SEARCH

SEARCH BY CITATION

Keywords:

  • capital markets;
  • valuation;
  • market efficiency;
  • portfolio strategies

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Literature Review
  5. 3. Valuation Measures
  6. 4. Data
  7. 5. Empirical Analysis
  8. 6. Implications
  9. 7. Conclusion
  10. Appendix
  11. References

Valuation signals have been among the most popular between equity portfolio managers. Given the large variation of techniques and theories with regard to how value is measured, this study investigates the efficacy of alternative value measures. We consider a cross section of simple and sophisticated alternative measures and focus on comparison metrics of primary interest to equity portfolio managers. Our results show that sophisticated valuation models are superior – although not universally – relative to simple valuation models in many respects. Thus, we conclude that sophisticated models have interesting attributes and, in general, should be considered as an additional if not primary perspective on equity valuation and portfolio management.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Literature Review
  5. 3. Valuation Measures
  6. 4. Data
  7. 5. Empirical Analysis
  8. 6. Implications
  9. 7. Conclusion
  10. Appendix
  11. References

Valuation is the cornerstone of the investment industry. Investments are made on the basis of what an investor believes is an asset's worth relative to its market price. The key is in knowing the value of the asset and the factors that determine its value. Given the developments in this field of research over the past years and the potential benefits associated with more sophisticated valuation methodologies, academics as well as practitioners have been interested in exploring the incremental information contained in sophisticated valuation measures.

Dissanaike and Lim (2010) provide the most comprehensive study in Europe, with a focus on the UK equity market. In particular, the authors investigate the efficacy of alternative valuation measures in the context of equity portfolio construction. Their valuation models include sophisticated approaches such as the Residual Income Model (RIM) pioneered by Ohlson (1991, 1995) and Feltham and Ohlson (1995), as well as variants. Dissanaike and Lim (2010) also include simple valuation measures such as book to market and cash flow to price. The authors conclude that, while the signals obtained from the sophisticated models are more informative for future cross sections of equity returns (in mid- to long-term horizons) than those obtained from simple valuation measures, the difference is marginal.

Our analysis further investigates whether sophisticated valuation models provide incremental information over simple valuation models. Our major contribution is that we study the practical implications of using different valuation models in valuation-related investment strategies. To achieve this objective, we organise our empirical investigation in a setting that largely resembles real-life equity portfolio management. This is an important aspect not fully addressed in Dissanaike and Lim (2010) or other studies.

In particular, we investigate stocks that are or have been constituents of a broadly used benchmark index, the Morgan Stanley Capital International (MSCI) Europe Index. Typical investment mandates are more likely to be pan-European as opposed to country specific (e.g., Dissanaike and Lim, 2010, who use a UK universe), which makes our investigation potentially more realistic. While this choice restricts our universe to about 450 stocks, on average, per rebalancing period, it guarantees that the stocks in our sample a) are those a typical equity portfolio manager would be allowed to invest in, b) have a reasonable history of data with which to compute all relevant variables, and c) are available for shorting either directly or through derivatives. From a methodological viewpoint, using a pan-European universe of stocks, which has not been previously done in the context of contrarian strategies, addresses the potential data-snooping bias addressed by Conrad et al. (2003) and stressed in Dissanaike and Lim (2010).

The portfolio construction and portfolio performance evaluation approaches we follow are also in line with the perspective of equity portfolio managers. Frequent, that is, monthly, rebalancing may be restrictive insofar as whether a month is sufficient for accounting-related information to be reflected in market prices; however, this choice is justified by the way practitioners, in general, manage money. Equity portfolios are typically revisited monthly or quarterly and the relevant question for the investment professional would relate to the measure that captures firm value best over a 1- or 3-month time horizon.

Moreover, our evaluation and robustness metrics extend the set of results presented in current published research by including measures such as signal consistency, alpha decay, and portfolio turnover, which are all critical in portfolio management. Our objective is to shed light on whether the valuation models we consider provide consistent valuation signals and to identify the model(s) that maximise after-cost profits. These are interesting issues not addressed in Dissanaike and Lim (2010).

One of the models we use, that of Hwang and Sohn (2010), has been argued to provide the most consistent valuation signal for up to three years. Hwang and Sohn (2010) highlight that shareholders possess an option to liquidate their investment if they believe the long-term business prospects of the firm are poor such that the level of capitalised future dividends is expected to be below the firm's net asset value (NAV). The authors suggest that this option, termed the ‘abandonment option’, should be incorporated in the firm's intrinsic value calculation, and thus the measurement of equity value should be undertaken with a real options model (ROM). Their analysis finds that including the abandonment option in the calculation of the firm's intrinsic value provides incremental predictability over the RIM. Our analysis includes the ROM and explores its properties within an applied setting with European data for the first time.

While our research focuses on the perspective of portfolio managers, we believe it addresses several important issues that are concerning academics and provides new useful insights beyond those relating to the investigation of previously unexplored equity universes. A recent article by Richardson et al. (2010) discusses the desirable properties of fundamental analysis and accounting anomalies research. The authors suggest that the most important features in this line of research are a credible alternative hypothesis, robust predictive power, a sound treatment of risk and transaction costs, ensuring the incremental forecasting additivity of newly discovered attributes, and the inclusion of supplemental non-price tests to strengthen inferences. We structure our analysis in a way that is consistent with (the most relevant of) these principles and discuss the academic implications of our study in a separate section (Section 6). These implications briefly involve a better understanding of the drivers of the information content of alternative firm fundamental value proxies, as well as new insights on sector biases of valuation-based portfolios, possibly relating to the biases inherent in analysts’ earnings forecasts (Cheng, 2005).

Our empirical analysis provides important findings that we group in three categories. First, sophisticated models, that is, the RIM and ROM, produce equity valuations that are better able to predict the cross section of short-term future equity returns than simple models. The superior quality of the signal translates to an improvement in the annualised monthly return that can be up to about 5% – that is, the difference in the hedge portfolio return between the RIM and a simple earnings forecast-related measure – or, in magnitude, about 50% or more in terms of the information ratio1 (IR), that is, the difference in the IRs of RIM- and ROM-based portfolios relative to all others (with the exception of a cash flow-related valuation measure that delivers less inferior results). The ranking of the models does not materially change when we consider quarterly rebalancing.

Second, the signal obtained through the RIM and ROM is characterised by superior strength. Equity valuations based on the RIM and ROM are better able to predict the cross section of short-term future equity returns than simple models, even after accounting for transaction costs (with the exception of a cash flow-related valuation measure that delivers only marginally inferior results). When we focus on the rank information coefficient2 (IC), we conclude that although the difference is small, sophisticated models and, in particular, the RIM outperform all others. In addition, when we examine the efficacy of the valuation measures in pre- and post-crisis periods, we conclude that sophisticated measures, particularly the ROM, present with a relatively consistent ranking.

