You are on page 1of 38

The Relation between Expected Returns, Realized Returns, and Firm Risk Characteristics

*
CHRISTINE A. BOTOSAN, University of Utah MARLENE A. PLUMLEE, University of Utah HE WEN, University of Utah 1. Introduction Existing literature employs two general approaches to assess the validity of alternative proxies for firm-specific cost of equity capital or expected return (hereafter Et)1(rt)). The first approach involves examining the association between the proxy for Et)1(rt) and future realized returns. The second approach focuses on the association between the Et)1(rt) proxy and contemporaneous risk characteristics of firms. The results of these two streams of literature are mixed. Easton and Monahan (2005) (hereafter EM) and Guay, Kothari, and Shu (2005) (hereafter GKS) focus on the association between alternative proxies for Et)1(rt) and future realized returns and conclude that none of the proxies they examine provide valid estimates of the construct of interest. In contrast, Botosan and Plumlee (2005) (hereafter BP) conclude that two common proxies for Et)1(rt) — rDIV (Botosan and Plumlee 2002) and rPEG (Easton 2004) — are valid, based on their finding that both are associated with firm-specific risk characteristics in a theoretically predictable and stable manner. Furthermore, Pastor, Sinha, and Swaminathan (2008) document a positive association between market-level implied cost of capital and risk as measured by the volatility of market returns, consistent with the estimates capturing time-varying Et)1(rt). In this paper, our goal is to reconcile the conflict between these two streams of literature and provide additional evidence pertaining to the construct validity of the proxies employed in extant research. Contrary to the results documented in EM and GKS, we document a positive association between ten of the twelve Et)1(rt) proxies included in our study and future realized returns after controlling for new information.1 We reconcile our findings to those in EM and GKS by demonstrating that the prior results are due to empirical misspecification. Finally, we show that two of the proxies, rDIV and rPEG, demonstrate not only the expected relation with future realized returns, but also with firm-specific risk. We also address several other issues regarding the use of implied cost of capital estimates including: (1) analysts’ forecast bias, (2) the efficacy of realized returns for Et)1(rt) before and after controlling for news, (3) the effectiveness of averaging several Et)1(rt) proxies, and (4) the substitution of realized values for analysts’ forecasts of cash flows or earnings. Our evidence suggests that deviations between analysts’ expectations and those of the market lead to potentially less powerful proxies but do not generate biased or
* Accepted by Steven Salterio. We gratefully acknowledge the financial support of the David Eccles School of Business. We also wish to thank Kin Lo, K. Ramesh, Matt Magilke, and the workshop participants at the London School of Business, Michigan State University, Rotterdam School of Management – Erasmus University, the University of North Carolina, the University of British Columbia, Arizona State University, Georgetown University, and University of Akron for helpful comments on previous drafts of the paper. 1. Of the twelve estimates we examine, nine are implied cost of capital proxies, two are popular averages of subsets of these proxies, and the final estimate is derived from the Fama-French four factor model.

Contemporary Accounting Research Vol. 28 No. 4 (Winter 2011) pp. 1085–1122 Ó CAAA doi:10.1111/j.1911-3846.2011.01096.x

1086

Contemporary Accounting Research

inconsistent results. Furthermore, we find that realized returns do not proxy for Et)1(rt) even after controlling for news, and that averaging several proxies does not yield an enhanced metric. Finally, substituting realized values for analysts’ forecasts of cash flows yields systematically biased estimates, which might yield biased and inconsistent results when such estimates are employed in empirical research. Given the current state of the literature, the validity of the various cost of capital estimates is unclear and it is not uncommon for similar studies to document dissimilar results because they employ different cost of capital estimates. For example, Ogneva, Subramanyam, and Raghunandan (2007) (hereafter OSR) conclude that firms with internal control weaknesses do not bear a higher cost of equity capital, while Ashbaugh-Skaife, Collins, Kinney, and LaFond (2009) (hereafter ACKL) conclude the opposite. These contradictory results are wholly attributable to differences in the authors’ choices of Et)1(rt) proxies.2 Thus, additional evidence regarding the validity of alternative Et)1(rt) proxies is needed to help guide researchers’ proxy selection. Based on our evidence, we recommend that researchers requiring a valid proxy for Et)1(rt) employ either rDIV or rPEG. We caution against the use of realized returns or the other implied cost of capital estimates we examine to proxy for Et)1(rt). Finally, we suggest that researchers introducing new proxies for Et)1(rt) to the literature subject their proposed measures to both of the construct validity tests employed in the current study, and provide support for how the measure enhances the existing technology for estimating Et)1(rt). Our findings should be of interest to researchers requiring a valid proxy for Et)1(rt). The need for such a proxy is far-reaching. It extends from accounting research that examines the impact of financial reporting and disclosure on required returns, to finance research that investigates market anomalies, asset pricing theory, and capital budgeting. These areas of research produce findings of interest to standard setters, regulators, investors, preparers, and auditors. Accordingly, the validity of the Et)1(rt) proxies employed in research is an important issue with pervasive implications. We organize the remainder of our paper as follows. Section 2 reviews the related literature and sets forth the development of our hypotheses. Section 3 presents our empirical models and proxies. Section 4 delineates our sample selection procedure and provides descriptive statistics for our sample. We discuss the results of our construct validity assessment in section 5. In section 6 we provide evidence pertaining to several other issues related to implied cost of capital estimates. Section 7 reconciles the results of our realized return analysis with those documented in prior literature. Finally, section 8 summarizes our conclusions. 2. Literature review and hypotheses development
Construct validity is woven into the theoretical fabric of the social sciences, and is thus central to the measurement of abstract theoretical concepts.. .. Fundamentally, construct validation is concerned with the extent to which a particular measure relates to other measures consistent with theoretically derived hypotheses concerning the concepts (or constructs) being measured.

2.

OSR replicate ACKL’s findings using ACKL’s proxy for cost of equity capital, but discount the results, citing conflicting evidence regarding the relative validity of alternative implied cost of capital measures. Our results provide little support for the Et)1(rt) proxy employed in OSR, but strong support for the proxy employed in ACKL.

CAR Vol. 28 No. 4 (Winter 2011)

Expected Returns, Realized Returns, and Firm Risk Characteristics

1087

The preceding quote describes the classic, well-accepted method routinely employed to examine construct validity.3 Consistent with standard practice, we rely on the preceding quote to guide our approach to examining construct validity. Specifically, we examine the relationship among alternative proxies for Et)1(rt) and future realized returns, which are a function of Et)1(rt). In addition, we examine the relationship among alternative proxies for Et)1(rt) and contemporaneous risk characteristics finance theory suggests are predictably associated with Et)1(rt). The following paragraphs detail the theories underlying our examination. Realized returns Realized return at time t (rREALt) is modeled as the expected return at time t conditional on information available at t ) 1 (Et)1(rt)) plus the unexpected (or abnormal) return due to new information (URt): rREAL;t ¼ EtÀ1 ðrt Þ þ URt ð1Þ:

Relying on prior research (e.g., Campbell 1991;Vuolteenaho 2002) further decomposes the unexpected return to new information (URt) into two components — the unexpected return due to cash flow news (Ncf,t), and the unexpected return due to expected return news (Nr,t). This gives rise to equation 2: rREAL;t ¼ EtÀ1 ðrt Þ þ ðNcf ;t À Nr ;t Þ where: rREAL,t = realized return from t)1 to t; Et)1(rt) = expected return at t, conditional on information at t)1; Ncf,t = return due to cash flow news from t)1 to t; and Nr,t = return due to expected return news from t)1 to t. Traditionally, research that employs realized returns to proxy for expected returns relies on the assumption that URt is mean zero, and that in-sample averaging of realized returns across firms or time purges URt to produce a valid proxy for Et)1(rt). Some more recent research goes further by using firm-specific (i.e., not averaged) realized returns to proxy for Et)1(rt) (e.g., Easley, Hvidkjaer, and O’Hara 2002; McInnis 2010). Nevertheless, a growing body of research questions the validity of realized returns as a proxy for Et)1(rt). Elton (1999) argues that averaging does not eliminate URt because unexpected returns tend to be large and correlated across firms and time. Vuolteenaho (2002) demonstrates that the unexpected component is the dominant factor driving firmlevel stock returns, and that cash flow news is largely firm-specific, whereas expected return news is linked to systematic macroeconomic factors. Consistent with the latter finding, Campbell and Ammer (1993) find that expected return news drives aggregate stock returns. These findings suggest that firm-level and portfolio-level realized returns could be poor proxies for Et)1(rt). At the firm level the URt due to firm-specific cash flow news, as well as the URt due to systematic expected return news, contaminate the realized return proxy. At the portfolio level sufficient averaging might mitigate the URt due to firm-specific cash flow news, but it is less likely to mitigate the URt due to systematic, macroeconomic expected return news. Moreover, if cash flow news is correlated across firms and ⁄ or time, averaging, even over large numbers of firms or long periods, might not purge either
3. Carmines and Zellner (CZ) 1979: 23; bolding added. Because theory rarely models abstract theoretical concepts completely, construct validation does not require the identification of an exhaustive list of the observable measures believed to be associated with the underlying unobservable construct. On the contrary, a theoretical basis for hypothesizing a directional association between the empirical proxy of interest and some set of measures associated with the underlying unobservable construct is paramount to credible construct validation.

ð2Þ;

CAR Vol. 28 No. 4 (Winter 2011)

we should observe a positive correlation between the proxy and rREAL. but a good (poor) predictor of rREAL. In this case.tÀ1 þ k ðE ðr Þ À rf . book-to-price ratio and growth. Kettler. no statistically significant correlation between rREAL.t after controlling for Ncf. Risk characteristics Ross’s 1976 Arbitrage Pricing Theorem (APT) models the expected return for a given period as a function of the risk free rate (rf) plus the risk premiums arising from K risk factors: XK EtÀ1 ðrt Þ ¼ rf . 5. This gives rise to our first hypothesis. Equation 3 above and this body of research gives rise to our second hypothesis. realized returns.e. rREAL. Sharpe 1964). Beaver. Mossin 1966. leading to a positive association between Et)1(rt) and growth.t and Nr. omitting the latter two variables reduces the power of the analysis. and a negative correlation with market value of equity provides support for the validity of that Et)1(rt) proxy. Our paper seeks to provide guidance to researchers seeking a valid proxy for Et)1(rt) by examining the extent to which implied cost of capital estimates proxy for expected returns (i.t.4 HYPOTHESIS 1. (2) and the research discussed in the preceding paragraph suggest that an examination of the association between rREAL. Berk (1995) argues that market value of equity is systematically decreasing in priced risk such that Et)1(rt) is inversely related to the market value of equity and positively related to the book-to-price ratio.t and Et)1(rt) might be observed even if one exists. The capital asset pricing model (CAPM) suggests that Et)1(rt) is increasing in market beta (Lintner 1965. and Scholes (1970) assert that abnormal earnings arising from growth opportunities are inherently more risky. 28 No. leverage. although existing research suggests several candidates. Modigliani and Miller (1958) support a positive association between leverage and Et)1(rt).t or Nr. Even if Et)1(rt) is not correlated with Ncf. market beta. Readers interested in a more in-depth discussion of these risk characteristics are referred to Botosan and Plumlee 2005.t).. (2) suggests that if an Et)1(rt) estimate is a valid proxy.t provides support for the validity of that Et)1(rt) proxy. a positive correlation between a proxy for Et)1(rt) and rREAL. Nonetheless. Finally.t and Nr. These roles for implied cost of capital estimates are quite different. After controlling for Ncf. If Et)1(rt) is correlated with Ncf.t the resulting correlated omitted variable bias could result in biased and inconsistent results. the 4. CAR Vol. In the presence of cash flow news and expected return news. and firm-specific risk characteristics. Et)1(rt)).e. as noted earlier.t and Nr. Lee.t are ignored. and Wang (2010) provides guidance to those seeking to predict future returns by investigating the ability of certain implied cost of capital proxies to predict future realized returns (i.t.1088 Contemporary Accounting Research component of the URt.5 HYPOTHESIS 2. however. So.t or Nr.t.t.. Related empirical research Prior empirical research examines the association between alternative proxies for Et)1(rt). Attempts to mitigate the problem by averaging over increasingly larger samples or longer periods invoke unpalatable stationarity assumptions (Chan and Lakonishok 1993).t and Et)1(rt) is vulnerable to omitted variables if Ncf.tÀ1 Þ: ð3Þ k¼1 k t À1 k Ross’s APT does not identify the risk factors. it is quite possible that a particular proxy might be a poor (good) proxy for Et)1(rt). A positive correlation between a proxy for Et)1(rt) and the risk free rate. 4 (Winter 2011) .

The conflict with respect to rDIV is implied as opposed to explicit since neither GKS nor EM include rDIV in their analyses. EM find that none of the proxies are positively associated with realized returns.t.t.t) and expected return news (Nr. 7. In contrast. Accordingly. 6. EWER_Nit= economy-wide expected return news during the 12-month realized return period. The Appendix summarizes the Et)1(rt) proxies examined in GKS. While GKS. CFN_TVit = news incorporated in target prices during the 12-month realized return period. = 12-month buy and hold return. as well as cash flow news (Ncf. to use both the realized return and risk-based approaches to examine the construct validity of the more recent additions to the set of alternative proxies for Et)1(rt). without a rigorous construct validity assessment of the same. CAR Vol. GKS regress realized returns on five alternative proxies for Et)1(rt). to augment BP’s risk-based construct validity analysis of rDIV with a realized return analysis and extend their findings over an additional period of time. = expected return proxy at t conditional on information at time t)1. EM. EM control for Ncf. and FSER_Nit = firm-specific expected return news during the 12-month realized return period. 3. and Firm Risk Characteristics 1089 findings from this research are mixed. a positive correlation between a given proxy for Et)1(rt) and rREAL. rREALit ¼ a0 þ b1 ERit À1 þ b2 CFN 1it þ b3 CFN TVit þ b4 EWER Nit þ b5 FSER Nit þ eit where: rREALit ERit)1 CFN_1it ð4Þ. the primary objectives of the construct validity portion of this study are threefold: first.t in their examination of the association between realized returns and seven proxies for Et)1(rt). and third. Accordingly.t and Nr. are valid proxies for Et)1(rt) as both are correlated with firm risk characteristics in a theoretically predictable manner. and BP examine different sets of Et)1(rt) proxies. rPEG is examined by all three. beginning the month after estimation of ER. They conclude that rDIV and rPEG. second. Realized Returns.7 Hypothesis 1 states that.t or Nr. since 2005 there has been an explosion in Et)1(rt) proxies employed in the literature. a positive ß1 coefficient) provides support for the validity of that Et)1(rt) proxy. 28 No. EM. Even so.6 The authors do not document the expected positive association between realized returns and their proxies. In addition. after controlling for Ncf. the findings of GKS and EM versus those of BP present an explicit conflict in the evidence regarding this metric’s construct validity.t) from t)1 to t as a function of expected returns at t conditional on information at time t)1 (Et)1(rt)). = news about period t to t+1 cash flows received during the 12-month realized return period.t and Nr.t (i. BP.. 4 (Winter 2011) . and the current study.t) received from t)1 to t. to investigate the source of the disparate results noted above.Expected Returns.e. In our specification ERit)1 is one of a number of alternative expected return estimates intended to proxy for Et)1(rt). GKS and EM conclude that their results do not provide support for the construct validity of the proxies they examine. We enumerate the expected return estimates included in our study in the Appendix and in section 3. Empirical models and proxies Empirical method for realized return analysis Realized return model Our empirical specification of the return decomposition model (2) is given below. BP examine the association among five proxies for Et)1(rt) and firm-specific risk characteristics. but their analysis does not control for Ncf. Recall that equation 2 models realized returns (rREAL.

