Negative EVA and value: a paradox?

Author:Fu, Richard

    Numerous studies have examined the relationship between stock returns and EVA or MVA metrics, both popularized by Stern Stewart & Company. Many previous studies by Abate, Grant and Stewart (2004), Grant (1996, 1997, and 2003), and Biddle, Bowen, and Wallace (1997) present a theoretical and empirical relationship between EVA and stock returns. Biddle et al. document that earnings have a stronger correlation to stock returns than EVA to returns. Grant (1996) find a strong relationship between MVA to EVA for the top fifty MVA firms in the Stern Stewart 1993 data, but the relationship is much weaker for the entire sample. Abate et al. (2004) also establish that forty of the top fifty MVA firms show positive EVA values when analyzing the Stern Stewart 2000 data set. However, ten of the top fifty MVA firms display negative EVA whereby they suggest that this "seemingly anomalous finding" can be explained by the market's confidence of the firms' ability to create positive EVA values from future investment opportunities. Two other studies utilize MVA or EVA to create portfolio investment strategies and also document similar anomalies. Yook and McCabe (2001) create portfolios using MVA and find that low MVA per share lead to higher average portfolio returns. Zaima (2008) examines returns generated from portfolios formed in deciles (P1 to P10) using EVA measures, and shows that the EVA to portfolio returns relationship is U-shaped rather than linear where the most negative EVA (P1) and the highest positive EVA (P10) portfolios display the greatest returns. Other portfolio returns (P2 through P9) fall significantly below P1 and P10. Despite numerous studies on EVA and stock returns, the results show that a straightforward EVA to value relationship is difficult to establish.

    Additional studies by O'Byrne (1999) and Grant (1996, 1997, 2003) show that negative EVA and positive EVA firms reveal different relationships to firm value. Grant (1997) suggests that firms with negative EVA include a mix of troubled firms that may be on its way to recovery versus ones that may not, thereby, adding significant "managerial noise". Taken together, these results indicate that the capital markets' expectations of firms, particularly those with negative EVA, require further investigation. By comparing the extreme deciles of firms with the highest EVA (P10) and those with the most negative EVA (P1), this study attempts to shed some light on the market's return predictability relative to the EVA measure. Hence, we are able to address a general question: What is the market paying for with negative EVA firms?

    To answer the question, we adopt a methodology that is different from those applied by past studies. Numerous studies have directly examined relationships between EVA to market value-added (MVA) or between EVA to stock returns. In contrast, we investigate the underlying asset pricing structure that drives the stock return with regard to EVA measure using Abate et al.'s (2004) argument. They suggest that firm value is comprised of "the EVA value (created) from current assets plus the EVA value generated by future opportunities". [Abate et al. 2004, p. 64] Hence, we investigate the relationship between stock returns to variables that proxy for "EVA value generated by future investment opportunities" instead of directly using current EVA. By excluding (current) EVA from the stock return relationship we circumvent the paradoxical relationship of negative EVA to stock value. Our method enables us to determine how subsequent stock returns relate to the expectation of EVA growth for firms with current negative EVA. Indeed, as advocated by Abate et al. (2004), we document that the stock returns of the two extreme EVA portfolios increase with the proxy of EVA growth opportunities; thereby providing evidence for the importance of EVA growth opportunities when analyzing stock valuations. Moreover, our results show that the return premium with regard to EVA growth is systematically higher for the P1 firms than P10 firms. Finally, the results are also consistent with Grant's conjecture that negative EVA firms are riskier due to "managerial noise", making it difficult to separate firms that are able to overcome its current negative EVA versus those that cannot.


    Abate et al. (2004) provide a thorough development of the relationship between EVA and firm value, which presents an insightful framework for investment strategies based on EVA/ME metrics. Using their notations EVA is defined as:

    EVA = NOPAT - WACC x C (1)


    NOPAT is the net operating profits after taxes or (1-tax rate)EBIT (earnings before interest & taxes); WACC is the weighted average cost of capital of debt and equity; and C is the total net invested capital including debt and equity.

    Abate et al. (2004) further show that EVA can be expressed as the EVA spread or the difference between the return on capital invested (ROC) and the cost of capital (WACC) multiplied by the capital invested (C), or:

    EVA = (ROC - WACC) x C (2)

    They further show that the market value of the firm, V, divided by capital, C, equals:

    (V/C) = 1 + ((EVA/WACC)/C) (3)

    = 1 + ((ROC - WACC)/WACC (4)

    = ROC/WACC (5)

    The equation shows that if a firm's value to capital (V/C) is higher than one, then it is a wealth creating firm where it engages in positive EVA (or NPV) projects. It also suggests that the EVA spread (ROC--WACC) must be positive for a wealth creating firm. In contrast, a wealth destroying firm has a value-to-capital (V/C) lower than one, which indicates that it engages in negative EVA (NPV) projects. Therefore, equation (5) measures the market "value-based price-to-book conditions for positive EVA or wealth-creating companies", implying that value-based price to-book ratio is positively related to EVA [Abate et al., 2004, p. 64].

    Equation (5), however, is inconsistent with the empirical evidence documented by numerous studies that show some negative EVA firms exhibit significant positive stock returns (see, for example, Abate et al. (2004), Grant (1996, 1997, and 2003), O'Byrne (1999), Yook and McCabe (2001), and Zaima (2008)). Therein lies the paradox. How can we reconcile the findings of subsequent high positive returns associated with negative EVA firms?

    Abate et al.(2004) is the first to argue that incorporating the effect of EVA generated by future growth opportunities may reconcile the apparent inconsistency between theory and empirical evidence. In their sample of 50 top MVA firms, 40 out of which have positive EVA, the remaining 10 firms have negative EVA measures. They argue that the firm value is comprised of "the EVA value (created) from current assets plus the EVA value generated by future investment opportunities" and that the market is willing to pay for the EVA value generated by future opportunities (Abate et al. 2004, p. 64). The bottom 50 MVA firms display a more consistent relationship where 46 out of 50 firms display negative EVA. In contrast, Yook and McCabe and Zaima find that the extreme negative EVA (or MVA) firms exhibit relatively high ex ante returns when forming portfolios with EVA(or MVA).

    In this paper, we investigate the role played by EVA of future opportunities in determining the future stock returns. Particularly, we focus on the stocks with the most negative EVA, which have subsequently generated high returns, as documented by Zaima (2008).

    To measure the EVA generated with future opportunities, we focus on the market value of equity because stockholders are residual claimants to the firm cash flows, and we assert that potential EVA from future growth of the firm will mostly...

To continue reading