Third, our examination of the purity, or uniqueness, of the valuation signals concludes that sophisticated models are less vulnerable to sector bias. When sector bias is neutralised, the changes to the IRs of portfolios obtained through simple valuation models are up to 78%, while the changes for sophisticated models are at most 20%. As expected, the correlations are almost identical among value factors, but there are differences with regard to the ex–value-factor correlations. The ROM and to a lesser extent the RIM appear to be the most appropriate valuation measures to combine with momentum signals. Our findings have important implications for equity portfolio managers as well as for scholars of equity valuation.

This article is organised as follows. Section 2 reviews the literature related to value investing. Section 3 discusses the valuation measures we use in our analysis. Section 4 details the dataset. Section 5 presents the empirical results of our analysis, Section 6 discusses the implications, and Section 7 concludes the paper.

2. Literature Review

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Literature Review
  5. 3. Valuation Measures
  6. 4. Data
  7. 5. Empirical Analysis
  8. 6. Implications
  9. 7. Conclusion
  10. Appendix
  11. References

Since Basu (1977) and shortly thereafter Ball (1978), concluded that earnings-to-price and dividend-to-price ratios, respectively, predict future returns in the cross section of stocks, a large number of papers have investigated the relation between firms’ fundamentals and their future stock returns. Earlier studies focus on straight measures of fundamental value, such as the debt-to-equity ratio (Bhandari, 1988), the book value of equity to the market value of equity (Chan et al., 1991), and cash flow-to-price ratio (Fama, 1990). The common ground of these studies has been to try to identify variables that better reflect the real value of a firm and thus better predict future equity or price returns.

Recent research efforts have been more sophisticated in the way they propose computing the real value of a firm. Ohlson (1991, 1995) and Feltham and Ohlson (1995) pioneered an alternative valuation model, the (discounted) RIM. The RIM assumes an accounting identity, the clean surplus relation, which states that the change in the book value of equity is equal to the difference between accounting earnings and dividends. The residual income, or abnormal earnings, is defined as the difference between accounting earnings and the previous period's book value multiplied by the cost of equity. The RIM maintains that the current stock price should equal the current book value of equity plus the present discounted value of (infinite) expected future residual income (Jiang and Lee, 2005). In contrast to a simple dividend discount model, the RIM accounts for all forms of cash payouts to shareholders, given that it broadly defines dividends as the difference between earnings and the change in book value.

Frankel and Lee (1998) extend this concept by proposing that for the empirical application of the model, the infinite sum of discounted residual incomes can be approximated with a truncated sum of short-horizon earnings forecasts of up to three  years – which are available with relatively good precision and, moreover, for a large cross section of stocks. The authors find that firms’ fundamental value estimates obtained through the RIM provide better forecasts of long-term cross-sectional returns than a straight book-to-price ratio. Their findings have motivated a large number of researchers since. Lee et al. (1999) find that value to price, where value is based on the RIM, has superior predictive power over traditional market multiples (e.g., book-to-price, earnings-to-price, and dividend-to-price ratios) for Dow 30 stocks. Francis et  al. (2000) find that valuations from the RIM are more accurate and explain more of the variation in security prices relative to valuations obtained with either the discounted dividend model or the discounted free cash flow model. In addition, Ali et al. (2003) find that RIM valuations predict mid- and long-term cross-sectional returns and that this is primarily attributed to the model's ability to identify mispriced stocks (as opposed to stocks with certain risk characteristics).

In a recent development in this strand of literature, Hwang and Sohn (2010) posit that a more accurate valuation model should incorporate the value of the abandonment option effectively held by shareholders. Motivated by Burgstahler and Dichev (1997), the authors suggest that shareholders have the option to liquidate net assets at all times. If the long-term business prospect of a firm is so poor that the level of capitalised future dividends, that is, cash flows for shareholders, is expected to be below the NAV, shareholders can opt to liquidate the firm's net assets. If the contrary is true, shareholders can exercise their call option to take the underlying asset with the payment of the exercise price of the firm's net assets. This model predicts that the value of a company is always higher than its net assets (or very close to its net assets for an impending company liquidation). Hwang and Sohn (2010) find that the value-to-price ratio from the ROM has enhanced predictability for future abnormal stock returns relative to the RIM. In particular, the authors find that for firms in the same RIM-based V/P quintile, those in the highest ROM-based V/P quintile have 14–50% higher future 36-month buy-and-hold abnormal returns than those in the lowest quintile. A follow-up study by the same authors (Hwang and Sohn, 2009) concludes that the predictability of this factor is stronger for stocks with high idiosyncratic volatility.

Overall, the literature that examines the predictability of valuation measures seems to conclude that sophisticated valuations models generally provide more accurate forecasts than simple models for the cross section of future equity returns, especially in mid- to long-term investment horizons. Dissanaike and Lim (2010) posit that the predictability of the RIM, which a priori may carry greater information content and captures the evolution of future earnings more realistically, should be stronger than that of simple models. Moreover, Hwang and Sohn (2010) argue that the ROM is expected to have improved return predictability over the RIM because it explicitly operationalises the value implications of the abandonment option, which is underappreciated by investors. A broader perspective is given in Richardson et al. (2010), who suggest that dividend discount models, variants of which are the RIM and to a lesser extent the ROM, could be successful in predicting future returns. The author argue that dividend discount models essentially represent a simple transformation stating that price is a function of expected returns and expected future financial statement variables (income and change in book value). This framework suggests that future returns should be positively related to expected profitability and negatively related to asset (or book) growth, and much empirical evidence supports this basic prediction. The incremental predictability of sophisticated models may, however, result in aggressive portfolio rebalancing, and therefore increased transaction costs. We study this aspect to determine whether the conjectured incremental predictability of sophisticated valuation models is readily explained away by trading costs.

3. Valuation Measures

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Literature Review
  5. 3. Valuation Measures
  6. 4. Data
  7. 5. Empirical Analysis
  8. 6. Implications
  9. 7. Conclusion
  10. Appendix
  11. References

This study investigates the predictive ability of certain valuation measures over the cross section of future equity returns. In line with Dissanaike and Lim (2010), our analysis involves valuation measures that can be classified as ‘simple’ or ‘sophisticated’.

Our simple valuation measures include the forward price-to-earnings ratio (PE), book yield, fair price-to-earnings ratio, and cash flow (CF), where the forward price-to-earnings ratio is the next 12  months’ (weighted between the current one-year ahead and two-year ahead) I/B/E/S consensus earnings per share (EPS) forecast divided by the month-end closing price, and book yield is the last 12  months’ (weighted between the current and prior year) reported book value per share divided by the month-end closing price. The fair PE, that is, the ratio of the forecast PE to the market PE, is a theoretical PE estimated by a cross-sectional model (similar to a Gordon growth framework) using expected earnings growth and market cap (risk and liquidity proxy) as inputs. The Appendix provides the details of this model. Finally, CF is the next 12  months’ (weighted between the current one-year ahead and two-year ahead) I/B/E/S consensus cash flow per share forecast divided by the month-end closing price.