Our empirical specification includes two proxies for Ncf. Accordingly. Our second proxy. Under such circumstances.e. ß1 > 0) might be valid for use in empirical research even if the magnitude of the proxy is biased (ß1 „ 1). We include this variable to capture cash flow news related to long-horizon cash flows.8 Since target prices are increasing in future cash flows but decreasing in the discount rate. The following paragraphs provide details of the measurement of all variables included in (4) except for the Et)1(rt) proxies. We compute CFN_TV as the difference between the midpoint of the forecasted target price range made 12 months after we estimate our expected return proxies and the midpoint of the forecasted target price range made at the date we estimate 8. As shown in (2). Target prices. expected return news is negatively associated with realized returns. a test of ß1 = 1 is an unnecessarily rigorous test of construct validity. we include the change in the risk free rate between t)1 and t to proxy for economy-wide expected return news linked to macroeconomic factors (EWER_Nt). less analysts’ forecasts of those annual earnings issued the day we estimate Et)1(rt). CFN_1t captures cash flow news related to near-term cash flows. CFN_TVt. Our proxy for the current year cash flow news (CFN_1) is the difference between the reported annual earnings per share for year t announced during the 12-month period we estimate realized returns. Accordingly. analysts’ revisions of target prices reflect changing beliefs about discount rates as well as future cash flows..t. note 1). Most empirical research employing Et)1(rt) proxies is concerned with cross-sectional variation in Et)1(rt). CAR Vol.1090 Contemporary Accounting Research In theory ß1 should equal 1. As noted above. or both. For this reason. As both types of news are important to the theoretical model. Cash flow news. Since expected returns are a function of the risk free rate. Thus. reflect the present value of infinite horizon cash flows beyond the forecast horizon. To calculate our cash flow news proxies we rely on Value Line analysts’ forecasts of annual earnings per share for the current year (Year t) and target prices at the end of Year t+4. CFN_1 captures ‘‘earnings surprise’’ similar to that employed in numerous other studies. Since expected returns are also a function of the amount of risk a particular firm’s stock presents. CFN_TVt is a particularly important control variable. All forecasts are made in the third calendar quarter of the year. we include the change in firm-specific market beta between t)1 and t to proxy for expected return news linked to changes in the amount of risk associated with the firm (FSER_Nt). Since realized returns are increasing in cash flow news (see (2)). Lambert (2009) highlights that target prices are ‘‘free to reflect whatever assumption regarding future discount rates are deemed appropriate’’ (261. our primary objective for including CFN_TVt in our empirical model is to capture long-horizon cash flow news. an Et)1(rt) proxy that captures cross-sectional variation in Et)1(rt) (i. We calculate realized returns using CRSP data as the buy and hold realized return computed over the 12 months beginning the month after we estimate expected returns. we expect the association between CFN_TVt and realized returns to be positive regardless of whether CFN_TVt captures cash flow news. in testing our first hypothesis we do not require ß1 to meet the more stringent test of equivalence to the theoretical value of one. we expect ß2 and ß3 to be significantly positive. and is measured as the earnings surprise during the realized return period. 4 (Winter 2011) .t. but we report the results of this test. Such a test is not only a test of the extent to which the proxy captures cross-sectional variation in Et)1(rt). we expect both ß4 and ß5 to be significantly negative. The Et)1(rt) proxies are outlined following our discussion of the empirical model we employ in our risk analysis. Realized returns. 28 No. Our model also incorporates two proxies for Nr. which we employ in the computation of CFN_TVt. but also the extent to which it captures the magnitude of Et)1(rt). expected return news. Nevertheless. Consistent with this. is the revision in analysts’ forecasts of target prices during the realized return period.

which is stock price as of the publication date or within three days after publication. finding c1. In our analyses. leverage. we expect ERit)1 to be increasing in the risk-free rate. stating that the risk-free rate has ‘‘nothing to do with risk’’ and claiming that a change in rf is a cross-sectional constant. UBETAit)1 = unlevered CAPM beta. rf is an economywide parameter that bears directly on Et)1(rt). Empirical method for risk analysis Expected return model (5) provides our empirical specification of the expected return model given by equation 3: ERit À1 ¼ v0 þ c1 rftÀ1 þ c2 UBETAit À1 þ c3 DMitÀ1 þ c4 LMKVLit À1 þ c5 LBPit À1 þ c6 EXGRWitÀ1 þ git À1 where: ERit)1 = expected return proxy.S. Our proxy for firm-specific expected return news (FSER_N) is calculated as the change in market beta (MBETA) over the 12-month realized return period.9 We measure rf as the five-year treasury constant maturity rate obtained from the U. Thus. CAR Vol. In so doing. c2. Each beta estimation period ends in June. which also captures firm-specific expected return news. 10. Monahan and Easton (2010) question the use of changes in MBETA to capture firm-specific expected return news. EXGRWit)1 = expected growth in earnings per share. rft)1 = risk-free rate of interest. from Value Line. Realized Returns. we measure the change in rf using the five-year treasury constant maturity rate as of the month the Et)1(rt) proxy is estimated. LMKVLit)1 = log of the market value of common equity. but the potential detrimental effect of this is mitigated by the inclusion in our model of the change in expected terminal value. Monahan and Easton (2010) question the use of this proxy. and c6 greater than zero and c4 less than zero provides support for a given 9.10 We use CRSP data to estimate MBETA via the market model with a minimum of 20 out of 60 monthly returns and a market index return equal to the value weighted NYSE ⁄ AMEX return. Specifically. we do not presume that market risk is the only relevant priced risk. market beta. market beta.gov. EWER_N is the rf during the last month of the realized return estimation period less rf twelve months earlier. book-to-price. From (3). and Firm Risk Characteristics 1091 our Et)1(rt) proxies. but we do assume that changes in MBETA are correlated with changes in the overall level of risk an investment presents to the market. In addition. Federal Reserve at http: ⁄ ⁄ www.federalreserve. bookto-price. our construct of interest. c3. Panel A of Table 1 provides a description of the cash flow news proxies employed in our analysis. We compute our proxy for economy-wide expected return news (EWER_N) as the change in the risk free rate (rf) over the 12-month realized return period. and growth and decreasing in the market value of equity. LBPit)1 = log of book-to-market ratio. we scale our cash flow news variables by stock price on the day we estimate expected returns. Hypothesis 2 states that a theoretically predictable relation between a given proxy for Et)1(rt) and the risk-free rate of interest. Specifically. 4 (Winter 2011) . rf is not a cross-sectional constant because. Accordingly. We employ the change in MBETA to proxy for firm-level changes in the amount of risk. a change in rf gives rise to a change in Et)1(rt). Expected return news. leverage. ð5Þ. market value of equity. 28 No. We obtain actual earnings per share from COMPUSTAT and forecast and stock price data. DMit)1 = leverage. Panel B of Table 1 provides a description of the expected return news proxies employed in our analysis. such that a change in rf constitutes economy-wide expected return news.Expected Returns. as stated in Table 1. The power of this proxy is reduced if this assumption is violated. and earnings growth provides support for the validity of that proxy. c5.

and firm-specific risk factors Panel A: Cash flow news proxies Variable CFN_1 Descriptiona = Current period cash flows news = (A_epsit – eps1it) ⁄ |Priceit|. = long-term liabilities (COMPUSTAT items dltt) at the end of the fiscal year prior to the date Et)1(rt) is estimated scaled by MKVL. CAR Vol. 4 (Winter 2011) . scaled by the absolute value of the Value Line reported stock price at the time r is estimated (in Year t). expected return news. FSER_N Panel C: Risk proxies MBETA = CAPM beta estimated via the market model using the value weighted NYSE ⁄ AMEX market index return. made in the third quarter of Year t. = Firm-specific expected return news = Change in market beta = mbetat+1 – mbetat. UBETA DM MKVL BP EXGRW Notes: a Throughout the description of the variables. Calculated as the five-year treasury constant maturity as of the month Et)1(rt) in Year t+1 less the five-year treasury constant maturity at the time Et)1(rt) is estimated (in Year t). 28 No. rf is drawn from the U.gov. = book value of equity at the end of the most recent quarter prior to the date Et)1(rt) is estimated scaled by MKVL. Federal Reserve at http: ⁄ ⁄ www. LBP is the natural log of BP.federalreserve. scaled by the absolute value of the price at the time r is estimated (in Year t). Year t refers to the year that the expected return estimates (Et)1(rt)) are made. Calculated as the Value Line forecast of target price made in Year t+1 less the Value Line forecast of target price made in Year t. = unlevered CAPM beta = MBETA ⁄ (1 + Debt ⁄ Equity) where debt is long-term liabilities (COMPUSTAT items dltt) and equity is total common stockholders’ equity (COMPUSTAT item ceq) as of the end of the fiscal year prior to the date Et)1(rt) is estimated.S. LMKVL is the natural log of MKVL.1092 Contemporary Accounting Research TABLE 1 Variable descriptions: cash flow news. If the market value of equity from the Center for Research on Security Prices (CRSP) is not available use the one from COMPUSTAT as of the end of the fiscal year prior to the date Et)1(rt) is estimated. Calculated as COMPUSTAT actual earnings per share for Year t+1 less the related Value Line forecast of Year t+1 earnings. Require a minimum of 20 monthly returns over the 60 months prior to June of the year Et)1(rt) is estimated. stated in millions of dollars. = earnings growth computed by dividing the difference in forecasted earnings five periods in the future less forecasted earnings four periods in the future by the absolute value of forecasted earnings four periods in the future. CFN_TV Panel B: Expected return news proxies EWER_N = Economy-wide expected return news = Change in discount rate = rft+1 – rft. = Terminal value cash flow news = (TPit+1 – TPit) ⁄ |Priceit|. = market value of equity on December 31st of the Year t)1 (prior to the date Et)1(rt)) is estimated.

We calculate book-to-price (BP) as the book value per share at the quarter end closest to a date on or before June 30th of the Value Line publication year 11. Leverage. Accordingly. Hamada 1972. Finally. Debti MBETAi ¼ UBETAi þ Equity UBETA. We ‘‘unlever’’ MBETA using the procedure described in standard finance textbooks (e. and consequently no benefit arising from the tax deduction of interest payments.g. 28 No. Chung 1989). Market value of equity. however.. We do not infer construct validity from the magnitude of the R2 of the cost of equity capital models because assumptions regarding the terminal value imposed by the researcher in the derivation of some implied cost of capital estimates can lead to induced spurious correlation between the proxy and certain risk characteristics. explains a significant portion of the variation in that expected return proxy. We collect market value of common equity from CRSP. although a test of c1 = 1 is a particularly rigorous construct validity test. Nevertheless. it is unclear whether both variables should achieve significance in the empirical model. We compute debt-to-equity by dividing long-term debt by stockholders’ equity using COMPUSTAT data measured at the end of the fiscal year prior to the time we estimate the Et)1(rt) proxies. Our proxy for the market value of equity is the natural log of the market value of common equity (LMKVL) and is collected as described above. Modigliani and Miller (1958) suggest that as debt in a firm’s capital structure increases.. For example.11 Briefly. Thus. it is not surprising that growth. CAR Vol. Employing unlevered beta circumvents this issue. Including the traditional levered beta (MBETA) in the model captures leverage risk as well as market risk (e. Accordingly. Panel C of Table 1 summarizes their descriptions. Berk (1995) argues that book-to-price and market value of equity play similar roles in an incomplete model of expected returns. The formula we employ to unlever beta assumes no certainty. there is no reason to expect the tax adjustment to play a role. and Firm Risk Characteristics 1093 proxy for expected return. In theory c1 should equal 1. or from COMPUSTAT if the data are unavailable on CRSP. if an assumption related to long-range growth in earnings is used to derive the terminal value (as in the PEG model). Realized Returns. Kester. and ignore the tax shield.Expected Returns. i which yields MBETAi UBETAi ¼ 1þ½Debt : i =Equityi Š MBETA is estimated as described earlier. Fruhan. Unlevered beta. In the context of our analysis. a firm-specific risk factor. it is particularly important to assess the association between the expected return estimates and various firm-specific risk factors after controlling for risk factors that are potential candidates for spurious correlation (primarily growth). Book-to-price. 4 (Winter 2011) . The ‘‘catch-all’’ nature of these variables. yielding a high R2 by construction.g. risk increases. Below is a detailed discussion of our measurement procedures for the risk proxies employed in our analysis. As discussed above. The Hamada 1972 formula is an alternative approach that incorporates the value of the tax shield on interest by including an adjustment for tax (1)t) in the denominator. in that both variables capture otherwise unmeasured risk. we log transform the data to mitigate skewness. estimating the model with UBETA allows us to predict unequivocally a positive coefficient on DM (c3). we follow the approach typically used by investment banks and practitioners. mitigates the concern that the model is susceptible to omitted risk factors. and Ruback 1997). which complicates the interpretation of the coefficients on both leverage and beta. Piper. We include unlevered beta (UBETA) to capture the theoretical relation between Et)1(rt) and CAPM beta. we also report the results of a test of c1 = 1. Consistent with prior research. We measure DM as the ratio of long-term debt (described above) to the market value of common equity measured on December 31 prior to the year we estimate expected return.

