating rate in the stock market must be considered in an efficient pricing. Under these conditions, the economic profit within the context of the cost of capital model of Gordon (1959) is stated below.8 EP = NOPAT - C ((d/P)+g)

[8]

where among the notations additionally used in the equation (8); d refers to the dividend of company that may be expected to be paid next period, g refers to the expected growth rate, and P refers to the price of stock at the beginning of period. Thus, the cost of capital is explained in terms of the earnings. As the dividend yield represents the cost of capital, the borrowing capacity of the companies will be related to the profitability, payout ratio, and the taxation. Once the borrowing instruments are considered in terms of either amount or taxation potential, the following economic profit equation is derived:

EP = NOPAT - C [wdRd+we (d/p+g)]

[9]

In this equation of economic profit, while the cost of equity will be dependent on the expected dividend, the borrowing cost will be described by the loan costs in the markets.9

The relation of economic profit to the market value added can be evaluated as Stewart (1990) and Grant (1996):

Market Value Added = Net present Value of Future Economic Profits

8

[10]

Equation (10) says that the investors would aim to increase the market value of the company’s assets over their book value if any wealth increase were expected with in the future. In a contrary case, the company’s assets value (equivalent to the book value of the debts and equity) will be realized over or below the market value. Since the specified explanations have attempted to reveal wealth increase/decrease of different companies, the non-dimensional measurements are considered in the study. We can then easily conclude that the empirical analysis is concordant for the MVA-to-Capital and EP-to-Capital as Grant (1996) and Rappaport (1997) have already used. (MVA/C)t =a+b(EP/C)t+1 +ε

[11]

According to the linear regression formula (11); the change in the MVA-to-C ratio is classically represented by b unit against one unit change of EP-to-C ratio of a company in the market, while t denotes the time, and є indicates the error term.

3 Data and Methodology The sample of companies for the study covers the non-bank firms for which detailed financial data are available for the 1993-1998 period and whose equity shares were traded on the Istanbul Stock Exchange between the period of 1991-1998. The combined balance sheets items for these firms is given in Table 1. When the aggregate market values of 123 companies in the study are examined on a yearly basis, it is observed that the market capitalization share of the sample companies in the stock market has never dropped below 50% as shown in Graph 1. At the end of 1998, it still represented a share over 52%. Thus we assume that the results of the analysis will be sufficient in reflecting the general performance of the market.

9

In order to calculate the β coefficients, the rate of returns for 48 months preceding the end of each accounting period are used. The β coefficients estimated in the study are presented in Table 2. As the rate of returns have been calculated ex-post by using the market values, no capital increase correction has been required while the correction is made for the payment of dividends. Average public borrowing rates with 1-year maturity is considered instead of 3-monthtreasury bills to represent the risk free rate (chart 2), because the longer the maturity, the more the stable pattern of interest rates. As is shown on chart 2, the interest rates of public borrowing have already appeared with certain variation on a yearly basis. Therefore, when costs at the beginning of the term are specified, the interest rates of the last month of the preceding year have been taken into the consideration.10 Moreover, the borrowing rate used in the calculation of economic profit is dependent on the foreign exchange fund representing the values that have been permitted to use by the development banks or the values that are specified with the inflation adjustment according to Fisher (1930) for the direct foreign currency loans.11 It is assumed that all companies have carried out all of their borrowings at the stated interest rates.

The analysis of variables used with the regression analysis has been performed under the models given in Table 3. NOPAT has been included in all of the models identically. On the other hand, the capital amount of investments is considered to be equivalent to the total assets. The basic difference of the models is in terms of the cost of capital. Model 1 and 2 are comparable to the economic value added of Stewart (1990). In model 3, 4 and 5 rather than using cost of capital, net present value of dividends for the following 4 years discounted for inflation is used. Thus, we include the inflation and growth on an ex-post basis and then 10

dividend to price ratios are estimated. Model 4 and 5 represent the models by considering WACC. Model 5 with market value weighted cost of capital differs from Model 4.