The sophisticated models include the RIM and ROM.3 Under the RIM, the value of a firm is a function of its current book value of equity (BV) – as a proxy for net assets – plus its residual income:

  • display math(1)

We express RIt as a finite sum of the present value of abnormal future earnings plus a terminal value. We adopt a three-year term model. The terminal value at the start of year three is treated as perpetuity by taking the third year's forecast of abnormal earnings and dividing it by the firm's industry-specific cost of equity. Hence, the firm's RIt is calculated as

  • display math(2)

where inline image is the forecast return on equity (ROE) at time t + n, n = 1, 2, 3, inline image is the reported book value at time t, and rt is the industry-specific cost of equity forecast at time t. We compute the industry-specific cost of equity following Fama and French (1997), as detailed in the Appendix. Following Hwang and Sohn's (2010) methodology, the forecast ROE is calculated as the I/B/E/S consensus EPS forecast for fiscal year t + n divided by the book value per share at time t + n – 1. Additionally, the forecast book value at time t + 1 is equal to the BV at t plus the EPS forecast at t + 1 less the payout ratio prt multiplied by the EPS forecast at t + 1, that is, inline image.

Under the ROM (Hwang and Sohn, 2010), a firm's value is not a linear function of accounting earnings but a convex function of both earnings and book value because of its option-like characteristics. As mentioned earlier, shareholders have the option to liquidate a firm's (net) assets if the firm's long-term business prospects are poor such that the level of capitalised future dividends is expected to be below the NAV. However, if the future dividend or cash flow stream is greater than the NAV, the shareholders ‘exercise’ the call option by ‘paying’ the exercise price and deploying the net assets for future business operations. The option is viewed as a call option4 to buy the future dividend or cash flow stream with the net assets. Shareholders consider this call option by continuously comparing the long-term expected levels of the future dividends/cash flow and the NAV. Formally, the value of a firm can be expressed as inline image, where CO is the call option premium of an option on the capitalised future cash flow stream (underlying asset), with strike price equal to the NAV (which is proxied by the BV).

The option price or valuation is modelled using the standard Black and Scholes (1973) option pricing model. The call option price is calculated as

  • display math(3)

where Vft is the ‘price’ (of the underlying asset) within the Black–Scholes equation and a function of the book value plus the residual income, that is, as in equation (1); BVt is the strike price and the current book value of the firm; RF is the risk-free rate at time t; T is the maturity of the option, during which shareholders compare the values of BVt and Vft; N(d) is the probability that a standardised normal variable will be less than or equal to d, where d1, d2 are computed as inline image, and inline image is the standard deviation of the historical percentage change in Vft, that is, inline image. The fundamental value of the firm and hence the value of the option is computed on a monthly basis. Each month we compute COt through equation (3), using the value of Vft obtained through equation (1) for that month, BVt, RF, and inline image, which we estimate through the standard deviation of monthly percentage changes in Vft for a period of five years preceding that month.

Within the ROM, the value of a firm can be expressed as

  • display math(4)

The difference in valuations obtained through equations (1) and (1), the RIM and ROM, respectively, depend on the factors determining the value of the liquidation option. For firms with high expected profitability and hence high RIM valuation, all else being equal, the call option in the ROM becomes in the money. Its value converges to its intrinsic value, VftBVt, and thus the ROM valuation converges to the RIM valuation. For firms with negative earnings, the contrary is true. The value of the option also depends on the volatility of the percentage changes in Vft. Equation (4) suggests that, all else being equal, the value of the liquidation option and hence the ROM valuation increases for firms with volatile business performance. For these firms, the RIM and ROM valuations are not expected to be similar. If the abandonment option value provides a better means of capturing more accurate intrinsic value through the ROM, the ROM's predictability should be more pronounced when the option value is a larger fraction of the intrinsic value. In fact, Hwang and Sohn (2010) confirm that the enhanced valuation effect of the ROM over a three-year period is more pronounced for firms with low profitability. Their tests regarding the effect of volatility suggest that the predictability of the ROM does not increase monotonically with volatility. This contradictory empirical finding is attributed to a possible measurement error in estimating volatility.

The valuation factors that we use are obviously not independent of each other, although they are based on different rationales. Book yield could be viewed as the least related to the others in a sense. Although the BV is a variable of both the RIM and ROM, valuations between the three approaches may vary, depending on the sign and magnitude of residual income. The forward PE and CF, on the other hand, rely on 1- and 2-year-ahead earnings and cash flow forecasts, respectively. The former are at the core of the RIM and ROM valuations, while the latter also impound analysts’ expectations. However, in the RIM and ROM, industry risk is accounted for explicitly through the industry cost of the equity discount factor. The fair PE is also based on (forecasted) earnings growth; hence, one of its constituents can be seen as being in common with the constituents of the forward PE, ROM, and RIM valuations. These observations suggest that quantitative strategies based on either of these factors should not be dramatically different in terms of their return-to-risk attributes, especially in the relatively short investment horizon considered, considering the motivation of each factor. However, any improvement we see should support the rationale and theory behind the construction of each factor and its capacity to reveal value not captured by its peers and not priced (or not initially priced) by the stock market.

4. Data

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Literature Review
  5. 3. Valuation Measures
  6. 4. Data
  7. 5. Empirical Analysis
  8. 6. Implications
  9. 7. Conclusion
  10. Appendix
  11. References

For the reasons discussed earlier, we focus our analysis on all stocks that are or have been constituents of the MSCI Europe Index during January 1990 to April 2010. The survivorship bias has been accounted for by using the MSCI Europe Index constituents at the time of backtest rebalancing. The fundamental data reported are from Worldscope, while the forecast data are from I/B/E/S.

We apply several standard filters to maintain a sensible dataset. We exclude firms with negative book values and missing equity prices. Companies that have missing return data after the date of portfolio formation, for example, due to bankruptcy, are not excluded in order to address potential look-ahead biases. To account for extreme payout ratios, we set all ratios greater than 100% equal to 100%, and all ratios less than 0% equal to 0%. We apply the same filter for FROEs. Forecast EPS data for three fiscal years out are limited. For firms that lack EPS data for three fiscal years out, we create a proxy EPS forecast for three years ahead (epsFY3) by multiplying the forecast EPS growth by the forecast for two years ahead (epsFY2); that is,

  • display math(5)

where

  • display math

One of the main potential restrictions with the ROM is the requirement to model a stock's residual income over time. This is required because we need to measure the volatility of the residual income, since it is a key input into the option pricing model. Following Hwang and Sohn (2010), we use five years of data to calculate the volatility of the residual income of a firm. This choice is consistent with the idea that five years of operation are necessary to convince shareholders of the real expected level of the ratio of long-term profitability. Given this requirement, before we can start to empirically test the ROM, we therefore require five years of data. We use five years of data, January 1990 to December 1994, to calculate the industry cost of equity as in Fama and French (1997), and an additional five years of data, January 1995 to December 1999, to calculate the volatility of the residual income of a firm. This reduces the available sample for out-of-sample investigation to January 2000 to April 2010. Given the data and the adjustments that we have made, on average, our stock coverage relative to the MSCI Europe is 86%, and in terms of market cap coverage it is slightly more, 88%.