rPEGST. Thus. Due to the popularity of this measure. (5) Barth. and Li 2007 to compute rDKL. rGLS. in our test of Hypothesis 2. rGOR. Krull. scaled by the absolute value of the Year t+4 forecast. realized returns (rREAL) are a function of Et)1(rt) determined prior to the period over which realized returns and news are measured. Table 3 provides details that are not repeated in the text below. Price per share is the stock price on the Value Line publication date or closest date thereafter within three days of publication. and rPEG. rDIV. with an emphasis on key similarities and differences. however. CAR Vol. we employ risk proxies estimated contemporaneously with our Et)1(rt) proxies. rGLS. The only empirical assumption underlying this method is that analysts’ forecasts of dividends during the forecast horizon and target price at the end of the forecast horizon capture the market’s expectation of those values. rOJN. We estimate expected earnings growth (EXGRW) using the forecasted growth in earnings five years hence. we also include estimates derived from the Fama-French four-factor model (rFF). we follow the procedures outlined in: (1) BP to compute rDIV. cash flow news. rOJN. and (7) Dhaliwahl. but our estimates of Et)1(rt) are measured at the end of the prior year. 4 (Winter 2011) . rDIV The target price method relies strictly on current stock price and analysts’ forecasts of dividends and target prices. which equates current stock price to a finite series of expected future cash flows and a terminal value. whereas rFF is not an implied cost of capital estimate. Before moving on to the measurement of the Et)1(rt) proxies. Expected return proxies We analyze twelve proxies for Et)1(rt). It employs a short-horizon form of the dividend discount model in which a forecasted terminal value truncates the infinite series of future cash flows at the end of Year 5. Table 2 summarizes the assumptions that underlie the nine unique implied cost of capital proxies and the Value Line forecasts needed to estimate each proxy. Konchitchki. Thus. Beliefs about the evolution 12.1094 Contemporary Accounting Research scaled by price per share. (6) Hail and Leuz 2006. (4) Gode and Mohanram 2003 to compute rGM. 28 No. 2009 to compute rHL. two of the proxies are based on an average of a subset of these. As noted in (2). Expected growth. EM. rGOR. we emphasize the importance of timing in our analyses. discounted to the present at the cost of equity capital. As noted in (3). and Landsman 2010 to compute rFF. as well as two popular ‘‘composite’’ proxies not examined in any of the three earlier studies (rHL and rDKL). BP) that contribute to the debate regarding the validity of expected return estimates: rCT. all drawn from extant research. (3) EM to compute rMPEG and rPEGST.12 We follow the estimation procedures employed in prior literature. We log transform these data to mitigate skewness (LBP). in our test of Hypothesis 1 we measure realized returns. Specifically. firm-specific risks existing at t)1 are part of the t)1 information set investors use to determine Et)1(rt). We provide a brief description of each of the proxies below. We calculate EXGRW for each firm-year as the forecasted earnings for Year t+5 less forecasted earnings for Year t+4. The implied cost of capital proxies are derived from the short-horizon form of the classic dividend discount model. Nine are implied cost of capital proxies included in at least one of the three studies (GKS. As noted in the text. (2) GKS to compute rCT. rMPEG. Alternative estimates arise from models that deal with the terminal value in different ways. and expected return news contemporaneously. rGM and rPEG (see the Appendix).

Value Line forecasts Short-horizon rDIV • During the forecast horizon. Realized Returns. • Analysts’ earnings forecasts in Years 1 and 2 equal the market’s expectation. 4 (Winter 2011) 1095 . • Zero dividends in Year 1. two year ahead. and long range. • Year 1 earnings and Year 2 ‘‘abnormal earnings’’ (specification 2b) are positive. current year. one and two year ahead and long range. • Year 1 earnings and Year 2 ‘‘abnormal earnings’’ (specification 2b) are positive. • Beyond the forecast horizon zero growth in ‘‘abnormal earnings’’. • Earnings per share. rPEGST • Analysts’ earnings forecasts in Years 1 and 2 equal the market’s expectation. rMPEG • Beyond the forecast horizon zero growth in ‘‘abnormal earnings’’. • Zero dividends in Year 1. • Long range minimum and maximum target price. 28 No. rPEG • Dividends per share. and Firm Risk Characteristics (The table is continued on the next page. • Analysts’ earnings forecasts in Years 1 and 2 equal the market’s expectation.) CAR Vol. • Beyond the forecast horizon zero growth in ‘‘abnormal earnings’’. one year ahead. • Year 1 earnings and Year 2 ‘‘abnormal earnings’’ (specification 2b) are positive. • Earnings per share. • Earnings per share. one and two year ahead. • Dividends per share.TABLE 2 Summary of forecast assumptions and data requirements for implied cost of capital expected return proxiesa Terminal value • Beyond the forecast horizon analysts’ forecasts of stock price equal the market’s expectation. Expected Returns. current year. analysts’ dividend forecasts equal the market’s expectation.

rGOR Contemporary Accounting Research rGLS • Dividends per share. rCT • Analysts’ earnings forecasts in Years 1 and 2 and of dividends in Year 1 equal the market’s expectation. current year. . • During the forecast horizon without forecasts. • Beyond Year 5. • Beyond the forecast horizon. and long range. Specification 2 is defined as r)1(eps2 ) Reps1). • Earnings per share. 28 No. and long range. • Book value per share. long range. ‘‘abnormal earnings’’ grow at a constant rate. one year ahead and long range. 4 (Winter 2011) • Growth in ‘‘abnormal earnings’’ is a constant rate for all t. current year. current year. For empirical purposes. • Book value per share. • Earnings per share.TABLE 2 (Continued) Terminal value Value Line forecasts • Dividends per share. firms have a 100% dividend payout ratio. analysts’ forecasts of earnings and book value equal the market’s expectation. additional assumptions are imposed regarding the length of the forecast horizon. • Analysts’ earnings forecasts from Years 1 to 5 equal the market’s expectations. Specification 1 is defined as r)1(eps2 + rdps1 ) Reps1). analysts’ dividend forecasts equal the market’s expectation. which is assumed to be the inflation rate. • Earnings per share. b Two definitions of abnormal earnings are used in the various models. one year ahead. • Year 1 earnings and Year 2 ‘‘abnormal earnings’’ (specification 1b) are positive. current year. • Beyond the forecast horizon. each firm’s ROE equals its cost of equity capital. • Earnings per share. one year ahead and long range 1096 Short-horizon rOJN and rGM CAR Vol. one year ahead and long range. • Analysts’ book value forecasts from Years 1 to 4 equal the market’s expectation. • During the analysts’ forecast horizon. firm ROE fades linearly to industry ROE. one year ahead. • Estimated abnormal earning growth rate equals the market’s expectation. • Dividends per share. one year ahead and long range. one year ahead and long range. Notes: a The assumptions above are stated in general terms and underlie the theoretical form of each model. current year. • Growth rate is less than the cost of equity and exceeds zero. firms earn their industry ROE in perpetuity. • Beyond the forecast horizon. • During the forecast horizon.

Finite horizon (Gordon and Gordon 1997). ð1 þ rDIV ÞÀt ðdpst Þ t ¼p 1 ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 5 P Variable þ ð1 þ rDIV ÞÀ5 ðP5 Þ Common name for method and original source rDIV Target price (Botosan and Plumlee 2002). 1097 . Modified economy-wide growth (Gode and Mohanram 2003).t Þ MOMt The average of rCT. and Firm Risk Characteristics CAR Vol. rCT Economy-wide (Claus and Thomas 2001). À ÁÀ1 ð1 þ rGOR ÞÀt ðdpst Þ þ rGOR ð1 þ rGOR Þ4 ðeps5 Þ t ¼1 11 P P0 ¼ b0 þ ð1 þ rGLS ÞÀt ððroet À rGLS Þbt À1 Þ þ ðrGLS ð1 þ rGLS Þ11 ÞÀ1 ððroe12 À rGLS Þb11 Þ t¼ 51 P P0 ¼ b0 þ ð1 þ rCT ÞÀt ðroet À rCT Þbt À1 þ ðrCT À gÞÀ1 ð1 þ rCT ÞÀ5 ðroe5 À rCT Þb4 ð1 þ gÞ rFF ¼ Rft þ bRMRFði. adjusted implied cost of capital (Dhaliwal et al. 4 (Winter 2011) rDKL Fama-French and momentum four-factor model (Barth et al. A ¼ ððc À 1Þ þ dps1 =P0 Þ=2 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi rGM ¼ A þ A2 þ ðeps1 =P0 Þà ½ðeps2 À eps1 Þ=eps1 À ðc À 1ފ.t Þ HMLt þ bMOMði.TABLE 3 Formulas for expected return proxies Formula P0 ¼ rPEG ¼ ðepsp eps4 Þ=P0 5À ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi rMPEG ¼ A þ A2 þ ðeps2 À eps1 Þ=P0 . Mean implied cost of capital (Hail and Leuz 2006). A ¼ dps1 =ð2P0 Þ rPEGST ¼ rOJN pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ðeps2 À eps1 Þ=P0 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ A þ A2 þ ðeps1 =P0 Þà f½ðeps3 À eps2 Þ=eps2 þ ðeps5 À eps4 Þ=eps4 Š=2 À ðc À 1Þg. 2010). rPEG rMPEG b rPEGST Price-earnings-growth ratio (Easton 2004). rOJN rGM P0 ¼ 4 P c rGOR Economy-wide growth (Ohlson and JuettnerNauroth 2003). 2007).tÞ SMBt þ bHMLði.5 rFF rHL Expected Returns. and rOJN The average of rCT. Modified price-earnings-growth (Easton 2004).tÞ ðRM À Rf ÞðtÞ þ bSMBði. Realized Returns. Price-earnings-growth ratio (Easton 2004). rMPEG. rGLS. Mean. A ¼ ððc À 1Þ þ dps1 =P0 Þ=2 rGLS t ¼1 a Industry method (Gebhart et al. 2001). rGLS. and rGM after limiting each implied estimate to 0. 28 No.

and bMOM(i.t). 4 (Winter 2011) . These factors explain some of the variation in realized returns. we employ near-term earnings forecasts (Year t+2 and Year t+1) to estimate rPEGST. as in EM. book-to-market-factor. Li. rPEG. however. rMPEG. This is a distinct difference vis-a which do not presuppose the set of priced risks. we believe. and Short 2009) employ expected return estimates based on the Fama-French four-factor model. In contrast. rFF Barth et al. and price momentum factor. Following BP we employ long-range earnings forecasts (Year t+5 and Year t+4) in lieu of the short-range earnings forecasts (Year t+2 and Year t+1) to estimate rPEG. We estimate the expected annual factor returns (bRMRF(i. This assumption simplifies the dividend discount model sufficiently that it can be solved for expected return directly.1098 Contemporary Accounting Research of cash flows beyond the forecast horizon are permitted to vary across firms as a function of analysts’ beliefs embedded in target prices. bHML(i. compensation. whereas rOJN is estimated with short. consistent with prior research. size factor.and long-term growth in earnings. and rPEGST The primary assumption underlying the models used to estimate these proxies is that the market expects zero growth in abnormal earnings beyond the forecast horizon.and long-term growth in abnormal earnings.t)) by first calculating each factor’s CAR Vol. which GLS estimate based on historical industry ROE. abnormal earnings. but use of this model to estimate Et)1(rt) assumes the factors they employ explain variation in Et)1(rt) and offer a complete representation of the sources of risk for which investors demand ` -vis implied cost of capital approaches.g. bSMB(i. The sole difference between rOJN and rGM lies in the empirical procedures employed in estimating the model: rGM is estimated with short-term growth in earnings. rCT and rGLS. GLS assume that the market anticipates ROE will linearly fade to an industry-based ROE 12 years hence. Doing so increases our sample size and.t). consistent with many studies (including EM). and Swaminathan (2001) (hereafter GLS) employ the residual income model derived from the dividend discount model (Ohlson 1995) to specify prices in terms of forecasted ‘‘return on equity’’ and expected return. rOJN and rGM Ohlson and Juettner-Nauroth (2005) derive an accounting-based valuation model that imposes a series of assumptions regarding the market’s expectations of near term earnings. We employ analysts’ forecasts to calculate short-term earnings growth rates and. (2010) among others (e. set the infinite growth in abnormal earnings equal to rf less 3 percent. which they set equal to the inflation rate. rGOR This estimate is derived from a finite specification of the Gordon and Gordon 1997 model. estimating rMPEG includes a modification for forecasted dividends. The two approaches deal with the terminal value issue in a different manner. Lee. The rPEG model also assumes that the market expects dividends in Year t+1 to be zero. and the rates of short.. Finally. The four factors included in the model are a market factor. 28 No. is more in keeping with the assumption regarding future cash flows. Kothari.t). which assumes that beyond the forecast horizon the market expects each firm’s ROE to revert to its expected return. Claus and Thomas (2001) (hereafter CT) and Gebhardt. CT assume that the market expects abnormal earnings beyond the forecast horizon to grow at a constant rate. whereas rMPEG relaxes this assumption.

we require: (1) CRSP data to calculate realized returns. and rGM. The primary sample for our analysis of the association among realized returns and the expected return proxies includes 14. if the researcher-imposed assumptions mirror those of market participants. Specifically. The primary sample for our analysis of the association among the expected returns proxies and firm-specific risk factors consists of 17. Realized Returns. 2007 calculate rDKL as the mean of rCT. after ‘‘winsorizing’’ each of these values to a maximum value of 0. and rOJN. 2010. (3) current stock price. and Firm Risk Characteristics 1099 average monthly return over the 60 months prior to month m. Whether this reflects variation in the underlying Et)1(rt) however. Finally. Nevertheless. whereas Dhaliwal et al. researcher imposed assumptions about the behavior of infinite horizon growth in cash flows can create a spurious correlation between the Et)1(rt) estimates produced and growth. (2) Value Line forecasts required to estimate expected return and cash flow news proxies.73 at most for rFF) is consistent with variability in the unexpected component of the return swamping variation in the expected return component.60 (14. and (4) COMPUSTAT data to calculate our firm-specific risk factor variables. such correlations need not be spurious.904 firm-years drawn from 1984–2004.Expected Returns. 4. All of the other implied cost of capital approaches assume market participants’ expectations regarding infinite horizon cash flows are consistent with the assumptions imposed by the researcher. Moreover. consistent with Barth et al.27) percent for rGLS to 15. and then compounding the resulting average monthly returns over the twelve months prior to the beginning of firm i’s fiscal year. CAR Vol. Mean (median) estimates of expected return range from 7. such that a firm must have sufficient data in adjacent years to be included in the sample employed in this analysis. rGLS. rMPEG. Because rDIV does not constrain the behavior of infinite growth in expected cash flows to be the same across firms. 2009) calculate rHL as the mean of rCT. Hail and Leuz (2006. The high volatility of realized returns (39. rGLS. rFF is calculated as an annualized predicted return computed using compounded monthly returns. and correlations Sample construction For each firm-year. depends on whether analysts’ forecasts adequately proxy for the market’s expectations.5. Descriptive statistics Expected and realized return estimates Panel A of Table 4 provides pooled descriptive statistics pertaining to the twelve Et)1(rt) proxies and realized returns (rREAL). 4 (Winter 2011) . Sample construction. This underscores the severity of the power issue that could arise if cash flow and expected return news are ignored.36 (7. rHL and rDKL The final two proxies we examine are averages of certain other proxies.521 firm-years from the same period. 28 No. Comparison of terminal value assumptions The rDIV approach assumes analysts’ beliefs about infinite horizon cash flows accord with market participants’ beliefs imbedded in stock price. rFF limits the set of priced risk factors to a predetermined set resulting in an association between rFF and the predetermined set of risk factors by construction. descriptive statistics.91) compared to the low volatility of expected returns (8. as discussed previously. we expect rDIV to display the greatest cross-sectional variation. Fewer observations appear in our realized return analysis because our cash flow news and expected return news variables are change variables.50) percent for rFF.