The findings related to the cross-sectional regression models for each year of 1993-1998 period are given on a yearly basis in Table 4. In the linear regression models, the determination coefficients a and b specify the relation between MVA/C and EP/C.12

4 Empirical Results When the contents of Table 4 are analyzed, the most interesting finding seems to be the lack of any significant relation for the first two models, while the model 3, 4, and 5 are significant at 95% confidence level. Except for the year 1994, the coefficients of linear models 1 and 2 are insignificant at 95% in terms of t statistics. On the other hand, model 4 has a higher explanatory power than model 3 on the basis of their respective adjusted R2 and F-statistics. For the year 1993, the value of R2 has been calculated 3.2% for Model 3, and 6.8% for Model 4. Again, for the year 1998, while the former is 9%, the latter is 17%. Once all years are considered, it is demonstrated that the measurement of cost of capital using dividends is fairly consistent, thus the cost of borrowing is required to be distinguished and considered separately. Since leverage is specified more realistically in model 5, it is available to reflect the priority of the market participants and more significant findings were obtained accordingly.

Looking at model 5; it seems that the (b) coefficient has changed each year.13 The rate has increased significantly to 2.5 in 1994 from 0.81 in 1993; decreased down to 1.62 in 1995, increased again to 2.04 in 1996 and with a rapid increase reached to 8.85 in 1997. Since the rate has a positive sign, it is assumed to be an indicator to show that an additional market value 11

is created during the current period against the expected economic profit. At the end of 1993, the shareholders had to produce market added value of 0.81 unit for the company which is to achieve and realize an economic profit of 1 unit. Because this rate is at a level of 8.85, a relatively high value in 1998, the investors had a favorable outlook on economic profitability. On the other hand, economic expectations were not as favorable in 1994 (see footnote 13). We imply that the change of the magnitude of the b coefficient year by year is as a result of the omitted variables from the regression. Since we do not investigate the effect of macroeconomic factors on the economic profits during the sample period, we should be cautious in concluding that the b coefficients are rather high or low.

The findings on Table 5 is calculated for the 35 largest firms out of a total 123 firms. 35 largest represent 75% of sample’s asset values. Results based on larger companies are more significant in terms of R squares. In terms of model 5, the results for larger companies are more significant compared to total sample especially for b coefficients for 1993 and 1995. These results confirm the studies of Grant (1996) and Graham (1996). To test an additional argument of this study, the opinion that the criterion of offering rate to public having restrictive impact on the correct measurement of capital costs, Model 5 is used to study the companies for which offering rate to public was over 30%. Such companies in question were of 68 out of 123. When the adjusted coefficient of determination of the regression, R2, is analyzed, the model did not turn out to have more explanatory power when compared to the findings of 123 companies. The general evaluation of the above results show that, due to the volatility of risk free rate of returns and market risk, a reliable capital cost based on CAPM and short term data cannot be used in emerging markets. The interest rate policy of government and the shortage of private

12

sector bonds make it difficult to calculate a consistent capital cost for the companies. According to the findings obtained, more significant models are made available by considering the dividend pay-out as the equity capital cost for the companies.

5 Conclusion This paper focuses on the economic profit measurement which modifies Stewart’s (1990) economic value added (EVA) approach. Considering the cost of capital in a different perspective, we recommend an economic profit model and compare it with the EVA in Turkish stock market.

Economic profit approach is mathematically equivalent to discounted cash flows. Therefore, using the performance measure in the accounting profit that is calculated at the end of the term by considering the weighted average cost of capital (WACC) is in line with the investment decisions. The market inefficiencies and the uncertainty related to future estimations restrict the practical use of the economic value added (EVA) as one of the economic profit measurements. From the investor’s point of view, in evaluating company performance, cost of capital and wealth maximization become the most important issues.