Our comparison adopts a common universe for all valuation measures. This is the universe that we maintain after applying all filters and data requirements for the model with the stricter data requirements, that is, the ROM. This ensures that any differences in performance or risk profile are a result of the model or technique rather than due to differences in sample data.

5. Empirical Analysis

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Literature Review
  5. 3. Valuation Measures
  6. 4. Data
  7. 5. Empirical Analysis
  8. 6. Implications
  9. 7. Conclusion
  10. Appendix
  11. References

We study the predictability of valuation signals in the context of calendar time portfolios. Stocks are ranked with respect to the ratio inline image, where Value is derived by either the forward PE, book yield, fair PE, CF, RIM, or ROM at the end of each month. We then split the universe into five portfolios. The top quintile portfolio based on forward PE, for example, comprises stocks with high earnings yield, while the bottom quintile comprises stocks with low earnings yield. We then calculate the subsequent month's total return for each group of stocks (equally weighted) and rebalance these portfolios monthly.

We organise our experiments and discussion into three groups. The first group is concerned with examination of the quality of the valuation signal, that is, the performance of the constructed portfolio. The second group of tests deals with the strength of each valuation signal, that is, the consistency of the signal across the cross section of stocks and through time, as well as its capacity after transaction costs are accounted for. Finally, we also concern ourselves with the uniqueness of the signals, that is, how the trading strategies arising from the different valuation measures overlap and correlate with other alpha sources.

5.1 Signal quality

Table 1 reports return and risk statistics for three portfolios focussed on: the top quintile portfolio, that is, Q5, of particular interest to long-only portfolio managers; the bottom quintile portfolio, that is, Q1, of interest to managers able to short stocks as well as for long-only managers wishing to identify/avoid future underperformers; and the top minus bottom quintile portfolio, that is, Q5 – Q1. We note that the returns reported for the bottom and top quintiles are spreads over the market returns, that is, over the MSCI Europe Index returns. The reported returns are geometric means (annual) over the entire period as per the rebalancing method discussed in Dissanaike (1994).

Table 1. Spread returns for portfolios formulated on value sorts This table presents the annualised spread returns and performance metrics of portfolios sorted on the basis of alternative valuation measures for the period January 2000 to April 2010 on a monthly basis (total of 124 observations). These measures include the forward PE, which is the next 12 months’ (weighted between fist forward year FY1 and second forward FY2) I/B/E/S consensus EPS divided by the month-end closing price; book yield, which is the last 12 months’ (weighted between current year FY0 and previous year FY-1) reported book value per share divided by the month-end closing price; the fair PE, that is, the ratio of the forecast PE to the market PE, which is a theoretical PE estimated by a cross-sectional model (similar to a Gordon growth framework), using expected earnings growth and market cap (risk and liquidity proxy) as inputs; CF is the next 12 months’ (weighted between FY1 and FY2) I/B/E/S consensus cash flow per share forecast divided by the month-end closing price; the RIM, which is obtained through the RIM (see Ohlson, 1991; Feltham and Ohlson, 1995); and the ROM, which is obtained through the ROM (1995 Hwang and Sohn, 2010). The top quintile portfolio, that is, Q5, comprises stocks with high earnings yield, while the bottom quintile, that is, Q1, comprises stocks with low earnings yield. The returns of the bottom and the top quintiles are spreads over the market returns, that is, over the MSCI Europe Index returns. The IR is the ratio of the annualised return over the annualised volatility. Maximum drawdown is the minimum return obtained for a strategy in the study period. The hit rate is the percentage of correct predictions. The turnover is computed as the total buys and sells divided by the twice the number of stocks in a portfolio; for example, if all the stocks were traded, the turnover would be 100%. We use 50 bps for fixed transaction costs through the period of analysis, which comprises commission, spread, borrow (where applicable), and market impact costs.
Panel A: Top quintile Q5 – market
 Forward PEBook yieldFair PECFRIMROM
Annualised returns7.06%8.92%7.34%8.70%9.02%9.92%
t-statistic1.741.841.862.202.582.40
Annualised volatility14.6%18.0%13.9%13.5%11.7%14.0%
IR0.480.500.530.640.770.71
Maximum drawdown−15.1%−10.4%−14.2%−9.7%−11.8%−10.4%
Hit Rate64.5%58.9%62.9%64.5%64.5%65.3%
Average monthly turnover15%10%18%9.5%15%15%
Returns after costs5.2%7.6%5.1%7.5%7.0%8.0%
% reduction in returns−26.4%−14.5%−30.4%−14.1%−22.1%−19.6%
Panel B: bottom quintile Q1 – market
 Forward PEBook yieldFair PECFRIMROM
Annualised returns−1.05%−0.89%−1.68%−1.63%−3.72%−1.80%
t−Statistic−0.16−0.29−0.57−0.61−1.00−0.50
Annualised volatility10.5%7.1%7.8%7.3%10.5%9.1%
IR− 0.10− 0.13− 0.22−0.22− 0.36− 0.20
Maximum drawdown9.7%6.3%9.2%5.3%8.8%9.5%
Hit Rate52.4%50.8%50.0%55.6%47.6%50.8%
Average monthly turnover13%8%16%9.7%15%12%
Returns after costs0.5%0.1%0.2%−0.5%−1.9%−0.4%
% reduction in returns−146%−111%−112%−71.0%−48%−76.1%
Panel C: Top–bottom quintiles Q5–Q1
 Forward PEBook yieldFair PECFRIMROM
Annualised returns7.05%9.07%8.34%9.90%11.94%10.98%
t-Statistic1.521.841.802.212.532.25
Annualised volatility17.8%18.0%16.9%15.5%16.0%16.9%
IR0.400.500.490.640.750.65
Maximum drawdown−19.1%−15.6%−16.8%−14.1%−16.5%−17.9%
Hit Rate55.6%59.7%62.1%58.1%60.5%56.5%
Average monthly turnover28%18%33%19.2%31%26%
Returns after costs3.6%6.7%4.1%7.4%7.9%7.5%
% reduction in returns−49.3%−26.1%−50.8%−25.3%−33.7%−31.5%

On a long-only basis, portfolios formulated with the sophisticated models, that is, the RIM and ROM, outperform those based on simple models, with spread returns, on average, 9.02% and 9.92% annualised with t-statistics of 2.58 and 2.40, respectively. Here CF provides a lower but statistically significant annualised return that is 8.70% with a t-statistic of 2.20. A less statistically significant return is obtained through the book yield; that is, the t-statistic is just 1.84, while the average return obtained through the forward PE and fair PE is only marginally significant, that is, the t-statistics are 1.74 and 1.86, respectively. In terms of the risk–return relation, RIM- and ROM-based strategies portfolios are better by a factor of almost 1.5 than all the strategies based on a simple valuation model, except for CF, which is inferior by only about 10% and 20% to the ROM and RIM, respectively.