11 )2.63 17.521 (The table is continued on the next page.14 13.44 )268.17 17.50 14.11 21.76%** )0.00 )133.02 134.55% 11.82 12.53 13.11 9.91 17.69%* 0.39 11.65% 10.98 8.21 26.13 84.80 27.14 0.93 7.19 33.19 143.48 3.24 14.83 3.52 10.49 19.01 )0.17 15.62 11.39% 9.26 25.33 )23.91 19.904 4.94 9.904 rDKL 10.13 5.98 17.98 0.94 14.20 14.01 36.99 5.36% 11.904 Contemporary Accounting Research Panel B: Cash flow newsb and expected return news proxiesc Cash flow news (scaled by price) CFN_1 FSER_N )0.521 CFN_TV 0.48 10.904 rREAL rDIV rPEG CAR Vol.521 Mean Median Perc.11 2.67 17.76 17.20% 12.521 .11 18.02 17.91 9.61 17.54 1.38 321.32** )0.904 0.60% 14.42 9.06 17.61 4.35 11.00 )36.72 )0.904 5.521 Expected return news EWER_N )0.25 8.12 8.20 14.27 9.36 1.98 5.54 4. 28 No.904 0.63 0. 1% 25% 75% 99% STD N 15.02** )0.521 CFN_TVA 2.26% 14.20 3.608 11. 4 (Winter 2011) Cash flow news (scaled by forecast) CFN_1A )0.904 4.06 1.63% 11.64 14.89 10.89 7.TABLE 4 Descriptive statistics 1100 Panel A: Expected return proxiesa rMPEG 12.48 8.49 13.34 8.11 14.904 5.96 28.) Mean Median Percentiles 1% 25% 75% 99% STD N )0.50 rPEGST rOJN rGM rGOR rGLS rCT rFF rHL 10.31 5.50 2.99 17.45 20.48 12.54 4.66 3.00 )57.23 10.92 17.37 15.39 3.33 16.65 39.09 0.42** )3.90 19.27% 11.57 17.904 4.48 41.92 9.39 )8.72 15.02 49.23% 8.98 30.36% 7.22 11.92 7.18 4.03** )0.904 2.48% 10.20 13.904 3.67 20.43%* 0.27 10.79 17.59 2.89 )13.41 23.09 )67.25 11.73 15.99%* )2.16** )16.

904. See detailed descriptions and calculations for each proxy in Tables 2 and 3. Realized Returns.55 0. but the variables labeled ‘A’ are scaled by the absolute value of the original forecast. The original values are scaled by recent stock price. 28 No.98 Notes: a rREAL is the buy and hold realized return computed over the 12 months beginning after the month expected return proxies are estimated.TABLE 4 (Continued) Panel C: Firm-specific risk proxiesd UBETA 0.14 0. b CFN_1A and CFN_TVA are analogous to CFN_1 and CFN_TV except for the scaling.50 DM MKVL BP EXGRW 13.54 12. c Cash flow news. respectively (2-tailed t-test). 4 (Winter 2011) 1101 .42 0.05 or better levels.64 0. **. * denotes significance at the 0. and Firm Risk Characteristics d Number of observations for these variables is 17. Expected Returns. firm-specific risk proxies variables are defined in Table 1.50 MBETA Mean Median 1.01 and 0.22 4106.73 0.62 909. CAR Vol.02 0.

Correlations Table 5 presents Spearman correlations among sets of variables employed in our regression analyses. while the median value is 0. rGLS.1102 Contemporary Accounting Research Explanatory variables Panel B of Table 4 provides descriptive statistics for our cash flow and expected return news proxies. Finally. 4 (Winter 2011) . The mean change in the target price is significantly positive (0. and rCT) correlate positively with rREAL in ten or more years but. consistent with analyst optimism. in four cases (rDIV. the variable included in our empirical analyses. The first row of panel A.54 percent average growth in expected earnings.42) indicating an annual decline in the risk free rate of one-third to almost one-half of one percent. CAR Vol. Table 5 presents univariate correlations among realized returns and the twelve proxies for Et)1(rt).00 (rPEGST and rFF) to 0.69 percent of stock price). and EXGRW. In contrast.13 CFN_1 captures current year cash flow news via an ‘‘earnings surprise’’ variable. Similarly.32 ()0. where the correlation is statistically positive in ‘‘only’’ 19 years. We provide these data to convey the economic magnitude of the revisions in cash flows. MKVL. CFN_1 has a mean (median) value of )0. and there is a strong positive correlation among rMPEG.03 ()0. Both the mean and median values are statistically negative. The correlations are quite small.90).16) percent of recent stock price.76 ()0.09 (rGOR). The mean (median) value indicates an earnings surprise of approximately )1. Except for the correlation between rDIV and rGLS. These statistics describe a sample similar in average market risk to that of the market portfolio with a mean debt-to-market ratio of 42 percent and a market value of equity that is heavily skewed to larger firms. Four of our Et)1(rt) proxies (rOJN. The mean (median) value of our economy-wide expected return news proxy is significantly negative at )0. Consistent with cash flow news playing an important role in explaining realized returns. this analysis fails to consider critical controls for cash flow and expected return news as modeled in (2). We also provide cash flow news statistics after scaling each component by the absolute value of the relevant forecast at the beginning of the period. We present the mean value of the year-by-year correlations along with the number of years (out of 21) that the annual correlation is significantly (positive ⁄ negative). the correlation between rFF and the implied cost of capital estimates is almost as likely to be negative as positive.96) despite being the product of an average of somewhat different subsets of alternative measures. the pair-wise correlations among the implied cost of capital proxies are statistically positive in all 21 years. UBETA. DM.02). Nonetheless. rPEGST.0. rPEG. BP. Panel B of Table 5 presents the correlations among the variables we employ in testing Hypothesis 1. panel C of Table 4 reports pooled mean and median statistics for our firmspecific risk factors including MBETA. and rGLS). We provide cash flow news statistics after scaling each component of cash flows news by recent stock price. This is consistent with ‘‘good’’ news on average with respect to infinite horizon discounted expected cash flows. the mean (median) value of our firm-specific expected return news proxy is significantly negative at )0. rHL and rDKL are highly correlated (q = 0. and rGM (q > 0.4) percent of the initial earnings forecast. CFN_TV captures the change in analysts’ expectations of target price over the 12-month realized return period. the correlation between rREAL and the cash flow news proxies is large 13.0 ()2. the proxies correlate negatively with rREAL in five or more years. rPEGST. The average book-to-price ratio of 55 percent is consistent with the relatively healthy rate of 13. This indicates that these groups of estimates capture essentially the same underlying construct. This suggests that rFF and the implied cost of capital estimates do not capture the same underlying construct. 28 No. ranging from 0. rGOR. CFN_1A measures the earnings surprise as a percentage of beginning earnings forecast.

56 (21 ⁄ 0) 0.98 (21 ⁄ 0) 0.43 (21 ⁄ 0) 0.70 (21 ⁄ 0) — rCT 0.56 (21 ⁄ 0) 0.05 (10 ⁄ 3) 0.70 (21 ⁄ 0) 0.04 (8 ⁄ 3) 0.91 (21 ⁄ 0) 0.00 (5 ⁄ 7) 0.66 (21 ⁄ 0) 0.57 (21 ⁄ 0) 0.07 (9 ⁄ 2) 0.01 (3 ⁄ 3) )0.56 (21 ⁄ 0) 0.51 (21 ⁄ 0) — rPEGST 0.00 (5 ⁄ 5) 0.01 (5 ⁄ 4) 0.03 (8 ⁄ 3) 0.29 (21 ⁄ 0) — rGM 0.09 (12 ⁄ 2) 0.40 (21 ⁄ 0) 0.04 (6 ⁄ 6) — rPEG 0.44 (21 ⁄ 0) 0.51 (21 ⁄ 0) 0.02 (6 ⁄ 4) 0.84 (21 ⁄ 0) 0.00 (7 ⁄ 6) )0.68 (21 ⁄ 0) 0.79 (21 ⁄ 0) 0.46 (21 ⁄ 0) 0.74 (21 ⁄ 0) 0.86 (21 ⁄ 0) 0.80 (21 ⁄ 0) 0.00 (4 ⁄ 4) )0.00 (21 ⁄ 6) 0.04 (2 ⁄ 7) 0.43 (21 ⁄ 0) 0.80 (21 ⁄ 0) 0.73 (21 ⁄ 0) 0.05 (9 ⁄ 3) 0.73 (21 ⁄ 0) 0.06 (9 ⁄ 2) 0.46 (21 ⁄ 0) 0.03 (7 ⁄ 3) — 0.30 (21 ⁄ 0) 0.63 (21 ⁄ 0) 0.50 (21 ⁄ 0) 0. Realized Returns.36 (21 ⁄ 0) 0.96 (21 ⁄ 0) (The table is continued on the next page. 4 (Winter 2011) 1103 .72 (21 ⁄ 0) 0.74 (21 ⁄ 0) 0.53 (21 ⁄ 0) — rGLS 0.22 (19 ⁄ 0) 0.08 (12 ⁄ 5) 0.08 (13 ⁄ 3) 0.04 (9 ⁄ 3) — Expected Returns.52 (21 ⁄ 0) 0.54 (21 ⁄ 0) — rFF 0.62 (21 ⁄ 0) 0.60 (21 ⁄ 0) 0.90 (21 ⁄ 0) 0.01 (8 ⁄ 5) 0.TABLE 5 Correlations Panel A: Correlations among expected return proxies rPEGST rOJN rGM rGOR rGLS rCT rFF rHL rDKL rDIV rPEG rMPEG rREAL rDIV 0.61 (21 ⁄ 0) — rMPEG 0. 28 No.91 (21 ⁄ 0) — rOJN 0.54 (21 ⁄ 0) 0.01 (5 ⁄ 7) 0.51 (21 ⁄ 0) 0.25 (21 ⁄ 0) 0.49 (21 ⁄ 0) 0.38 (21 ⁄ 0) — rGOR 0.87 (21 ⁄ 0) 0.22 (21 ⁄ 0) 0.72 (21 ⁄ 0) 0.55 (21 ⁄ 0) 0.) CAR Vol.62 (21 ⁄ 0) 0.64 (21 ⁄ 0) 0.39 (21 ⁄ 0) 0. and Firm Risk Characteristics rHL 0.

00 )0.08 (0 ⁄ 6) )0. 28 No.13 (3 ⁄ 14) )0.00 (7 ⁄ 6) )0.01 (7 ⁄ 5) )0.01 (2 ⁄ 2) 0.01 (4 ⁄ 4) )0.TABLE 5 (Continued) 1104 Panel B: Correlations among realized returns.01 (6 ⁄ 4) 0.02 (4 ⁄ 7) )0.14 (0 ⁄ 19) )0.00 (2 ⁄ 3) )0.13 (0 ⁄ 15) 0.01 (0 ⁄ 1) — EWER_N FSER_N 0.16 (0 ⁄ 19) )0.01 (3 ⁄ 3) )0.02 (3 ⁄ 3) 0.06 (7 ⁄ 1) )0.01 (6 ⁄ 9) 0.05 (8 ⁄ 1) )0.06 (0 ⁄ 8) )0.06 (0 ⁄ 7) )0.14 (0 ⁄ 18) )0.04 (0 ⁄ 7) )0.01 (0 ⁄ 1) )0. CFN_TV 0.33 (21 ⁄ 0) — CAR Vol.) .03 (0 ⁄ 5) 0.06 (0 ⁄ 8) )0.01 (5 ⁄ 3) 0.35 (21 ⁄ 0) 0. 4 (Winter 2011) rMPEG rPEGST rOJN rGM rGOR rGLS rCT rFF rHL 0.12 (0 ⁄ 13) )0.03 (0 ⁄ 6) 0.02 (4 ⁄ 4) 0.04 (2 ⁄ 4) )0.07 (0 ⁄ 6) )0.01 )0.01 0.01 (5 ⁄ 6) 0.01 (3 ⁄ 3) )0.01 (6 ⁄ 6) (The table is continued on the next page.06 (0 ⁄ 5) )0.01 (4 ⁄ 4) 0.03 (1 ⁄ 6) 0. rDKL rDIV rPEG CFN_1 Contemporary Accounting Research CFN_TV EWER_N FSER_N )0. and expected return news proxies.01 (6 ⁄ 5) Panel C: Correlations among expected return and cash flow news and expected return new proxies.46 (21 ⁄ 0) — )0.25 (0 ⁄ 21) )0. cash flow news.00 (2 ⁄ 3) )0.00 (4 ⁄ 3) 0.01 (7 ⁄ 8) )0.00 (5 ⁄ 6) )0.09 (0 ⁄ 11) )0.04 (0 ⁄ 6) )0.03 (1 ⁄ 4) 0.01 (0 ⁄ 3) 0.01 (4 ⁄ 5) 0.05 (3 ⁄ 10) 0.04 (6 ⁄ 1) )0.01 (2 ⁄ 4) )0.01 (10 ⁄ 4) (4 ⁄ 5) (2 ⁄ 3) (2 ⁄ 4) CFN_1 rREAL CFN_1 CFN_TV EWER_N 0.

