Operating income, residual income and EVA: which metric is more value relevant--a comment.

This comment concerns two main issues that have been presented and discussed by Chen and Dodd (2001). Firstly, epistemologically, economic value added (EVA) is of dubious worth because it is construct deficient in an efficient market hypothesis (EMH) world and also in a non-efficient market hypothesis (non-EMH) world. This has important implications for valuations and value creation. Moreover, the transmission mechanism of EVA into market value added (MVA) is problematic. Secondly, contrary to the importance that Chen and Dodd (2001) ascribe to the relevance of accounting information, there is a growing body of research that shows that accounting information is becoming progressively less relevant to stock returns and stock price changes, and this trend is expected to continue.

THE EMH, EPISTEMICS AND EVA

Epistemologically, EVA is a non sequitur because it is construct deficient in both EMH and non-EMH-worlds. Furthermore, the fundamental economic and financial factors that constitute the main drivers of EVA have been shown to have only a minor impact on changes in share price, and hence share returns.

EVA in an EMH-World--a Financial Fiction

In an EMH-world, where assets plot on the Security Market Line (SML) or Capital Market Line (CML), and asset prices (market prices) correspond to asset values (intrinsic values), it is not possible to meaningfully talk about a measure such as EVA. On the SML and CML, by definition net present value (NPV) equals zero and the required rate of return (PER) or cost of capital equals the internal rate of return (IRR), in which case EVA must equal zero. Since EVA measures the difference between RRR and IRR, in an EMH-world EVA is attempting to measure a quantum that by definition cannot exist, except perhaps as noise. Arbitrage and competitive action ensure that abnormal profit cannot consistently occur. If the phenomenon of EVA were to be observed, its occurrence would be random, statistically non-significant, would not be serially correlated and, on the average, positive-EVA would be offset by negative-EVA. In an EMH-world, it is not possible to consistently earn excess returns (i.e., abnormal or super-profits) e xcept at the price of higher risk, measured by the beta coefficient. Thus, within the logic of the EMH, EVA is a fiction.

In an EMH-world, EVA would be impounded in an unbiased way into stock prices so that prices corresponded to intrinsic value. In other words, the market would recognize the impact of EVA and would impound it into stock price. The increase in stock price would create market value and is known as MVA. If EVA, an internal metric of performance, is not transmitted into MVA, an external metric of performance, shareholders do not benefit from improvements in EVA as far as asset pricing is concerned. The question of the transmission of EVA into MVA is an area that requires additional research. Since EVA is concerned with economic and financial fundamentals, it is important to know to what extent changes in EVA will move stock price. The question of what moves stock prices has been investigated for a considerable period of time, and the findings suggest that financial and economic fundamentals play only a minor role in the changes in asset prices (Cutler et al,, 1989; Haugen et al., 1991; Shiller, 1981; Roll, 1984, 19 88; Frankel and Meese, 1989).

EVA in a Non-EMH-World

There are three issues I would like to raise with regards to EVA in a nonEMH-world, namely, the validity of using the capital asset pricing model (CAPM) and beta, the matter of what moves stock prices and hence MVA, and the matter of dividends and earnings in relation to stock price.

The validity of using the CAPM. In a non-EMH world, the validity of using the CAPM as the basis for the calculation of EVA is questionable because the CAPM is derived from the EMH and is dependant upon the existence and functioning of the EMH. In a non-EMH world, the CAPM and beta should not be used to calculate the cost of equity that forms part of the weighted average cost of capital (WACC). Numerous studies over more than a decade have shown that using the GAPM and beta is an undesirable way of calculating the cost of capital and should not be used for valuation purposes, and this sentiment has been very clearly and unambiguously stated by Fama and French (1992, 1996). Thus, an important part of the basis for the calculation of EVA, namely CAPM, has been rejected because of the poor relationship between the cross section of returns and the systematic risk coefficient, beta.