The results in the bottom quintile are far from impressive across all valuation metrics. The RIM shows the best predictability, but, in general, the wealth curve for all valuation factors suggests that all models are poor in identifying overvalued companies in the short-term investment horizon. Moreover, the returns for all strategies are statistically insignificant. As per our earlier discussion, the ROM's incremental predictability over the RIM (as well as the other valuation models) through the incorporation of the abandonment option should be more prominent for financially distressed firms or firms with negative earnings or residual income with good NAV. In theory, the ROM would identify these firms, classify them out of the bottom quintile, and, if the model's prediction is correct, outperform all other models for the firms in the bottom quintile. Although statistically we observe no difference, the difference we observe economically supports this argument for all alternative specifications, except for the RIM. This is in contrast with the ROM's prediction and the empirical findings of Hwang and Sohn (2010), who find that the ROM's outperformance is stronger for firms with residual income below book value. We believe this finding is related to the nature of the firms in our sample, that is, typically long-lived companies with financial performance not in either of the extremes of residual income versus book value. Our correlation analysis on the returns of the hedge portfolios obtained through the different valuation methods in Section 0 below supports this argument.

When we assess the performance of hedge portfolios, we conclude that RIM- and ROM-based portfolios rank higher than the portfolios obtained through the simple measures. Portfolios constructed on the basis of the RIM have the highest annual return, about 1% and 2% higher than the second and third best factors, that is, the ROM and CF, respectively. In terms of return volatility, CF ranks best, with only marginally lower volatility than the RIM. Thus, with respect to return–risk RIM ranks best, presenting with an IR of 0.75 (relative to 0.65 and 0.64 for the second and third best, that is, the ROM and CF, respectively).

The results based on quarterly rebalancing (untabulated) are qualitatively similar to those based on monthly rebalancing.5 However, we need to interpret the results with caution, given the relatively small number of observations, 42 in total.

To obtain additional insight in our baseline monthly rebalancing investigation with regard to what drives the return of each portfolio and to rule out the portfolio return being pure compensation for known risks, we calculate the risk-adjusted return of the long–short portfolio (Q5–Q1) as in Fama and French (1993) and Carhart (1997). For this analysis, Table 2 reports alphas and risk loading estimates for a time series regression of the long–short portfolio (Q5–Q1) monthly return on market risk, the value and size premiums, and momentum returns. The results indicate that the long–short (Q5–Q1) risk-adjusted returns (alphas) of the RIM and ROM are significant at the 5% and 10% significance levels, respectively. We interpret this finding as an indication that the RIM and ROM are able to unhide value not captured by traditional measures, such as those used to construct value-to-growth ratio indices. Another observation worth highlighting is that the size risk factor is not statistically significant for any of the valuation measures except CF. The implication is that the strategies’ positive returns are not a result of small cap exposure, and therefore trading costs may not significantly reduce their after-cost returns.

Table 2. Risk-adjusted returns for the Q5–Q1 portfolio This table presents the monthly risk-adjusted returns and risks of the long–short portfolio (Q5–Q1) sorted monthly on the basis of alternative valuation measures for the period January 2000 to April 2010 (total of 124 observations). These measures include the forward PE, which is the next 12  months’ (weighted between FY1 and FY2) I/B/E/S consensus EPS divided by the month-end closing price; book yield, which is the last 12  months’ (weighted between FY0 and FY-1) reported BPS divided by the month-end closing price; the fair PE, that is, the ratio of the forecast PE to the market PE, which is a theoretical PE estimated by a cross-sectional model (similar to a Gordon growth framework), using expected earnings growth and market cap (risk and liquidity proxy) as inputs; CF is the next 12  months’ (weighted between FY1 and FY2) I/B/E/S consensus cash flow per share forecast divided by the month-end closing price; RIM, which is obtained through the RIM (see Ohlson, 19911995; Feltham and Ohlson, 1995); and ROM, which is obtained through the ROM (see Hwang and Sohn, 2010). Risk adjustment is pursued through the Fama and French (1993) and Carhart (1997) factor models that account for market risk, the value/growth, large/small cap, and momentum premiums. Market risk is measured through the MSCI Europe Index returns, value premiums are computed through the spread of the MSCI Europe Value and Growth Index returns, and size premiums are computed through the spread of the MSCI Europe Large Cap and Small Cap Index returns. Momentum returns are proxied by the spread return difference between the high and low portfolios, formulated on the basis of the 12-month price return of companies in the MSCI Europe Index. The table reports alphas (alpha), and loading estimates with respect to the market-wide equity risk (market), the value premium (value growth), the size premium (large–small cap), and the momentum return (momentum). Here ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively.
 AlphaMarketValue growthLarge–small capMomentum
ROM0.66%**−0.29***1.00***−0.05***−0.32***
Forward PE0.28%**−0.17***1.08***−0.19***−0.12***
Book yield0.40%**−0.07***1.23***−0.20***−0.32***
CF0.36%0.060.63***−0.54***−0.24***
RIM0.86%**−0.36***0.69***0.06***−0.13***
Fair PE0.36%**−0.04***1.07***−0.23***−0.21***

In summary, our analysis of the signal quality concludes that sophisticated models, that is, the RIM and ROM, produce equity valuations that are better able to predict the cross section of short-term future equity returns than simple models. The superior quality of the signal translates to an improvement in the annualised monthly return that can be up to about 5% – that is, the difference in the hedge portfolio return between the RIM and forward PE – or about 50% or more in terms of IR – that is, the difference in the IRs of RIM- and ROM-based portfolios relative to all others, except CF, where the improvements are of 20% (in the long-only and long–short portfolios) and 10% (in the long-only portfolio), respectively. All models have been poor in identifying overvalued companies in the short-term investment horizon. Moreover, the ranking of the models does not change when we consider quarterly rebalancing.

5.2 Signal strength

We examine the valuation factor strength or robustness from three different angles. First, to address concerns suggesting that transaction costs may have a significant impact on the profitability of possibly high turnover strategies, we examine how the turnover of the portfolios based on each valuation measure affects returns. Second, we carry out an analysis of a typical measure of ex post signal strength evaluation, the IC. The IC enables us to see if the signal is consistent across the cross section of stocks that we are considering at a point in time. Third, given the economic uniqueness of the global financial crisis (GFC) and its tremendous impact on valuation-based equity strategies, we study how each valuation measure performs both before and after the crisis.