31 (21 ⁄ 0) 0.61 (21 ⁄ 0) 0.02 (2 ⁄ 4) )0.23 (21 ⁄ 0) )0.05 (8 ⁄ 9) )0.35 (21 ⁄ 0) 0.36 (20 ⁄ 0) 0.01 (7 ⁄ 5) )0.24 (0 ⁄ 20) 0.33 (21 ⁄ 0) 0.27 (20 ⁄ 0) 0.63 (21 ⁄ 0) 0.03 (6 ⁄ 4) )0.19 (19 ⁄ 0) 0. CAR Vol. rMPEG rPEGST rOJN rGM rGOR rGLS rCT rFF rHL rDKL rREAL rDIV rPEG MBETA UBETA DM LMKVL LBP EXGRW )0.16 (19 ⁄ 0) )0.20 (21 ⁄ 0) )0.36 (21 ⁄ 0) 0.03 (10 ⁄ 8) 0.01 (3 ⁄ 4) )0.17 (0 ⁄ 18) 0.12 (1 ⁄ 13) 0.62 (21 ⁄ 0) 0.09 (12 ⁄ 0) )0. and Firm Risk Characteristics Table values are the mean of year-by-year correlations.25 (0 ⁄ 20) 0.05 (1 ⁄ 7) 0.09 (13 ⁄ 1) Notes: Expected Returns.39 (21 ⁄ 0) )0.24 (20 ⁄ 0) 0.08 (11 ⁄ 1) 0.16 (0 ⁄ 18) 0.27 (21 ⁄ 0) 0.16 (1 ⁄ 17) 0.26 (21 ⁄ 0) )0. Realized Returns.04 (8 ⁄ 10) 0.30 (21 ⁄ 0) 0. 28 No.18 (17 ⁄ 0) 0.00 (4 ⁄ 4) )0.08 (11 ⁄ 0) )0.06 (10 ⁄ 5) 0.05 (4 ⁄ 10) 0.23 (0 ⁄ 21) )0.TABLE 5 (Continued) Panel D: Correlations among expected return and risk proxies. All variables are defined in Tables 1 or 3.39 (21 ⁄ 0) )0.07 (10 ⁄ 1) 0.19 (14 ⁄ 2) 0.06 (9 ⁄ 3) )0.30 (21 ⁄ 0) )0.10 (10 ⁄ 4) 0.25 (0 ⁄ 21) 0.02 (7 ⁄ 5) 0.14 (18 ⁄ 0) 0.05 (7 ⁄ 1) )0.39 (21 ⁄ 0) )0.16 (0 ⁄ 20) )0.29 (0 ⁄ 21) 0.15 (0 ⁄ 18) 0.78 (21 ⁄ 0) )0.24 (0 ⁄ 20) 0.18 (17 ⁄ 1) 0. 4 (Winter 2011) 1105 .20 (20 ⁄ 0) 0.09 (0 ⁄ 15) 0.14 (0 ⁄ 18) 0.24 (0 ⁄ 21) 0.07 (8 ⁄ 3) )0.69 (21 ⁄ 0) )0.45 (21 ⁄ 0) )0.03 (6 ⁄ 9) 0.14 (0 ⁄ 17) 0.40 (21 ⁄ 0) 0.10 (12 ⁄ 0) )0.09 (1 ⁄ 13) 0.02 (2 ⁄ 6) 0.11 (15 ⁄ 0) 0.02 (3 ⁄ 7) )0.43 (21 ⁄ 0) )0.37 (21 ⁄ 0) 0. Figures in parentheses are the number of years the correlation is significantly (positive ⁄ negative) at 5% level in year-by-year correlations.25 (0 ⁄ 21) 0.29 (0 ⁄ 21) 0.46 (21 ⁄ 0) 0.

highlights the importance of separating leverage risk and market risk. rGM. EXGRW is also highly correlated with rDIV (q = 0. Panel C of Table 5 presents correlations between the Et)1(rt) proxies. rHL. and serves to underscore the importance of controlling for growth in our firm-specific risk analysis. and rFF) correlate positively with MBETA in at least 11 years. while six of the proxies (rOJN. rMPEG. Panel D of Table 5 presents the correlation coefficients among rREAL. rGOR.46). rCT. suggesting that ‘‘good’’ current period cash flow news tends to be associated with ‘‘good’’ long-horizon cash flow news.30) and significantly positive in all 21 years.14 In contrast. and the firm-specific risk factors. Taken in their entirety. which is consistent with growth being an important risk factor in its own right. Nevertheless. Despite the counterintuitive nature of the negative association between rGLS and EXGRW. rCT. documented correlations with MBETA might be due to associations with either risk characteristic. All Et)1(rt) proxies other than rDIV and rFF are significantly positively related to leverage in at least 11 years.. the correlation between rREAL and the risk proxies are frequently contrary to expectations if rREAL is viewed as a proxy for expected returns. There is little relationship between our macroeconomic expected return news proxy (EWER_N) and our firm-specific expected return news proxy (FSER_N). in light of the association among the risk proxies. We also document a strong positive correlation between our cash flow news proxies (0. The particularly strong correlation between EXGRW and rPEG (q = 0. rPEGST. 4 (Winter 2011) . CAR Vol. and the fact that neither is strongly correlated with rREAL supports the conclusion in prior research that cash flow news is the primary driver of realized returns. combined with the fact that several of the Et)1(rt) proxies are positively correlated with MBETA but not UBETA. This finding. and rPEGST). rPEG. rOJN. rPEG. this finding is consistent with prior research (e. There is a fairly strong negative correlation between CFN_TV and several of the Et)1(rt) proxies (rDIV.. Finally.23) and an unexpected positive association between rFF and LMKVL. 28 No. and our proxies for cash flow and expected return news.1106 Contemporary Accounting Research (greater than 0. our Et)1(rt) proxies. and rPEGST are associated with all of the firm-specific risk factors in a theoretically predictable manner. rGOR. The low correlation among most of the Et)1(rt) proxies and our two proxies for expected return news is reasonable since there is no obvious basis for expected return news to be correlated with the initial level of expected returns.36). Gebhardt et al. and rHL). who study a more limited set of proxies over an earlier time period (1983–1993). only three of the Et)1(rt) proxies are positively related to UBETA in at least 11 years (rDIV. while one (rGLS) correlates negatively with MBETA in 20 years. the remaining correlations among the Et)1(rt) proxies and the risk factors accord with expectations. the results presented in panel D suggest that only rDIV. Although we use UBETA in our regression model. Except for an unexpected negative association between rGLS and EXGRW ()0. BP 2005. Six of the proxies (rDIV. we include MBETA in this table for discussion purposes. rPEG. which does not have an algebraic relationship with growth. the expected return news proxies are not highly correlated with the cash flow news proxies. Nonetheless. rGLS. Because MBETA captures both market and leverage risk.78) is explained in part by the algebraic relationship between rPEG and growth. it is important to examine the association between the expected return and firm-specific risk proxies in a multivariate setting. Finally. and rDKL) are negatively related to UBETA in 11 or more years. 2001). suggesting that cash flow news is distinct from expected return news. 14. With respect to rDIV and rPEG this finding is consistent with BP.g. rPEG. which suggests that terminal values tend to decline when cost of equity capital is high at the beginning of the realized return period.

Full regression model Panel B of Table 6 presents the results of estimating twelve specifications of the realized return model with the various Et)1(rt) proxies. but statistically indistinguishable from 1. but the coefficient on firm-specific expected return news (FSER_N) is insignificant. We document a strong positive relation between realized returns and the cash flow news variables (CFN_1 and CFN_TV). the coefficient on EWER_N is significantly negative. We also present the number of years the proxy coefficient is (1) significantly positive and not different from one.9 percent. While most of the expected return proxies correlate positively with realized returns. The average R2 of the baseline model is 25. These findings are consistent with prior research. and rDKL) the average coefficient is statistically indistinguishable from the theoretical value of one. and the coefficient on FSER_N is not statistically different from zero. Except for the coefficients on rPEGST and rFF. Alternatively. rPEG. The average coefficient on rDIV is significantly positive. For several of the proxies (rDIV. We estimate this model to assess the incremental explanatory power attributable to the Et)1(rt) proxies. Also.Expected Returns. We report the time-series averages of the coefficients from annual crosssectional regressions and t-statistics based on the standard error of the coefficient (Fama and MacBeth 1973). (2) significantly positive. if the empirical model is well-specified. In all specifications the coefficients on CFN_1 and CFN_TV are significantly positive. Campbell and Ammer 1993). consistent with expectations.30 (rGM) to 2. The models employing rDIV and rGOR show the greatest increase in R2— an increase of 15 percent to almost 30 percent in both cases. which concludes that expected return news is driven by macroeconomic. with average values ranging from 0. Empirical results Test of Hypothesis 1 1107 Baseline model Panel A of Table 6 provides the results of estimating a baseline realized return model that includes our cash flow and expected return news proxies. the intercept should be zero. CFN_TV might capture firm-specific expected return news. as well as results in Voulteenaho 2002. The sign of the coefficient is most stable across years for rDIV (significantly positive in 19 of 21 years). adding Et)1(rt) proxies to the model has no effect on the associations between realized returns and cash flow and expected return news. It is significantly positive in the greatest number of years (19) and indistinguishable from one in close to half the years (9). 4 (Winter 2011) . leaving no role for FSER_N. in all but one specification (rFF) adding Et)1(rt) to the baseline models improves explanatory power. Below the coefficient on the Et)1(rt) proxy. and (3) significantly negative. Thus. rDIV seems to rise to the forefront in terms of strength of results. rOJN. 28 No. we present t-statistics related to whether the mean proxy coefficient is significantly positive and whether the mean proxy coefficient is equal to 1. providing further support for the appropriateness of our models and proxies. indicating that our proxies for cash flow and expected return news capture one-quarter of the variation in rREAL. rGLS.14 (rGOR). the mean coefficients on the Et)1(rt) proxies are significantly positive. which is the case in all but one of the specifications (rGOR). CAR Vol. rDIV and rOJN perform best in this respect with coefficients not different from one in nine years. rCT. not firm-specific factors (Vuolteenaho 2002. The model estimated with rDIV is also tied for the greatest increase in R2 (15 percent increase). We document a significant negative relation between rREAL and economy-wide expected return news (EWER_N). and Firm Risk Characteristics 5. Realized Returns. Moreover. but excludes the Et)1(rt) proxy. rDKL.

50**) (0.16**) 0.002 (15.104 ()1. R2 0.17) 26.014 (16.99**) CFN_TV (+) 0.85*) Panel B: Model 2: rREALit ¼ a0 þ b1 ERitÀ1 þ b2 CFN 1it þ b3 CFN STit þ b4 CFN TVit þ b5 EWER Nit þ b6 FSER Nit þ eit CFN_1 (+) 0.17) 27.002 (14.68**) {4 ⁄ 9 ⁄ 2} 2.045 ()0.7 )0.37) 28.3 )0.018 (0.015 (17.77**) {5 ⁄ 10 ⁄ 2} 0.92) {4 ⁄ 12 ⁄ 0} 1.17) 29.157 (4.33**) {5 ⁄ 17 ⁄ 0} 0.002 (0.29**) 0.015 (18.15) 26.67**) (0.002 (16.23*) (0.14**) 0.51**) 0.52) {8 ⁄ 16 ⁄ 0} )0.008 (0.002 (15.014 (16.125 ()1.67**) 0.06*) 0.581 (5.322 (1.131 ()1.91*) FSER_N ()) 0.017 (0.94*) 0.29) 27.40) {9 ⁄ 19 ⁄ 0} 0.54) ()3.007 (0. R2 29. Adj.014 (0.133 ()1.014 (17.13**) )0.87*) 0.99**) (3.47**) 0.2 (The table is continued on the next page.131 ()1.0 )0.883 (6.69) {7 ⁄ 13 ⁄ 3} 0.016 (0.83*) ()2.002 (15.9 )0.42**) )0.67) 0.002 (15.22**) 0.68**) 0.66**) )0.31 **) )0. 28 No.014 (15.136 ()1.144 ()1.118 ()1.304 (1.114 ()2.6 )0.434 (2.122 ()1.44) Avg. Adj.18) CFN_1 (+) CFN_TV (+) EWER_N ()) FSER_N ()) Avg.43) 0.161 ()1.36) 28.13) ER (+) 0.63**) 0.90*) 0.008 (0.48**) )0.05**) {7 ⁄ 10 ⁄ 2} 1.002 (15.8 )0.60) {9 ⁄ 14 ⁄ 0} 0.141 (6.95*) 0.44) 0.137 ()1.056 ()0.177 ()2.44**) ()2.22**) EWER_N ()) )0.056 ()0.94*) 0.88*) 0.1108 Contemporary Accounting Research TABLE 6 Regressions of realized returns on cash flow news and expected return news proxies.69**) )0.893 (3. 4 (Winter 2011) .05) 26.33**) 0.42) 0.015 (17.014 (16.09**) )0.008 (0.018 ()0.014 (16.964 (3.07**) (1.002 (16.14**) )0.53) 0.126 ()1.008 (0.124 ()1.) CAR Vol.014 (14.9 0.15) 25.90**) )0.002 (15.34) 0.57) 0.46**) (0.7 Intercept rDIV >0 =1 =1 ⁄ + ⁄ ) rPEG >0 =1 =1 ⁄ + ⁄ ) rMPEG >0 =1 =1 ⁄ + ⁄ ) rPEGST >0 =1 =1 ⁄ + ⁄ ) rOJN >0 =1 =1 ⁄ + ⁄ ) rGM >0 =1 =1 ⁄ + ⁄ ) rGOR >0 =1 =1 ⁄ + ⁄ ) rGLS >0 =1 =1 ⁄ + ⁄ ) rCT >0 =1 =1 ⁄ + ⁄ ) )0.95*) 0.044 ()0. Panel A: Model 1: rREALit ¼ a0 þ b1 CFN 1it þ b2 CFN TVit þ b3 EWER Nit þ b4 FSER Nit þ eit Intercept )0.002 (16.6 )0.