One of the main shortcomings in the calculation of EVA in inefficient markets is related to the CAPM approach. Even though the cost of equity calculation can be performed by including β, it is clear that the future expectations are not realized in practice with the ex-post analysis in the inefficient markets. Should the rate of returns as the basis for the opportunity costs become considerably high and volatile in the emerging markets,14 the cost of capital will be high. In such a case, since the cost of capital in the EVA formula would be high also, and the view that companies decrease wealth will be dominant. Moreover, in the inflationary economies such as 13

Turkish capital market, the unstable pattern of the market risk premium (Rm-RF) and the insufficient use of a long-run model by using relatively short term rate of returns restrict the use of CAPM in economic profit measurement.

In the calculation of economic profit in the study; EVA, MVA, and SVA measures are considered. Required descriptions in connection with the efficiency of use have been given, after suggesting static CAPM in the measurement. The rate of returns according to CAPM have been calculated in two ways: first by taking the tax deductibility of borrowing cost for the company into consideration, and secondly by considering the cost of equity in the calculation of weighted average cost of capital. Almost all of the company EVAs for the period of 1994-1998 have resulted in negative values. Thus, from these results, it may be concluded that the companies in sample have not created any economic value added in terms of macro economic significance.

The cash flow and the impact of inflation resulting from term differential and investment periods should be considered in the calculations. Besides, the performances of companies operating in different industries may be changed with respect of inflation. In this study, instead of an inflation adjustment for all variables, the economic profit calculation with the context of capital cost model of dividend discount has been favored. Hence, the current period expected profit values are specified after having adjusted the profit figures that are already realized by the companies through inflation. Within such a framework, in the environments where the risk-free rate of return is high, the Gordon model of weighted average cost of capital (WACC) is included in EP formula. Regressing the MVA to Capital ratio with the EP to Capital, we demonstrate that the model offered is more suitable than the EVA approach especially in an emerging market such as Turkey. Moreover, the findings based on larger companies in the

14

sample are more explanatory and significant.

For further research, the borrowing skills of the companies, and the incentive and investment credits granted may also be included in the analysis. However, such studies must be done in accordance with the efficiency of the market. On the other hand, as in the developed markets, it is required to have credit rating institutions focused on joint stock companies. As a matter of fact, had such ratings existed, it would have served to carry out more accurate specification of borrowing costs of the companies, while the investors would have trade with more accurate information accordingly.

When any economic profit criterion is used, adjustments for different types of markets become necessary. As demonstrated in this study, when an emerging market is considered, instead of the economic value added approach, the economic profit approach which contains dividends adjusted for inflation is recommended.

References Attanasio, Orazio P., 1991, "Risk, Time Varying Second Moments and Market Efficiency," Review of Economic Studies, No.58, 479-494. Beaekert, G., Harvey, C.R., 1997, “Emerging Equity Market Volatility,” Journal of Financial Economics, Vol. 43, No.1, (January), 29-77. Berk, N., 1999, Financial Management, (in Turkish) Fifth Edition, Türkmen Yayınevi, Istanbul. Chow, G.C., 1960, “Tests of Equality between Sets of Coefficients in Two Linear Regressions,” Econometrica, Vol. 28, No.3, 591-605. Cleary, 1999, “ De-Villiers, J., 1997, “The Distortions in Economic Value Added (EVA) Caused by Inflation”, Journal of Economics and Business, Vol.49, No.3, May-June, 285-300 Downie, Jack, 1958, The Competitive Process, Gerald Duckworth & Co. Ltd., London. Erb, C.B., Harvey, C.R., 1995, “Country Risk and Global Equity Selection,” Journal of Portfolio Management, Vol.21, No.2, 74-83. Erdogan, O., 1996, Comparable Approach to the Theory of Efficient Markets, A Modified Capital Asset Pricing Model, Capital Markets Board, No. 53, Ankara. Fama, E.F., French, K.R., 1992, “The Cross Section of Expected Stock Returns,” Journal of Finance, Vol. 47, 427-465. 15