Determining a strategy's likely transaction costs within an empirical framework is difficult. Transaction costs are a function of fund size, the alpha signal, and market conditions, and in some cases they are firm specific. However, we can make some broad assumptions to gauge a strategy's potential costs. The key variable within this research framework is turnover. We assume in our analysis, that all things being equal, the higher the turnover, the higher the transaction costs. While we recognise that transaction costs are nonlinear relative to transaction size and fund size, for this exercise and for simplicity, we apply a fixed, ‘all-in’ transaction cost to our analysis. We simply assume transaction costs are a function of turnover6 and subtract the cost of trading from the strategy's monthly return. For the fixed transaction cost through the period of analysis we use 50 basis points (bps), which comprises commission, spread, borrow (where applicable), and market impact costs.7

Table 1 shows the results of this analysis and, in particular, the pre- (headlined annualised returns) and post-transaction cost (headlined returns after costs) strategy returns. This analysis concludes that the ROM presents with the third lowest average turnover (after book yield and CF), and the RIM with the second highest (the fair PE is associated with the highest-turnover portfolio strategy). In terms of the long–short portfolios, although the RIM, in general, has the highest turnover, given our transaction cost assumptions, the higher returns of the RIM-based strategy more than compensate for this. The long–short returns are higher for the RIM, although not materially different from those for the ROM and CF, that is, 7.9% versus 7.5% and 7.4%, respectively. Focusing on the top quintile (long), we conclude that, after costs, the ROM has the highest risk-adjusted returns. The returns in the bottom quintile (short), although statistically insignificant, are more compelling for the RIM and almost indifferent across all other models.

Our second test of signal strength involves the analysis of the IC. We calculate the rank IC for all valuation measures. Table 3 shows the average IC over the period of analysis, its standard deviation, and the t-statistic corresponding to the t-test of the null hypothesis that the mean IC is zero. While, in general, the ICs are relatively low, most of this can be attributed to the deep value underperformance through the GFC. Restricting our analysis to the period spanning 2000 to mid-2007 (available upon request), most of the ICs are twice what they are for the entire sample. This having been said, it is also apparent that the ICs were already falling prior to the GFC (see Jones, 2010, for a detailed analysis of falling IRs).

Table 3. Rank ICs: descriptive statistics This table presents statistics of the rank IC, which is the cross-sectional correlation of the ranks of stocks upon portfolio formation and the next rebalancing. We present the average IC, its standard deviation, and the t-statistic for a t-test of the null hypothesis that the mean IC is zero for stocks sorted monthly on the basis of alternative valuation measures for the period January 2000 to April 2010 (total of 124 observations). These measures include the forward PE, which is the next 12 months’ (weighted between FY1 and FY2) I/B/E/S consensus EPS divided by the month-end closing price; book yield, which is the last 12 months’ (weighted between FY0 and FY-1) reported BPS divided by the month-end closing price; the fair PE, that is, the ratio of the forecast PE to the market PE, which is a theoretical PE estimated by a cross-sectional model (similar to a Gordon growth framework), using expected earnings growth and market cap (risk and liquidity proxy) as inputs; CF is the next 12 months’ (weighted between FY1 and FY2) I/B/E/S consensus cash flow per share forecast divided by the month-end closing price; RIM, which is obtained through the RIM (see Ohlson, 1991, 1995; Feltham and Ohlson, 1995); and ROM, which is obtained through the ROM (see Hwang and Sohn, 2010).
 AverageSt. Dev.t-Statistic
Forward PE3.9%14.7%2.93
Book yield2.5%14.4%1.90
Fair PE3.6%13.8%2.88
CF3.5%13.8%2.84
RIM4.1%14.3%3.16
ROM3.6%13.6%2.93

We complement this analysis with an investigation of the IC decay. Traditionally, accounting-based valuation methodologies such as the RIM and ROM are advocated to provide a better value estimate when value needs to be measured in the mid to long term. To this extent, prior studies such as Hwang and Sohn (2010) and Dissanaike and Lim (2010) have shown (or have challenged) the merits of these methodologies for portfolio strategies that rebalance yearly or every two or three years. To investigate this angle, we calculate the IC decay. As in Section 0 we calculate the rank correlation in the cross section of stocks for up to 24 months in the future. The analysis (available upon request) indicates that the IC decay of the forward PE is more pronounced, whereas the ROM seems to possess ICs that maintain their magnitude for up to 24 months after portfolio formation.

Our last test explores the predictability of each valuation measure in the pre- and post-GFC period. Our aim is to infer if our conclusions with respect to the relative ranking of valuation measures are driven by prevailing market conditions. We focus on the returns of the long–short portfolio. Table 4 reports the results of our investigation, where we split the sample into two periods: 2000 to 2006 and 2007 to 2010.8 As shown, the difference in the spread returns of the value strategies between the two periods is stark. We observe that the ROM does fairly well in capturing value in the pre-crisis period, although it does not rank best among its peers. Simple valuation measures perform better on a risk-adjusted basis over this period. In the second period the ROM also does fairly well, in the sense that it is (along with the RIM and CF) among the least value-destroying factors. Interestingly, the factors that perform strongly in the pre-crisis period – that is, the fair PE, forward PE, and book yield – are the worst performers in the crisis and post-crisis periods. Therefore, this analysis concludes that the RIM and ROM, in particular, maintain a reasonably good and consistent rank during the entire study period.9 Notably, CF also does.