01 level or better. 28 No. which it is in these three specifications. if the empirical model is well-specified the intercept should be zero.117 ()1. and Firm Risk Characteristics TABLE 6 (Continued) CFN_1 (+) 0.0 )0. **. rPEG.Expected Returns. and EXGRW). A second set of ‘‘other measures’’ also useful for this purpose is the set of firm-specific risk factors that theory predicts should be associated with Et)1(rt).002 (16.002 (16.107 ()1.05 p-value) (=1 ⁄ + ⁄ )) in {}.430 (4. rREAL is not correlated CAR Vol. We investigate the source of the difference in our results in section 7. R2 25.014 (18.65) {8 ⁄ 12 ⁄ 0} Avg. In addition.30) {5 ⁄ 13 ⁄ 0} 1. we include the t-statistic for whether the mean coefficient is equal to one (=1) and the number of times the coefficient in the year-byyear regressions is (significantly positive and equal to one ⁄ significantly positive ⁄ significantly negative) (based on a 0. The model includes the risk-free rate along with five risk proxies (UBETA.131 ()1.92**) )0. Adj. In addition.45) 1109 Intercept rFF >0 =1 =1 ⁄ + ⁄ ) rHL >0 =1 =1 ⁄ + ⁄ ) rDKL >0 =1 =1 ⁄ + ⁄ ) Notes: 0.04**) {1 ⁄ 5 ⁄ 7} 1. rPEG and rPEGST are related to all of the proxies consistent with theory.009 (0. 0. Our results provide support for the construct validity of all the proxies we examine except rPEGST and rFF. This conclusion is inconsistent with GKS and EM. We find that only rDIV.013 (0. All variables are defined in Tables 1 or 3.17**) 0.120 ()1.7 This table includes the time-series averages of the coefficients of the 21 annual cross-sectional regressions (1984–2004) and t-statistics for whether that mean coefficient is statistically positive ⁄ negative (>0) using the standard error of the coefficient estimates across the 21 years (Fama and MacBeth 1973).04**) )0.33) 28. LMKVL.74**) (0. Sample size is 14.111 ()1. rREAL is only one of the ‘‘other measures’’ that can be used to assess the construct validity of alternative Et)1(rt) proxies.95*) 0. Although the results in Table 6 provide support for the construct validity of 10 of the Et)1(rt) proxies. which it is in the models estimated with rDIV and rPEGST. No other expected return proxy performs as well as rDIV.032 (0.051 ()0. LBP.05 level or better (1-tailed t-test).83**) EWER_N ()) )0. 4 (Winter 2011) . Test of Hypothesis 2 Table 7 presents the results of estimating regression equation 5. Figures in bold are significant at the 0.08) 0. but this is expected if LMKVL and LBP both serve to capture unmeasured risk (Berk 1995). and in some cases significantly negative relationships between the implied cost of capital estimates they examine and realized returns. Realized Returns. rPEG or rPEGST with respect to the association with firm-specific risk. and rPEGST are related as expected to the risk proxies included in the model.29) 27.177 (3.01.521.33**) (0.36) 0. and the coefficient on the risk-free rate should be indistinguishable from 1.002 (15.014 (15.56) ()6. DM.56**) CFN_TV (+) 0.2 )0. rDIV is related to all but LMKVL. who document insignificant.64**) 0.* denotes significance at the 0.89*) FSER_N ()) 0.014 (0. The next section presents the results of this analysis.014 (17.45) ER (+) )0. for the ER coefficients.93*) 0.

81) ()3.60) Avg.63**) {4 ⁄ 6 ⁄ 2} UBETA (+) 1.06) 0.49) {6 ⁄ 7 ⁄ 1} 0.98) (0.5 6.94**) ()10.86 )0.34) DM (+) 1. R2 8.11 (12.73 (3.60 (9.63 (1.74 )1.82**) 1.19) 1.91 ()9.28**) ()0.27 (8.16 (11.15 (0.59 (2.01**) 2.34 (4.09) 1.61) (2.23) 0.07 ()0.59 (0.04) )0.10 (1.13 (8.52 (3.45 (2.19) 1.85 (8.12**) 1.85**) {3 ⁄ 2 ⁄ 2} 0.40 (2.49) {3 ⁄ 0 ⁄ 1} 0.38**) 28.31 (0.03**) 22.36**) ()1.07 (4.54 )0.48 )0.3 7.71#) )0.30 (1.3 9.07**) {6 ⁄ 6 ⁄ 5} 0.17**) )0.11**) ()2.70) 0.57**) {7 ⁄ 6 ⁄ 0} 1.56**) 67.00 (7.17 (1.02**) ()0.6 16.37 )0.27**) 0.80**) 0.38 (3.55) ()1.07**) 0.85**) 30.77*) )0. Adj.19**) 0.43) 15.21**) )0.07) 2.90 (3.37 )0.57 (2.07 (2.72#) 17.35**) 0.16**) ()1.01**) 0.87) 2.96**) ()9.07 (6.28) LBP (+) 2.65) 0.12 (0.5 4.1110 Contemporary Accounting Research TABLE 7 Regression of expected return specifications on firm-specific risk factors.2 5.21 (13.25**) ()0.29 (0.06) 0. 28 No.36**) 0.00 (0.06 ()2.07 ()0.61 (2.51 (8.46**) 39.16 (8.16 (2.33 )0.16**) {5 ⁄ 0 ⁄ 3} 0.14 ()0.18**) {4 ⁄ 6 ⁄ 2} )0.34 ()0.15**) )0.3 4.98**) 2.18 (7.42**) 0.99**) 28.07**) 4.26) 0.21**) 0.43 (8.06 (15.31 (7.03 ()0.34 (2.24) 2.35**) 0.78**) {4 ⁄ 5 ⁄ 2} 0.96 )0.06 ()1.48) 0. Intercept (?) rREAL >0 =1 =1 ⁄ + ⁄ ) rDIV >0 =1 =1 ⁄ + ⁄ ) rPEG >0 =1 =1 ⁄ + ⁄ ) rMPEG >0 =1 =1 ⁄ + ⁄ ) rPEGST >0 =1 =1 ⁄ + ⁄ ) rOJN >0 =1 =1 ⁄ + ⁄ ) rGM >0 =1 =1 ⁄ + ⁄ ) rGOR >0 =1 =1 ⁄ + ⁄ ) rGLS >0 =1 =1 ⁄ + ⁄ ) rCT >0 =1 =1 ⁄ + ⁄ ) rFF >0 =1 =1 ⁄ + ⁄ ) )111.37 (7.52) rf (+) 14.56 (7.45 (1.97**) ()3.8 6.55**) ()4.12**) 0.21 (12.81 (4.36 (1.75**) 0.40 (4.19) 0.0 7.30**) 0.08 (0.22 (9.08 ()1.18) ()2.02 (2.97) ()4.) CAR Vol.9 (The table is continued on the next page.43) LMKVL ()) )0.92 (1.40) {4 ⁄ 2 ⁄ 2} 0.02 (1.22**) 0.59 (5.3 )2.45 (3.07 ()0.51**) 0.78#) 40.27 (7.51 (9.30 (1.21 ()1.81 )0.81#) 1. 4 (Winter 2011) .71**) ()4.59 ()2.17**) EXGRW (+) )0.16*) {7 ⁄ 0 ⁄ 6} )0.18 (7.36**) 2.30**) ()4.62**) 41.

39 (1. we focus on rDIV and rPEG because we find the greatest support for their construct validity. we are sympathetic to concerns regarding noise and the argument that averaging several valid estimates.70) ()3.37) LBP (+) 1. could yield a superior proxy. the two sets of analyses provide support for the validity of rDIV and rPEG alone. in a reasonable manner with any of the risk factors except for LBP. Realized Returns.Expected Returns.02 ()0.98 (10. rPEG and rPEGST. Adj. # denotes significant in the wrong direction.55**) {5 ⁄ 1 ⁄ 4} 0.41 (4. To combine these measures into one proxy we use factor analysis (to isolate the covariance between the two original proxies) and a simple average. Based on results (not tabled) from realized return and risk-based analyses we find that neither measure yields a proxy that dominates rDIV or rPEG alone.61 (7. CAR Vol.90**) LMKVL ()) )0. Figures in bold are significant at the 0.04 ()0. **.99**) EXGRW (+) 0.87#) DM (+) 0. Other issues In this section. Sample size is 14. each individually measured with error.05 (7.01. R2 39.02**) )0. In addition.20 (4. and (3) substituting realizations for analysts’ forecasts. 28 No. and Firm Risk Characteristics TABLE 7 (Continued) Intercept (?) rHL >0 =1 =1 ⁄ + ⁄ ) rDKL >0 =1 =1 ⁄ + ⁄ ) Notes: 7.55 (3.05 level or better (1-tailed t-test).9 7.89**) ()2. but our analysis suggests that neither is superior to their inputs.521. the results of this analysis provide support for the validity of rDIV.77**) 0.06 ()1. Nonetheless. while the results of the realized return analysis provide support for the construct validity of all of the Et)1(rt) proxies except rPEGST and rFF.17 ()1. we consider three other empirical issues that arise frequently in the literature regarding Et)1(rt) proxies: (1) the impact of analyst forecast bias.47 (1.5 This table includes the time-series averages of the coefficients of the 21 annual cross-sectional regressions (1984–2004) and t-statistics for whether that mean coefficient is statistically positive ⁄ negative (>0) using the standard error of the coefficient estimates across the 21 years (Fama and MacBeth 1973).93 (9. * denotes significance at the 0. All variables are defined in Tables 1 or 3. 6. Amalgamated proxies rHL and rDKL attempt to control for noise in the Et)1(rt) proxy by averaging several proxies.08 (7. To address this issue.57**) {6 ⁄ 2 ⁄ 5} UBETA (+) )0. 4 (Winter 2011) .71**) 1111 Avg.38**) 34. we include the t-statistic for whether the mean coefficient is equal to one (=1) and the number of years the coefficient in the year-by-year regressions is (significantly positive and equal to one ⁄ significantly positive ⁄ significantly negative) (based on a 0.50**) rf (+) 0.01 level or better. 0.53) 1.32) 0.08**) )0. further supporting the conclusion that firm-level realized returns are not a valid construct for Et)1(rt). (2) the efficacy of realized returns for expected returns after controlling for cash flow news.05 p-value) (=1 ⁄ + ⁄ )) in {}. In summary. since these are the only proxies associated with future realized returns and firm-specific risk in the manner predicted by theory. for the rf coefficient. Taken together.

4 percent. 4 (Winter 2011) . We split our sample into ‘‘consistent’’ and ‘‘inconsistent’’ subsamples. The decrease in explanatory power might also be a result of a decreased relation between CFN_1 and realized returns for the inconsistent sample.8 percent and the R2 of the rPEG specification decreases by 26 percent to 20. overstated) Et)1(rt) estimates. In addition. 28 No. although a couple of risk factors that are statistically significant in Table 7 lose significance when the sample is split. the R2 of the rDIV specification estimated with the inconsistent subsample decreases by 23 percent to 22. Consistent (inconsistent) firm-year observations are those for which the sign of the earnings surprise is the same as (different from) the market reaction to the earnings surprise. The coefficients on the risk factors are similar for the consistent and inconsistent subsamples. we identify firm-years for which we have reason to believe that analysts’ beliefs might have deviated from those of market participants at the time rDIV and rPEG are estimated.0 percent. For this to occur.. GKS and Hou. optimism relative to the market’s beliefs) might lead to biased (e.g.7 percent (Table 6) to 34. In contrast. Because the market’s beliefs about future cash flows at the time the proxies are estimated are not observable. we reestimate the realized return model (equation (4)) and expected return model (equation (5)) and present the results in Table 8. These results suggest that deviations between analyst expectations and those of the market do not lead to biased or inconsistent results with respect to the coefficients on the proxies for Et)1(rt). This finding holds for both proxies. which gives rise to the concern that deviations between the market’s expectations embedded in stock price and analysts’ forecasts lead to measurement error in the Et)1(rt) proxies. Splitting the sample has little impact on the tenor of our conclusions. the R2 of the rPEG specification also increases by 14. Since the findings are consistent across the 15.g. even though systematic bias in analyst forecasts (e.1112 Contemporary Accounting Research Impact of analysts’ forecast bias Most Et)1(rt) estimates employ analysts’ forecasts to proxy for the market’s expectations of future cash flows. estimated with the consistent and inconsistent subsamples. and accordingly our analysis of this issue focuses on these measures.15 Panel B presents the results of estimating the risk model with the two subsamples. it need not induce spurious correlations when employing the resulting Et)1(rt) proxy in empirical analyses. and Zhang (2009) express the concern that stale forecasts and ⁄ or bias in analysts’ forecasts could lead to measurement error in the Et)1(rt) proxies. The coefficients on the Et)1(rt) proxies continue to exhibit strong positive correlations with rREAL after controlling for cash flow and expected return news in both subsamples. Employing the data in each subsample. there is an increase (decrease) in the explanatory power of the model estimated with the subset of consistent (inconsistent) observations.6 percent (Table 6) to 31. It is important to note that error or bias in analysts’ forecasts relative to reported earnings is not the issue. measurement error in the Et)1(rt) proxy would need to be systematic and correlated with other variable(s) of interest.5 percent from 27. van Dijk. but do reduce the power of the analysis. Nevertheless. Even so. CAR Vol. The R2 of the rDIV specification estimated using the consistent subsample increases by 14.6 percent.. We expect the observations in the consistent (inconsistent) subsample to be those with the least (greatest) risk of a deviation between analysts’ forecasts and the market’s expectations. Panel A presents the realized return model. unsystematic deviations between analysts’ and the market’s expectations could create noise and reduce the proxy’s power. The issue is whether analysts’ beliefs are consistent with those of the market at the time expected returns are estimated. Our earlier analysis provides support for the construct validity of rDIV and rPEG. In the model estimated with rDIV (rPEG) DM (UBETA) is no longer statistically significant.5 percent from 29.