Fama, E.F., French, K.R., 1999, “The Corporate Cost of Capital and the Return on Corporate Investment,” Journal of Finance, Vol. 54, No. 6, 1939-67. Fama, E.F., Miller, M.H., 1972, The Theory of Finance, New York. Fisher, I., 1930, The Theory of Investment, Augustus M. Kelley, Publishers (1965), New York. Fisher, Anne B., 1995, “Creating Stockholder Wealth”, Fortune, Chicago, Vol. 132, No. 12, Dec 11, 105. Friedman, M., 1956, “The Methodology of Positive Economics”, Essays in Positive Economics, University of Chicago Press, Chicago. Gapenski, L., 1996, “Using MVA and EVA to measure financial performance”, Healthcare Financial Management, Vol.50, No.3, March. 50. Gordon, M.J., 1959, “Dividends, Earnings, and Stock Prices,” Review of Economics and Statistics, Vol.41, No.2, May, 99-105. Graham, J.R., 1996, “Taking Stock of Economic Value Added”, Consultants News, Vol.26, No.11, November, 6-7. Grant, James L., 1996, “Foundations of EVA for Investment Managers,” Journal of Portfolio Management, Fall, 41-47. Grant, James L., 1997, Foundations of Economic Value Added, F. J. Fabozzi Associates, USA. Grinold, Richard C., 1993, "Is Beta Dead Again?" Financial Analysts Journal, July-August, 28-34. Gultekin, N.Bülent, 1983, "Stock Market Returns and Inflation: Evidence from Other Countries", the Journal of Finance, Vol. 38, No.1, March, 49-65. Jensen, M.C., Meckling, W., 1976, “Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure”, Journal of Financial Economics, Vol. 3, October, 305-60. Kaplan, S.N., Zingales, L., 1997, “Do Investment-Cash Flow Sensitivities Provide Useful Measures of Financing Constraints?” The Quarterly Journal of Economics, February, 169-213. Katırcıoğlu, E., 1987, “The Concept of Money Capital and the Theory of Firm,” The Proceedings of the XI. International Congress on Operational Research, Vol.2, Istanbul, 92-110. Kwon, Shin, Bacon, 1997, “The Effect of Macroeconomic Variables on Stock Market Returns in Developing Markets,” Multinational Business Review, Vol. 5, No. 2, 63-70. Lawrence, S., 1995, “Comment on "EVA: Reinventing the Wheel", Chartered-AccountantsJournal-of-New-Zealand, Vol. 74, N.4, May, 36-38. Levonian, Mark E., 1994, “The persistence of bank profits: What the stock market implies”, Economic Review - Federal Reserve Bank of San Francisco, No.2, 3. Lintner, J., 1965, “The Valuation of Risk Assets and The Selection of Risky Investment In Stock Portfolios and Capital Budgets”, Review of Economics and Statistics, 47, 1347. Markowitz, H., 1999, “The Early History of Portfolio Theory: 1600-1960,” Financial Analysts Journal, Vol.55, No.4, (Jul/Aug), 5-16. Modigliani, F., Miller, M.H., 1958, “The Cost of Capital, Corporation Finance, and the Theory of Investment", American Economic Review, 261-97. Modigliani, F., Miller, M.H., 1961, “Dividend Policy, Growth, and the Valuation of Shares", The Journal of Business, Vol.34, No.4, 411-33. Mookerjee, R., 1992, “An Empirical Investigation of Corporate Dividend Pay-Out Behaviour in an Emerging Market,” Applied Financial Economics, Vol. 2, No. 4, December, 243-47. Mossin, J., 1966, “Equilibrium in a Capital Asset Market,” Econometrica, Vol. 35, No.4, 16