Table 4. Hedge portfolio (Q5–Q1) performance relative to the GFC: 2000–2006 versus 2007–2010 This table presents the annualised spread returns and performance metrics of portfolios sorted on the basis of alternative valuation measures for the period January 2000 to April 2010 on a monthly basis (total of 124 observations). These measures include the forward PE, which is the next 12 months’ (weighted between FY1 and FY2) I/B/E/S consensus EPS divided by the month-end closing price; book yield, which is the last 12 months’ (weighted between FY0 and FY-1) reported BPS divided by the month-end closing price; the fair PE, that is, the ratio of the forecast PE to the market PE, which is a theoretical PE estimated by a cross-sectional model (similar to a Gordon growth framework), using expected earnings growth and market cap (risk and liquidity proxy) as inputs; CF is the next 12 months’ (weighted between FY1 and FY2) I/B/E/S consensus cash flow per share forecast divided by the month-end closing price; RIM, which is obtained through the RIM (see Ohlson, 1991, 1995; Feltham and Ohlson, 1995); and ROM, which is obtained through the ROM (see Hwang and Sohn, 2010). The top quintile portfolio, that is, Q5, comprises stocks with high earnings yield, while the bottom quintile, that is, Q1, comprises stocks with low earnings yield. This table reports results for the long–short portfolio Q5–Q1. Here IR is the ratio of the annualised return over the annualised volatility. Maximum drawdown is the minimum return obtained for a strategy in the study period. The hit rate is the percentage of correct predictions.
Panel A: Analysis covering the period January 2000 to December 2006
 Forward PEBook yieldFair PECFRIMROM
Annualised returns20.35%17.91%19.79%17.23%19.38%19.45%
t-statistic3.242.913.463.382.983.17
Annualised volatility16.3%14.3%14.7%13.2%17.1%16.0%
IR1.251.251.341.311.131.22
Maximum drawdown−19.1%−15.6%−16.8%−14.1%−16.5%−17.9%
Hit rate65.5%72.6%71.4%66.7%63.1%64.3%
Panel B: Analysis covering the period January 2007 to April 2010
 Forward PEBook yieldFair PECFRIMROM
Annualised returns−16.28%−7.39%−12.26%−4.04%−2.20%−4.91%
t-statistic−1.55−0.40−1.05−0.23−0.21−0.35
Annualised volatility18.7%23.5%19.5%19.0%12.5%18.1%
IR− 0.87− 0.31− 0.63−0.21− 0.18− 0.27
Maximum drawdown−14.1%−10.9%−13.6%−9.2%−7.5%−9.3%
Hit rate35.0%32.5%42.5%40.0%55.0%40.0%

In summary, our analysis of the signal strength concludes that sophisticated models, that is, the RIM and ROM, produce equity valuations that are better able to predict the cross section of short-term future equity returns than simple models, even after accounting for transaction costs. When focussing on the ICs, we conclude that, although the difference is only marginal, sophisticated models and, in particular, the RIM outperform all others. Finally, when we examine the efficacy of the valuation measures in the pre- and post-crisis periods, we conclude that the sophisticated measures present with a relatively consistent ranking, particularly the ROM.

5.3 Signal uniqueness

The ROM as well as the other value factors may be systematically selecting attributes of the stocks that are not related to stock-specific value. This section investigates the influence that sector exposure has on the returns for the various strategies we are considering. We also address the concern that signals overlap as well as correlate with other predictive factors.

First, we repeat our analysis but neutralise sector biases in our results. The methodology we employ to create sector-neutral portfolios is straightforward. We create sector-relative portfolios by sorting the stocks in each of the 10 MSCI sectors by their respective valuation metric, grouping the stocks within each sector into quintiles and then aggregating the stocks from the same quintiles. Thus, the first aggregated, or sector-relative, quintile contains the stocks from the first quintile of each sector, while the fifth sector-relative quintile contains the stocks from the fifth quintile of each sector. We then equally weight the sector-relative quintiles and measure their total returns during the following month. Table 5 shows the results from this analysis. The returns of the hedge portfolios fall for all factors except for the fair PE. These changes have little impact on the relative rank of each model. The varying degrees of influence that sector neutrality has on the performance of each portfolio are, however, a surprise.

Table 5. Sector-neutral portfolio returns This table presents the annualised spread returns and performance metrics of portfolios sorted on the basis of alternative valuation measures for the period January 2000 to April 2010 on a monthly basis (total of 124 observations). These measures include the forward PE, which is the next 12 months’ (weighted between FY1 and FY2) I/B/E/S consensus EPS divided by the month-end closing price; book yield, which is the last 12 months’ (weighted between FY0 and FY-1) reported BPS divided by the month-end closing price; the fair PE, that is, the ratio of the forecast PE to the market PE, which is a theoretical PE estimated by a cross-sectional model (similar to a Gordon growth framework), using expected earnings growth and market cap (risk and liquidity proxy) as inputs; CF is the next 12 months’ (weighted between FY1 and FY2) I/B/E/S consensus cash flow per share forecast divided by the month-end closing price; RIM, which is obtained through the RIM (see Ohlson, 1991, 1995; Feltham and Ohlson, 1995); and ROM, which is obtained through the ROM (see Hwang and Sohn, 2010). The top quintile portfolio, that is, Q5, comprises stocks with high earnings yield, while the bottom quintile, that is, Q1, comprises stocks with low earnings yield. The returns of the bottom and the top quintiles are spreads over the market returns, that is, over the MSCI Europe Index returns. Portfolio returns are neutralised with respect to sector biases. We create sector-relative portfolios by sorting the stocks in each of the 10 MSCI sectors by their respective valuation metric, grouping the stocks within each sector into quintiles, and then aggregating the stocks from the same quintiles. Thus, the first aggregated, or sector-relative, quintile contains the stocks from the first quintile of each sector, while the fifth sector-relative quintile contains the stocks from the fifth quintile of each sector. We then equally weight the sector-relative quintiles and measure their total returns during the following month. Here IR is the ratio of the annualised return over the annualised volatility. Maximum drawdown is the minimum return obtained for a strategy in the study period. The hit rate is the percentage of correct predictions.
Panel A: Top quintile Q5 – market
 Forward PEBook yieldFair PECFRIMROM
Annualised returns9.01%9.43%9.59%7.64%10.37%10.00%
t-statistic2.542.273.012.242.932.63
Annualised volatility11.8%12.4%10.4%11.9%11.9%12.7%
IR0.760.760.920.640.870.79
Maximum drawdown−10.5%−10.1%−9.9%−9.3%−10.0%−9.6%
Hit rate61.3%61.3%63.7%63.7%66.9%62.1%
Panel B: Top–bottom quintiles Q5–Q1
 Forward PEBook YieldFair PECFRIMROM
Annualised returns7.04%8.31%9.44%6.23%10.77%10.38%
t-statistic2.052.272.841.812.932.59
Annualised volatility11.8%12.4%10.9%12.5%12.0%13.4%
IR0.600.670.870.500.890.78
Maximum drawdown−11.5%−11.7%−9.4%−11.6%−11.8%−13.2%
Hit rate59.7%56.5%65.3%55.6%69.4%64.5%

For the long–short portfolio, the largest deviations in returns are observed for CF – with a drop from 9.90% to 6.23%, that is, a relative change of −37.07% – and for the fair PE – with a relative change of 13.19%, from 8.34% to 9.44%. Significant deviations are observed in return volatility, which, combined with the deviations in returns, result in substantial changes in IRs. The changes in IRs are 0.2, 0.17, 0.38, −0.14, 0.14, and 0.13 for the forward PE, book yield, fair PE, CF, RIM, and ROM, respectively. These figures represent changes of 50%, 34%, 78%, −22%, 19%, and 20% of the initial IRs. Changes are clearly more pronounced for the forward PE, book yield, fair PE, and, to a lesser extent, CF. This observation suggests that in the unconstrained space, the difference in performance arises from the capacity of models such as the RIM and ROM to correctly identify stock-specific value and efficiently incorporate sector-wide risk in valuations. In line with this observation, from a different angle, though, Kim et al. (2009) provide theoretical motivation and empirical evidence on why an implementation of the RIM that accounts more explicitly for industry-wide valuations improves stock-specific valuations.