94**) CFN_1 (+) 0.17) )0.104 ()1.29) 3. R2 22.71**) 0.92) Inconsistent forecasts (6794 obs.32 (9.991 (3.28) DM (+) 0.88**) 0.141 ()2.43 (14.81) )0.01**) LMKVL ()) )0.69**) 0.36) Panel B: Regression of expected return on risk factors Consistent forecasts (7527 obs.8 20.03) 0. 4 (Winter 2011) .06 (4.32 (5.002 (12.016 (15.17*) FSER_N ()) 0.109 ()1.81 (1.32 (9.31 (1.24**) EXGRW (+) 0.28 (0.007 (7.39 ()0. Realized Returns.68) )0.09) 1113 Avg.77**) EWER_N ()) )0.09) ER (+) 1.94*) FSER_N ()) 0.88**) EWER_N ()) )0.75**) 0.002 (15.0 31.010 (9.39 (8.40 (7.164 ()1.82) 0.84) (7. R2 22.32) )0.002 (13.49 ()2.Expected Returns.08**) Avg.68**) 1.44) 0.21) DM (+) 0.44) )0.59 (1.) Intercept rDIV rPEG )0.70) Avg. and Firm Risk Characteristics TABLE 8 Impact of forecast bias Panel A: Realized return regressions Consistent forecasts (7527 obs.28**) CFN_TV (+) 0.90*) )0.) rf (+) 1.) rf (+) 1. Adj.92**) 1. Adj.93) 0.16**) Avg.02**) CFN_TV (+) 0.78) UBETA (+) 1.03 (4.41 (17.40 (0.09 ()0.12 2.) CAR Vol.30 (1.004 ()0. 28 No.138 ()1.002 (14.14**) 0.12 ()0. R2 34. Adj.75**) 0.009 (9.016 (17.) ER (+) 1.4 Intercept (?) rDIV rPEG )3.5 71.02) 0.93**) )0.16 (0.007 (0.13 ()3.12**) 0.4 Intercept rDIV rPEG )0.58 (1.56 (8.27) 3.97) UBETA (+) 1.132 ()1.047 (0.5 Intercept (?) rDIV rPEG )3.18 (0.28**) LBP (+) 2. R2 22.80**) CFN_1 (+) 0.07 (0.050 (0.17**) )0.18**) LMKVL ()) LBP (+) EXGRW (+) 0. Adj.127 ()1.21**) 0.83**) (The table is continued on the next page.21 ()0.13 1.18**) 0.6 Inconsistent forecasts (6794 obs.132 ()2.97**) (10.030 (4.4 68.200 (2.

respectively (1-tailed t-test). we estimate the residuals (rRESID) from the baseline realized return model shown in panel A of Table 6. That is. subsamples. we estimate the expected return model (equation (5)) with rRESID as the dependent variable. Figures in bold are significant at the 0. Panel A includes the realized return regressions after splitting the sample into observations where the sign of the earnings surprise is consistent with the sign of market response to earnings (consistent) and those where the sign of the earnings surprise is inconsistent with the sign of market response to earnings (inconsistent). Ogneva 2008). This can be quite problematic. 4 (Winter 2011) . in Chen.05 or better levels. there is no data advantage to using rREAL as a proxy for Et)1(rt). rRESID should be rREAL purged of the unexpected component of realized returns. and the coefficients on LBP and EXGRW are significant but the wrong sign. In the first stage regression. rRESID should be a proxy for Et)1(rt). with or without controlling for news.01 level or better. we adopt a two-stage approach. **. The coefficient on UBETA is not significant. We include the time-series averages of the coefficients in the 21 annual cross-sectional regressions (1984–2004). * denotes significance at the 0. In the second stage regression. the number of times the coefficient on the ER proxy is (statistically positive and equal to one ⁄ significantly positive ⁄ significantly negative) in the year-by-year regressions. since the data needed to estimate the cash flow news proxies is the same data needed to estimate rDIV and rPEG. Accordingly. presented in panel B.01 and 0. and Wang 2011 the authors substitute future realizations of ROE for the market’s expectations in estimating rPEG. Even after controlling for cash flow and expected return news. CAR Vol. Taken together.g. In theory. Moreover. mirror those presented in panel A. Panel A of Table 9 presents the results of estimating this model. we find no support for the construct validity of rREAL as a proxy for Et)1(rt) after controlling for cash flow and expected return news. We examine the validity of this approach by estimating the expected return model (equation (5)) with rREAL as the dependent variable and augmenting the explanatory variables to control for cash flow and expected return news proxies. For example. We include the time-series averages of the coefficients in the 21 annual cross-sectional regressions (1984–2004). t-statistics using the standard error of the coefficient estimates across the 21 years (Fama and MacBeth 1973). Panel B includes the risk model after splitting the sample into observations where the sign of the earnings surprise is consistent with the sign of market response to earnings (consistent) and those where the sign of where the sign of the earnings surprise is inconsistent with the sign of market response to earnings (inconsistent).. The results. 28 No. Lobo.1114 Contemporary Accounting Research TABLE 8 (Continued) Notes: This table includes two robustness tests. See Table 3 for detailed definitions of all variables. however. rREAL after controlling for cash flow news Some recent studies use realized returns after controlling for cash flows news to proxy for Et)1(rt) (e. As an alternative approach. Substituting realizations for analysts’ forecasts When analyst forecasts are unavailable some prior work employs realized values of future earnings and ⁄ or cash flows in the estimation of implied cost of equity capital. however. Chen. the association between rREAL and many of the firm-specific risk factors contradicts theory. t-statistics using the standard error of the coefficient estimates across the 21 years (Fama and MacBeth 1973). these results are also more consistent with a power issue than biased and inconsistent results. we see no benefit to using realized returns as a proxy for expected returns.

01 level or better.13) 0.02 (0. The values presented are the time-series averages of the coefficients in the 21 annual cross-sectional regressions (1984–2004).97**) 0.01 ()4. * denotes significance in the predicted direction at the 0.40) 0.69#) Avg. CAR Vol.02 (0. 28 No. Adj. Adj.017 (0.81) )14.13**) )0.TABLE 9 Additional analyses Panel A: Controlling for CFN EWER_N ()) )0. respectively (1-tailed t-test). t-statistics using the standard error of the coefficient estimates across the 21 years (Fama and MacBeth 1973).45#) FSER_N ()) rf (+) UBETA (+) DM (+) LMKVL ()) LBP (+) EXGRW (+) )0.4 Intercept (?) rf (+) rRESID 1.04 (5.14 ()8.002 (15.04 (5.93#) EXGRW (+) )0. Realized Returns.02 ()2.16**) )0.15 ()9.94**) 0.02 ()3.54#) Avg.077 ()1.78) 0.4 INTERCEPT CFN_1 (+) CFN_TV (+) rREAL 1.05#) Notes: Expected Returns. and Firm Risk Characteristics This table includes two robustness tests. **. Figures in bold are significant at the 0. Panel A is based on regressing realized returns on the cash flow and expected return news variables (as controls) and the risk factors included in Table 7.011 (11.19 (2.25) 0. Panel B is based on regressing the residuals from regression of realized returns on the cash flow news and expected return news variables.06) )0.25**) Panel B: Cost of capital estimate (= residuals from realized returns) regressed on risk proxies UBETA (+) 0. 4 (Winter 2011) 1115 .02 ()3.54 (1. See Table 3 for detailed definitions of all variables.87**) )13. R2 35.49**) DM (+) LMKVL ()) LBP (+) )0.01 and 0. # denotes significant in the wrong direction. R2 15.66 ()1.01 ()7.05 or better levels.

A more complicated issue explains the difference between our results and those of EM.t ¼ EtÀ1 ðrt Þ þ ðNcf . rREAL. after controlling for cash flow and expected return news. solving (7) for Et)1(rt) yields:16 EtÀ1 ðrt Þ ¼ f ðDCF Þ À f ðDCF Þ þ f ðDPÞ À f ðDP Þ ð8Þ: The right hand side of (8) implies a product that is close to zero. As noted earlier. CAR Vol. and therefore.g. This difference is particularly stark with respect to rPEG. all ICC $ f(CF. EM’s proxy for expected return news is a scaled measure of the difference in consecutive implied cost of capital estimates. equation (2)) is shown below for convenience. The theoretical specification of the realized return model (i. which confounds the coefficient on the Et)1(rt) proxy included in the model. EM also conclude that the GKS results suffer from a severe omitted variable bias. it is empirically problematic because it provokes circularity in the empirical model.t Þ ð6Þ: Empirically. which are also included in the model as dependent and explanatory variables. Nevertheless. by construction. DP Þ ð7Þ: EM’s proxy for expected return news (DICC) is by construction a function of DCF and DP. All implied cost of capital estimates (ICC) are internal rates of return that equate current stock price (P) to some series of expected future cash flows (CF). The implied cost of capital estimates are biased upward (downward) for firms with ex post good (bad) cash flow news. since all three studies examine close variants of this proxy. DP).. DP). although our empirical specifications are critically different.t = DICC $ f(DCF. Consistent with this. 7. but.t $ f(DCF). while it 16. P). In EM’s empirical specification Nr. rREAL.t À Nr . growth) vary systematically with firms’ cash flow news. Expected returns are increasing in cash flows (holding price constant) and decreasing in price (holding cash flows constant). all DICC $ f(DCF. we demonstrate that although EM’s proxy for expected return news is theoretically defensible. In the remainder of this section of the paper. respectively.t $ f(DP) and Ncf. we find the expected relation. We employ the change in the risk-free rate to proxy for macroeconomic expected return news. The theoretical specification of our realized return model is the same. Employing these estimates in empirical research might yield biased and inconsistent results if other variables of interest (e.e. the model EM estimate can be described by the following set of relationships: f ðDPÞ ¼ EtÀ1 ðrt Þ þ f ðDCF Þ À f ðDCF . 4 (Winter 2011) .1116 Contemporary Accounting Research because it leads to systematic error in the estimates that is related to ex post cash flow news. whereas GKS and EM find that none of the proxies they examine are positively associated with realized returns. ICCs vary across approaches as different CF assumptions arise from different terminal value assumptions. Thus. Expected return is not likely to explain realized returns under this empirical specification. and the change in market beta to proxy for firm-specific expected return news.. we document little or no correlation between rREAL and the expected return proxies in a univariate setting. Stated another way. Our results differ from GKS because their model does not include necessary controls for new information. Reconciliation with prior research Almost all the implied cost of capital proxies we examine are positively correlated with realized returns after controlling for cash flow and expected return news. Consequently. 28 No.

presented in Table 10. of the ICC estimate employed. and Firm Risk Characteristics 1117 is theoretically defensible to use the change in true Et)1(rt) to capture expected return news. panel B. we estimate the regression model using EM’s cash flow news proxy. in the absence of this substitution we would expect to observe no association between rREAL. Conclusions Existing literature employs two approaches to assess the validity of alternative proxies for firm-specific cost of equity capital or expected return (Et)1(rt)). rCT. DICC). and consistent with EM’s results. We estimate the model with the five implied cost of capital proxies that overlap with the prior work (rPEG. Finally. regardless of the validity. panel B for the rDIV specification. rMPEG. The former employs our empirical proxy. That is. the coefficients on the Et)1(rt) proxies are significantly positive. One approach relies on the theoretical link between realized returns and Et)1(rt). we reestimate our realized return model using EM’s cash flow and expected return news proxies (hereafter CFN_EM and ERN_EM. demonstrate that. the rDIV model in Table 6 has an R2 of 29. For example. and rGLS) plus rDIV. The cash flow news proxy EM employ in the estimation of their realized return model differs from the cash flow proxies EM employ in the estimation of ICC. it is empirically problematic to use the change in an Et)1(rt) proxy measured via an implied cost of capital approach for this purpose. the coefficient on EM_ERN is positive and highly significant in all specifications (t-statistics ranging from 3. 8.t. We find it quite plausible that among the alternatives.t and EM’s expected return news proxies (i. This breaks the cycle of near perfect circularity suggested by our analysis. which are significantly positive and negative.17 To provide further evidence of the impact of EM’s empirical specification for expected return news on the coefficient on the Et)1(rt) proxy. and as a result. once the circularity issue is resolved.7 percent in Table 10. 4 (Winter 2011) . it is interesting to note the difference in the R2s of the models estimated in Table 10 panel B versus Table 6. but merely masks the underlying problem. Panel A of Table 10 presents these results. The primary assumption underlying rDIV is that analysts’ beliefs regarding short-term cash flows and terminal value concur with 17. The only difference between these models is the empirical proxy for cash flow news. Because our measures of expected return news are independent of the derivation of the implied cost of capital estimates. the circularity we are concerned with not only manifests in no association between rREAL. and further complicates the interpretation of the results. respectively. The resulting provoked circularity in the empirical model provides no role for Et)1(rt) to contribute to the explanation of rREAL. since we find strong support for the construct validity of the latter proxy. As predicted by our analysis above..Expected Returns.7 percent – 68 percent higher than the R2 of 17. The higher R2 achieved with our cash flow news proxy provides evidence of its greater explanatory power. rDIV and rPEG consistently demonstrate the greatest degree of construct validity.56). Finally. 28 No. they do not provoke circularity in the empirical specification of the model. These results. but with this substitution the expected outcome is less clear. or lack thereof.e. Moreover.86 to 15. but in a strong association between rREAL. while the latter employs EM’s proxy. except for the coefficients on rCT and rDIV. Based on results from both approaches we conclude that there is support for the construct validity of two of the Et)1(rt) proxies we examine: rDIV and rPEG.t and the Et)1(rt) proxies. while the second approach relies on the theoretical relation between Et)1(rt) and priced risk characteristics. In addition. any ICC estimate included in the model to proxy for Et)1(rt) will be statistically insignificant. rGM. CAR Vol. Realized Returns. but our measure of expected return news. the coefficients on the Et)1(rt) proxies are statistically insignificant. respectively).t and the Et)1(rt) proxies.