(October), 76-83. Pastor, L., Stambaugh, R.F., 1999, “Costs of Equity Capital and Model Mispricing,” The Journal of Finance, Vol. 54, no.1, 67-121.February, Rappaport, A., 1979, “Strategic Analysis for More Profitable Acquisitions,” Harvard Business Review, July/August. Rappaport, A., 1998, Creating Shareholder Value, The Free Press, USA. Schnabel, J.A., 1985, “On Cash Demands, Dividend Yields, and the CAPM,” Journal of Business Research, June, 259-66. Sharpe, F. W., 1964, ”Capital Asset Prices: A Theory of Market Equilibrium under Conditions of Risk,” Journal of Finance, Vol.19, No.3, September, 425-442. Sharpe, William F., 1991, "Capital Asset Prices With and Without Negative Holdings," the Journal of Finance, Vol. XLVI, No.2, June, 489-509. Shanken and Smith, 1996, “Implications of Capital Markets Research for Corporate Finance”, Financial Management, Vol. 23, No.1, Spring, 98. Stewart III, G. Bennett, 1990, The Quest for Value, Harper Collins Publishers, USA. Young, D., 1999, “Some Reflections on Accounting Adjustments and Economic Value Added,” Journal of Financial Statement Analysis, Vol. 4, No. 2, Winter, 7-19.

17

ENDNOTES * Oral Erdogan,

Ph.D.,

is a staff member in the Istanbul Stock Exchange, and is a visiting lecturer in the Yildiz

Technical University and Marmara University ([email protected]); Niyazi Berk is a Professor of Accounting and Finance in Marmara University; Erol Katırcıoglu is a Professor of Economics in Marmara University. Comments from Noyan Arsan of the State University of West Georgia were very helpful. Vedat Akgiray, Nazim Engin, Burak Saltoglu, and conference participants at the 1999 METU International Conference on Economics III also provided useful comments. We also wish to thank the Editors and especially the referee whose comments and suggestions greatly improved the paper. We are responsible for any remaining errors. 1

See Shanken & Smith (1996) and Levonian (1994) for Economic Profit; See Jensen & Meckling (1976), and

Rappaport (1998) for Shareholder Value Added, SVA; See Stewart (1990), and Grant (1996, 1997), for Economic Value Added, EVA; see Young (1999) for the early studies on the residual income from 1920 to 1950s, and see Lawrence (1995) for the similarity of the EVA and the residual income. 2

Stewart (1990) uses the CAPM for measuring the cost of capital.

3

Although the Capital Asset Pricing Model (CAPM) of Sharpe (1964), Lintner (1965), and Mossin (1966) has

been supported with many researches so far, there have also been many opponents [See. Cheng & Grauer (1980), Fama and French (1992), Grinold (1993)]. Erdogan (1996) and Markowitz (1999) also dedicate it to Tobin (1958). 4

Cost of Capital (COC) in Equation 2 is tax adjusted through multiplying by (1-TC D/C); where TC is the

corporation tax rate, D is the book value of debt, and C is the equity . The value of T used in this study is 46 % for all companies. 5

For the theoretical explanations of profit concept in Accounting and Finance see Downie (1958); Stewart (1990,

68-117); Rappaport (1998), and Berk (1998). 6

For a comparison see Erdoğan (1996).

7

Bekaert and Harvey (1997) indicate that the volatility is an important factor for determining the cost of capital in

emerging markets. 8

In the empirical part of the study, as expected (average) net income is calculated ex-post from the historic

realized net incomes and adjusted to inflation rates, the growth rate is assumed to be included indirectly.

18

9

Moreover, the borrowing rate in the calculation of economic profit is subject to the foreign exchange fund

representing the values that have been permitted to use by the development banks. It is assumed for the simplicity that all companies have carried out all of their borrowings with the stated interest rates. 10

The related interest rates are 87.1%, 104.8%, 99%and 92% for the last month of 1993, 1994, 1996, and 1997

respectively. Since there was no treasury bill outstanding in December 1995, the monthly average of 1995 is calculated as the risk free rate of return. That is 105%. The values are available in the Web-site Central Bank of Republic of Turkey (www.tcmb.gov.tr). 11

We are grateful to Sınai Yatırım Bankası and Türkiye Sınai Kalkınma Bankası for providing the related data.