The uninspiring performance of traditional asset pricing models such as the CAPM and arbitrage pricing theory (APT) has given an impetus to behavioral finance because the traditional asset pricing models require that all of the predictable patterns in asset returns, both short and long run, be traceable to differences in loadings in economically meaningful risk factors. There is little evidence to show that this can be done (Hong and Stein, 1999).

What Drives Share Price? Extensive empirical research has shown that an equilibrium pricing theory, such as the EMH, is not a satisfactory descriptor of the real world. Market price seldom corresponds to intrinsic value and this disequilibrium can continue for extensive periods of time. Moreover, whereas economic and financial fundamentals will affect value, they are not the main movers of stock prices. In this regard, Fama (1981) found that a substantial fraction of return variation cannot be explained by macroeconomic news. Roll (1984) found that news about weather conditions, the principal source of variation in the price of orange juice, explains only about 10% of the movement in orange juice futures prices. Roll (1988) further found that it is difficult to account for more than one-third of the monthly variation in individual stock returns on the basis of systematic economic influences. When investigating which factors moved share price, Cutler et al. (1989) found that macroeconomic news explains only ab out one-fifth of the movement in stock prices, and they state: "The view that movement in stock prices reflect something other than news about fundamental values is consistent with evidence on the correlates of ex post returns (1989: 9). Haugen et al. (1991) established that the main driver of stock returns was changes in volatility, and that fundamental economic and financial factors were not the main drivers of changes in volatility. In fact, they found that as few as one-quarter of the volatility shifts are associated with the release of significant (financial and economic) information.

For EVA to deliver the beef in terms of stock returns, market price must recognize in an unbiased way the value creation process that derives from implementing NPV positive (i.e., EVA positive) investment, financing, and dividend decisions. Empirical evidence shows that notable drivers of price include volatility (Shiller, 1989, 2000; Cutler et at., 1989; Haugen et at., 1991), momentum (Chan et al., 1996; Moskowitz and Grinblatt, 1999; Hong and Stein, 1999; Hong et al., 2000; Lee and Swaminathan, 2000) and financial herding (Graham, 1999).

Research into volatility itself has stimulated research into momentum and financial herding. Grinblatt et al. (1995), and Wermers (1999) came to the conclusion that a large part of herding behaviour occurs when investors "momentum-follow," and Nofsinger and Sias (1999) found evidence that implicates the use of momentum strategies by growth-oriented funds as an important source of herding. What is worthy of note is that momentum and herding have a notable impact on market price that is not related to economic or financial fundamentals.

When volatility, herding and momentum are substantially present, as so often seems to be the case, and when these phenomena have a much greater impact on stock prices and stock returns than fundamental economic and financial factors, a number of problems emerge for EVA. Firstly, market values will not correspond to intrinsic values because pricing is being driven by non-fundamental factors. Secondly, the use of WACC, based as it is on market values and weights, is a highly questionable practice, because of the manic-depressive nature of market prices (Hagstrom, 1995; Loewenstein, 1995). In this regard, consider the excesses of the Nikkei-Dow in the 1980s and early 1990s in comparison with its levels in 2001. Consider also the excesses of the NASDAQ in 1999 and early 2000 in comparison with its levels in 2001, and the gyrations of US blue chip stocks as reflected in the DJIA in recent times.

In short, without a reliable CAPM and without a reliable WACC, how can EVA be determined and implemented and associated with changes in stock returns via MVA?

Dividends and Earnings. Fama and French (2001) report that the proportion of firms paying cash dividends has declined from 66.5% in 1978 to 20.8% in 1999, and that the evidence shows this to be part of a long-term process. For example, in every year from 1943 to 1962, more than 82% of NYSE firms paid dividends; in 1951 and 1952 more than 90% paid dividends, but with the addition of AMEX in 1963 the proportion of dividend payers declined to 69.3%. This decline continued with the addition of NASDAQ in 1973 (Fama and French, 2001). Clearly, unprofitable firms and those with low earnings cannot pay dividends, Investors have become more willing to hold the shares of small, relatively unprofitable growth companies (Fama and French, 2001). As dividends diminish, stocks returns are being driven increasingly by swings in capital values.