7 26.07) ()0. 4 (Winter 2011) .1118 Contemporary Accounting Research TABLE 10 Regressions of various specifications of expected returnb on EM cash flow news and expected return news proxies Panel A: rREALit ¼ a0 þ b1 ER it À 1 þ b2 CFN EM it þ b3 EM ER N it þ eit Intercept 0.92*) (2.18 12.46**) (5.29) ()1. t-statistics in the table are based on the time-series averages of the coefficients in the 22 annual cross-sectional regressions 1984–2004 (Fama and MacBeth 1973). Importantly.01 (0.52 0.34**) (4.51 0.56**) (7. Figures in bold are significant at the 0.11**) EWER_N (+) 13. in their discussion of terminal values.01 level or better.94**) (1.15 0.61 0.18) 0.23**) (2.76 0.00 (0. # denotes significant in the wrong direction.03 0.01**) (4.15**) (4.07 )0. Adj.31 0.83) (2. one would typically not expect a firm to extend its supernormal profits to new additional projects year after year.48 0.18**) EM_ERN (+) Avg.45 0. R2 11. Harris.01 and 0. those of market participants embedded in stock price.03 0.41 0.18) 0.15 0. This assumption is not unique to our study. because this is the only researcher assumption imposed on terminal value in the estimation of rDIV.31 )0.01 (0. terminal values are free to reflect whatever assumptions analysts make with regard to infinite horizon cash flows and future discount rates. and is supported by existing research (Barron. Palepu.46**) Panel B: rREALit ¼ a0 þ b1 ERitÀ1 þ b2 CFN EMit þ b3 EWER Nit þ b4 FSER Nit ER (+) CFN_EM (+) 0.06 )0.79**) ER (+) )0..08 ()0.2 25.19 )0.0 17.07 0. .15 (6.16 0.5 18.31 10.46 ()2.09 )0.86**) (6. R2 10.7 13.24**) (0.12 0.99**) (6. Since the benefits of the CAR Vol.36**) *. respectively (1-tailed t-test). See Table 3 for detailed definitions of all variables.49 (6.83*) (1.96 12. Healy.48 0.85) ()0.05) Avg.30) ()1.64) (1.94**) (16.16 0.30 0.9 17. growth in abnormal earnings is zero.36) CFN_EM (+) 0.34**) (13.60 0.6 16. rDIV is not constrained across firms or industries by researcher-imposed assumptions regarding the behavior of terminal values.01 (0. Each new project would generate cash flows with a present value no greater than the cost of investment — the investment would be a ‘zero net present value’ project.02**) (2.44 0.94**) (13. since it is commonly taught in financial statement analysis courses.4 17. It is reasonable to expect this researcher-imposed assumption mirrors the assumption frequently employed by analysts and market participants.50 0.10**) (8.8 16.19 11. Further.72) ()1.83**) (2.09 )0.87**) (11.03 )0.17) rDIV rPEG rMPEG rGM rCT rGLS Notes: 0.94*) (2.43 )0.59**) (16.92**) (14..49 0.08 (6.11 0.25**) (5.80 (1.003 (0. the primary researcher-imposed assumption underlying rPEG is that.90**) (16. * denotes significance at the 0.39 0.49**) (2.01*) (1.05**) (6.01 (0.93*) (1.3 16.01 (0.80) ()2.59**) (4.18) 0.4 31.87 1.83**) (3.67**) (13.3 32.37) ()1.04**) FSER_N (+) 0.05 or better levels.44 1.13 12.25**) (3.57 (4.04 0. Accordingly.75 11.56 0.03 0. and Bernard (1999) state: ‘‘But in the face of competition. For example.49**) ()1.02) 0. Adj.29 )0.9 rDIV rPEG rMPEG rGM rCT rGLS 0.27**) (10.45#) ()1. beyond the forecast horizon.52 0.04) 0.12) 0.9 Intercept )0.18 (15. and Stanford 2005). 28 No.98**) (8.

Finally. We advocate that researchers assess the validity of any new Et)1(rt) proxies by demonstrating a consistent and predictable association between the proxy and future realized returns. we demonstrate that their results are prompted by circularity in their empirical model generated by their empirical approach to measuring expected return news. and (4) the substitution of realized values for analysts’ forecasts of cash flows or earnings. This is one example. but similar instruction can be found in almost any financial statement analysis text. it does nothing to enhance the current value of the firm. we recommend that researchers requiring a valid Et)1(rt) proxy employ either rDIV or rPEG estimated with analysts’ forecasts and we caution against the use of realized returns with or without controlling for cash flow news to proxy for Et)1(rt). (1999. 12–6). we note that substituting realized values for analysts’ forecasts in the estimation of implied cost of equity capital yields estimates that are systematically biased upward (downward) for firms with ex post good (bad) cash flow news. (2) the efficacy of realized returns for expected returns before and after controlling for cash flow news. Finally. CAR Vol. With respect to EM. we consider several other issues raised in the literature regarding implied cost of capital estimates. Our evidence suggests that the impact of deviations between analysts’ expectations and those of the market is limited to potentially less powerful proxies. including (1) the impact of analysts’ forecast bias. 18. Finally. we note that the primary difference between rDIV and rPEG is that rDIV effectively allows the terminal value assumption to vary across firms. and the associated growth can be ignored. 4 (Winter 2011) .Expected Returns. Palepu et al. On the third point we find that the act of averaging several proxies does not yield an enhanced metric. we find that realized returns are not a reliable proxy for expected returns even after controlling for cash flow news. We leave an investigation of this supposition for future research. Realized Returns. 28 No. In conclusion. which could yield biased and inconsistent results if the resulting measurement error is correlated with other variables of interest. Consistent with EM and our univariate analysis we conclude that the results in GKS are attributable to an omitted variable bias arising from a lack of adequate controls for new information. On the second point. while rPEG imposes an assumption of zero growth in abnormal earnings beyond the forecast horizon on all firms regardless of their circumstances. and Firm Risk Characteristics 1119 project are offset by its costs. as well as established risk measures. This suggests that rDIV might be superior even to rPEG for firms with nonzero growth in abnormal earnings beyond the forecast horizon.’’18 The results of our realized return analysis differ markedly from those documented in prior research. (3) the effectiveness of averaging several proxies to produce superior measures.

L. Chen. and Plumlee. Scholes. Stanford. J. rPE is the expected return estimate imputed from the price to forward earnings model. Working paper. C. W.. and J. J. Assessing alternative proxies for the expected risk premium. 28 No. 2011. and R. Kothari and Shu (2005) Easton and Monahan (2005) Botosan and Plumlee (2005) x x x x x x x x Current study x x x x x x x x x x x x x x x x x x x x x x References Ashbaugh-Skaife. Cost of capital and earnings transparency. A variance decomposition for stock returns. A re-examination of disclosure levels and expected cost of equity capital. The Journal of Finance 48 (1): 3–37... 2005. and Plumlee. Lakonishok. and M. Beaver. Harris. Barth. The Accounting Review 45 (4): 654–81. D. Stanford University. LaFond. Are the reports of beta’s death premature? Journal of Portfolio Management 19 (4): 51–62. 1995. W. 1979. O. The effect of SOX internal control deficiencies on firm risk and cost of equity. 2010. and J. Wang. 2009. Guay. 1991. Forthcoming. J. Lobo. Zellner.. Collins. 1993. Journal of Accounting Research 40 (1): 21–40. Thousand Oaks. Chen. The Economic Journal 101 (405): 157–79. Campbell. W. Z. Konchitchki. 1993. Reliability and validity assessment. A critique of size-related anomalies. Y.1120 Contemporary Accounting Research Appendix Summary of Et)1(rt) proxies examined in related research Et)1(rt) proxy rDIV rPEG rMPEG rPEGST rOJN rGM rGOR rGLS rCT rFF rHL rDKL rdagr rPE Notes: rdagr is the expected return estimate imputed from Easton’s 2004 implementation of the Ohlson and Juettner-Nauroth 2005 model. The Accounting Review 80 (2): 403–21. All other expected return estimates are defined in Tables 2 and 3. M. D.. Carmines. Chan. Evidence that investors trade on private event-period information around earnings announcements. Effects of audit quality on earnings management and cost of equity capital: Evidence from China. 4 (Winter 2011) . M. Barron. Kettler.. CA: Sage. M.. G. C. Review of Financial Studies 8 (2): 275–86. Campbell. What moves the stock and bond markets? A variance decomposition for long-term asset returns.. G. Kinney Jr. A. P. The Accounting Review 80 (1): 21–53. and W. and M. H. and Y. Ammer. 1970. Botosan. 2005. Botosan. E. The association between market determined and accounting determined risk measures. Landsman. Journal of Accounting Research 47 (1): 1–43. CAR Vol. H. Berk. R. 2002. Contemporary Accounting Research. and R. J... J.

C. New York: McGraw-Hill ⁄ Irwin. Towards an ex-ante cost of capital. Working paper. 2004. Lintner. Working paper. A.. and analyst forecasts: A study using content analysis.. 28 No. T. Gordon. J. Hvidkjaer.. D. Fruhan. L. P. and implied cost of equity capital. W. International differences in the cost of equity capital: Do legal institutions and securities regulation matter? Journal of Accounting Research 44 (3): 485–531. J.Expected Returns. and Firm Risk Characteristics 1121 Chung.. Working paper. and financial press on cost of capital. and J. M. Zhang.. Discussion of ‘‘On the relation between expected returns and implied cost of capital’’. Evaluating accounting-based measures of expected returns: Easton and Monahan and Botosan and Plumlee redux. D. Debt and risk: A technical note. Journal of Political Economy 81 (3): 607–36. Mohanram. Equity premia as low as three percent? Evidence from analysts’ earnings forecasts for domestic and international stock markets. Working paper.. The valuation of risk assets and the selection of risky investments in stock portfolios and capital budgets. and B. Hail. S. and C.. Monahan. R. PEG ratios. The Accounting Review 80 (2): 501–38. 4 (Winter 2011) . and J. and M. average returns. The Journal of Finance 54 (4): 1199–220. Hou. Guay. Modigliani. X. Financial Analysts Journal 53 (3): 52–61. L. 2006. Lee. Review of Accounting Studies 8 (3): 399–431. K.. Leuz. 2010. Ruback. 1997. Cost of capital effects and changes in growth expectations around US cross-listings. and S. Journal of Accounting Research 39 (1): 135–76. 2009. Miller. and O. 2009. The expected cost of equity capital. W. Fama. E. Journal of Business Finance and Accounting 16 (5): 719–27. 2010. 2007. Stanford University. Kothari. and M. An evaluation of the reliability of accounting based measures of expected returns: A measurement error perspective. 2002. Journal of Financial Economics 93 (3): 428–54. Easton. J. 2009. van Dijk. and estimating the implied expected rate of return on equity capital. A. 2001. Case problems in finance. and M. Hail. R. return and equilibrium: Empirical tests. Elton.. 1972. Hamada. Properties of implied cost of capital using analysts’ forecasts. K. The finite horizon expected return model. Krull. Gode. Wang. Easton. 1973. O’Hara. Lambert. 1999. Review of Economics and Statistics 47 (1): 13–37. The Journal of Finance 56 (5): 1629–66. and C. 1958. American Economic Review 48 (June): 261–97. 2001. Gebhardt. So. Easton. P.. Monahan. Evaluating implied cost of capital estimates. and asset pricing tests. W. Is information risk a determinant of asset returns? The Journal of Finance 57 (5): 2185–221. J. The Accounting Review 79 (1): 73–96. and the theory of investment. Leuz. Risk. and J.. 2009. P.. Thomas. 1965. The Accounting Review 85 (1): 315–41. and S. P. Li. return volatility. realized return. 2005.. Massachusetts Institute of Technology. 2010. The effect of the firm’s capital structure on the systematic risk of common stocks. CAR Vol. S. Did the 2003 Tax Act reduce the cost of equity capital? Journal of Accounting and Economics 43 (1): 121–50. The implied cost of capital: A new approach. Realized Returns. Inferring the cost of capital using the Ohlson-Juettner model. The Ohio State University. 2003.. Notre Dame University. and P. The Accounting Review 82 (5): 1255–97. Expected return. The Journal of Finance 27 (2): 435–52. Review of Accounting Studies 14 (2–3): 260–68. Gordon. MacBeth. Earnings smoothness. Lee. L. analysts. corporation finance. Kester. 2005. Claus. PE ratios. and P.. D. The effect of disclosures by management. F. Easley. C. E. Kothari. and C. McInnis. and Y. J. S. 1997. Short.. Li. W. 1989. S. Shu. Swaminathan. and R.. Dhaliwal. Piper.

and B. Mason. Ohlson. J. 2008. Palepu. 2002. Business analysis and valuation: Using financial statements. J. 2007. Ogneva. Vuolteenaho. Accrual quality and expected returns: The importance of controlling for cash flow shocks. 1966. Juettner-Nauroth. P. Swaminathan. 2008. Pastor. Internal control weakness and cost of equity: Evidence from SOX section 404 disclosures. Sinha. S. Expected EPS and EPS growth as determinants of value. 2005. Review of Accounting Studies 10 (2–3): 349–65. The Journal of Finance 19 (3): 425–42. Ross. 4 (Winter 2011) . Econometrica 34 (4): 768–83. 28 No. Ogneva.. Healy. Subramanyam. Earnings. W.. M. Contemporary Accounting Research 11 (2): 661–87. K. Capital asset price: A theory of market equilibrium under conditions of risk. The Accounting Review 82 (5): 1255–97. 1976. OH: Thomson ⁄ South-Western. R. K. The Journal of Finance 63 (6): 2859–97. Equilibrium in a capital asset market. M. Journal of Economic Theory 13 (3): 341–60. CAR Vol. 1995. book value. and B. Stanford University. and V.1122 Contemporary Accounting Research Mossin. The arbitrage theory of capital asset pricing. Bernard. What drives firm-level stock returns? The Journal of Finance 57 (1): 233–64. 2nd ed.. Ohlson. Working paper. L. 1964. Estimating the intertemporal risk-return tradeoff using the implied cost of capital. J. T. and K.. 1999. and dividends in security valuation. M. Raghunandan. Sharpe.