The investment credit rates at the beginning of 1994, 1995, 1996, 1997, and 1998 are 63.9%, 86.4 %, 81 %, 105 %, and 90 % respectively. 12

R square is the coefficient of multiple determination to identify what effect the independent variables have on

the dependent variable. t refers to tail significance for the coefficients; N shows the number of observations. 13

According to Chow’s (1960) test we apply for each successive year; we infer that there is a statistically

significant change, at 1% level, in the economic profit functions from 1993 to 1994 and from 1996 to 1997. For the other pair of regressions, we cannot reject the null hypothesis that both regressions propose the equal functions. For each pair: N=246; df=244. 14

Bekaert and Harvey (1997) indicate that the volatility is an important factor for determining the cost of capital

in emerging markets.

19

Graph 1: The Share of the Sample in Market Capitalization

100 90 80 70 60 50 40 30 20 10 0

72%

69%

65% 51%

52%

1997

1998

%

77%

1993

1994

1995

1996 Years

Note: Total market capitalization includes banks as well.

Graph 2: Public Internal Borrowing Rate for 1-Year Maturity

200

Interest R ates %

180 160 140 120 100 80 60 3 93 93 3 94 4 4 5 95 95 -95 -96 -96 -96 -96 -97 -97 -97 -98 t r n b y g n-9 y- g- v-9 r- n-9 p-9 n-9 r- ulc p n Ja Ma Au No Ma Ju Se Ja Ap J Oc Fe Ma Au De Ma Ju Se Ja Source: Central Bank of Republic of Turkey, CBRT, Ankara 1999.

20

Table 1: Main Items of Consolidated Balance Sheet and Capitalization of 123 Companies (Billions, Turkish Lira) Year

Total Assets Current Assets Fixed Assets Equity

Net Profit

Capitalization

1993

183,9

102,3

81,6

85,8

19,0

420,4

1994

407,3

218,7

188,5

188,2

40,3

605,2

1995

796,5

441,2

355,3

393,5

100,1

868,4

1996

1.468,8

816,9

651,9

669,0

165,9

2.115,4

1997

2.866,2

1.635,2

1.231,0

1.287,2

317,7

6.491,2

1998

4.937,1

2.668,3

2.247,6

2.328,2

405,0

5.487,3

Source: ISE Company Yearbooks 1994-1998.

Table 2: Characteristics of βs

Mean Standard Error Median Standard Dev. Kurtosis Skewness No. of Obs. Confidence Lev. (95%)

1,18 0,05 1,15 0,55 12,99 2,45 123 0,098

1,15 0,03 1,13 0,38 6,92 1,44 123 0,068

1,17 0,03 1,18 0,37 3,31 0,75 123 0,065

0,98 0,03 1,00 0,28 0,13 0,11 123 0,051

0,86 0,02 0,86 0,25 1,20 0,00 123 0,044

0,96 0,02 0,97 0,24 2,93 0,66 123 0,044

Note: Number of firms included 123. Responded market portfolio return is the change in the ISE 100. security i is the covariance of the security with the market, divided by the variance of the market. Cov [R i , R M ] . β = Var

β of

[R M ]

Table 3: The Models Model

Model

Cost of Capital

Model

(MVA/C)t=a+b(EP1/C)t+1

(RF + (E(Rm)- RF) βim)(1-TCD/(C-D))C

Model

(MVA/C)t=a+b(EP2/C)t+1

[we(RF + (E(Rm)- RF) βim)+wdRd]C

Model

(MVA/C)t=a+b(EP3/C)t+1

(d/P)C

Model

(MVA/C)t=a+b(EP4/C)t+1

(we(d/P)+wdRd)C

Model

(MVA/C)t=a+b(EP5/C)t+1

(we(d/P)+wdRd)C

*: In model 5, the leverage ratios are market value weighted.

21

Table 4: Regression Results for Linear Model between MVA/C and EP/C Model Year: 1993 Model 1 Model 2 Model 3 Model 4 Model 5

a

b

ta

tb

Adj.R2

F value

N

1,412820 1,313928 0,826676 0,880746 0,757522

0,277224 0,055458 0,604078 0,781327 0,810838

7,71* 7,85* 3,20* 4,39* 3,36*

1,15 0,22 2,26** 3,14* 3,21*

0,00 0,00 0,032 0,068 0,071

1,32 0,1 5,1** 9,9* 10,3*

123 123 123 123 123

Year: 1994 Model 1 Model 2 Model 3 Model 4 Model 5

0,652848 0,6407591 0,188026 0,748584 0,464131

-0,519959 -0,461604 1,533994 2,131048 2,504026

5,42* 5,35* 1,11 6,88* 4,61*

-2,28** -2,20** 3,13* 5,84* 7,18*

0,03 0,03 0,07 0,21 0,29

5,2** 4,8** 9,8* 34,1* 51,5*

123 123 123 123 123

Year: 1995 Model 1 Model 2 Model 3 Model 4 Model 5

0,061080 0,050443 -0,037207 0,197864 0,068403

-0,403781 -0,402085 0,657350 1,619388 1,617495

0,55 0,44 -0,25 2,19 0,77

-1,75 -1,74 1,80 4,67* 5,53*

0,02 0,02 0,02 0,15 0,20

3,1 3,0 3,3 21,8* 30,5*

123 123 123 123 123

Year: 1996 Model 1 Model 2 Model 3 Model 4 Model 5

0,293883 0,268463 0,066351 0,619042 0,318379

-0,44325 -0,426090 0,961082 1,912219 2,037988

1,97 1,66 0,28 5,07* 2,78*

-1,57 -1,48 1,88 5,04* 5,99*

0,01 0,01 0,02 0,17 0,22

2,5 2,2 3,5 25,4* 35,9*

123 123 123 123 123

Year: 1997 Model 1 Model 2 Model 3 Model 4 Model 5

0,988267 1,016540 -0,07995 3,326262 1,598727

-0,538672 -0,360459 6,817981 6,602580 8,854257

1,94 1,79 -0,15 5,98* 4,62*

-0,49 -0,32 3,53* 5,05* 6,15*

0,00 0,00 0,09 0,17 0,23

2,5 0,03 12,5* 25,5* 37,9*

123 123 123 123 123

Note: Dependent variable is MVA/C, Explanatory Variable is EVA/C or EP/C; a is the intercept, b is the slope coefficient; ta and tb denotes the t values. They are simply the ratios of the estimated coefficients to their standard errors. F test values show the statistically significance of the regression functions. N refers to the number of the observations. * : significant at the 0.01 level. **: significant at the 0.05 level.

Table 5: Results for Larger Firms Years 1993 1994 1995 1996 1997

A

0,424747 0,562102

0,145718 0,550572

1,574155

B 1,728295 2,925896 2,330958 2,100074 7,968062

ta 0,71 2,88* 0,90 2,24** 4,59*

tb 2,38** 4,84* 4,44* 3,20* 5,14*

R2 0,12 0,40 0,36 0,21 0,43

F 5,6* 23,5* 19,8* 10,3* 26,5*

N 35 35 35 35 35

Note: Model 5 is applied here. The total asset value of the sample is 75-76 % out of 123 firms. Dependent variable is MVA/C, Explanatory Variable is EP/C. * : significant at the 0.01 level. **: significant at the 0.05 level.

22

Table 6: Results for Companies with Higher Offering Rates Year (MVA) 1993 1994 1995 1996 1997

a 0,7360263 0,3191521

-0,016987

b 0,659272 1,924181 1,2383764

0,1508509

1,4651409

0,6976725

2,9020326

ta 2,99* 3,32* -0,17 1,45 5,42*

* : significant at the 0.01 level. **: significant at the 0.05 level.

23

tb 2,70* 5,98* 3,91* 4,42* 5,68*

R2 0,09 0,34 0,18 0,22 0,32

F 7,3* 35,7* 15,3* 19,5* 32,3*

N 68 68 68 68 68