Our second concern in the context of examining signal uniqueness relates to the correlation of the returns for portfolios obtained with the different valuation measures, as well as with the returns of portfolios obtained through other known predictive factors. The interaction of factors has important implications when constructing multifactor models in portfolio management. So far our analysis has been completed on a univariate basis. We now consider the correlation in the returns of long–short portfolios based on the simple and sophisticated valuation metrics and the returns of long–short portfolios based on other common styles/factors. In general, we expect the five valuation signals to be highly correlated and, as we see in Table 6, all of them are indeed highly and positively correlated.

Table 6. Correlations between valuation signals and other signals This table presents the correlations between the monthly returns of the long–short portfolio obtained with monthly rebalancing through alternative valuation measures, as well with other known quantitative equity portfolio management signals for the period January 2000 to April 2010 (total of 124 observations). The valuation measures include the forward PE, which is the next 12 months’ (weighted between FY1 and FY2) I/B/E/S consensus EPS divided by the month-end closing price; book yield, which is the last 12 months’ (weighted between FY0 and FY-1) reported BPS divided by the month-end closing price; the fair PE, that is, the ratio of the forecast PE to the market PE, which is a theoretical PE estimated by a cross-sectional model (similar to a Gordon growth framework), using expected earnings growth and market cap (risk and liquidity proxy) as inputs; CF is the next 12 months’ (weighted between FY1 and FY2) I/B/E/S consensus cash flow per share forecast divided by the month-end closing price; RIM, which is obtained through the RIM (see Ohlson, 1991, 1995; Feltham and Ohlson, 1995); and ROM, which is obtained through the ROM (see Hwang and Sohn, 2010). Price momentum is a long–short portfolio strategy based on stocks’ total return over the previous 12 months. Estimates momentum is a long–short portfolio strategy based on stocks’ average of analysts’ upward revisions minus downward revisions for FY1 and FY2. Quality is a long–short portfolio strategy based on accruals. Risk is a long–short portfolio strategy based on stocks’ betas with the MSCI Europe, computed using monthly prices over 50 months. Growth is a long–short portfolio strategy based on stocks’ growth of reported earnings.
 Forward PEBook yieldFair PECFRIMROM
Forward PE1     
Book yield0.861    
Fair PE0.880.761   
CF0.830.900.881  
RIM0.870.840.600.641 
ROM0.860.930.840.860.741
Price momentum-0.45-0.30-0.59-0.49-0.12-0.49
Estimates momentum-0.30-0.16-0.49-0.400.05-0.32
High quality-0.25-0.10-0.48-0.370.07-0.28
Low risk-0.25-0.13-0.54-0.470.14-0.37
Growth-0.65-0.60-0.70-0.65-0.54-0.60

This analysis shows some interesting results. First, the ROM presents with a high correlation with all other factors, 0.74 being the lowest, with the RIM. The RIM, on the other hand, correlates well with the forward PE and book yield, with correlation coefficients of 0.87 and 0.84, respectively. The RIM correlates poorly, however, with the fair PE and CF. The correlation of the RIM with book yield, which is 0.84, taken together with the correlation of the ROM with book yield, which is 0.93, can partly enlighten us on the residual income to book value ratio of the firms in our sample. A negative residual income should result in different RIM versus book value valuations, whereas it should not have an impact on the ROM versus book value valuations; even when the residual income is negative, the value of the firm within the ROM will not fall below the book value.

The magnitudes of these correlations are consistent with the hypothesis that only a small fraction of firms in our sample fall in the (extreme) negative residual income domain, where Hwang and Sohn (2010) find the strongest incremental predictability of the ROM over the RIM (Hwang and Sohn, 2010, Table 8, pp. 393–394). This, we believe, explains why the RIM and ROM do not deliver significantly different valuations for the bottom quintile portfolio firms, or at least different to the extent that they dramatically differentiate the bottom quintile portfolio performance. The RIM presents with the lowest pairwise correlations with the fair PE and CF.

Table 7. Spread returns for portfolios formulated on combined value-to-momentum ratio signals This table presents the annualised spread returns and performance metrics of portfolios sorted on the basis of alternative valuation measures and momentum for the period January 2000 to April 2010 on a monthly basis (total of 124 observations). These measures include the forward PE, which is the next 12 months’ (weighted between FY1 and FY2) I/B/E/S consensus EPS divided by the month-end closing price; book yield, which is the last 12 months’ (weighted between FY0 and FY-1) reported BPS divided by the month-end closing price; the fair PE, that is, the ratio of the forecast PE to the market PE, which is a theoretical PE estimated by a cross-sectional model (similar to a Gordon growth framework), using expected earnings growth and market cap (risk and liquidity proxy) as inputs; CF is the next 12 months’ (weighted between FY1 and FY2) I/B/E/S consensus cash flow per share forecast divided by the month-end closing price; RIM, which is obtained through the RIM (see Ohlson, 19911995; Feltham and Ohlson, 1995); and ROM, which is obtained through the ROM (see Hwang and Sohn, 2010). The top quintile portfolio, that is, Q5, comprises stocks with high earnings yield, while the bottom quintile, that is, Q1, comprises stocks with low earnings yield. The returns of the bottom and top quintiles are spreads over the market returns, that is, over the MSCI Europe Index returns. Here IR is the ratio of the annualised return over the annualised volatility. Maximum drawdown is the minimum return obtained for a strategy in the study period. The hit rate is the percentage of correct predictions.
Panel A: Top quintile Q5 – market
 Forward PEBook yieldFair PECFRIMROM
Annualised returns7.9%8.2%7.2%7.64%8.2%8.2%
t-statistic2.581.832.622.242.892.93
Annualised volatility10.2%14.6%9.1%11.9%18.2%9.2%
IR0.780.560.800.640.450.90
Maximum drawdown-12.0%-9.6%-12.1%-9.3%-11.1%-9.8%
Hit rate66.1%69.4%70.2%63.7%68.5%69.4%
Panel B: bottom quintile Q1 – market
 Forward PEBook yieldFair PECFRIMROM
Annualised returns0.2%-0.6%-2.1%1.13%-1.8%-2.9%
t-statistic0.250.00-0.500.81-0.23-0.60
Annualised volatility13.3%11.3%10.2%5.1%9.2%12.0%
IR0.01-0.06-0.200.22-0.20-0.24
Maximum drawdown14.4%11.4%9.5%3.5%13.75%11.8%
Hit rate49.2%50.8%45.2%58.1%46.8%48.4%
Panel C: Top–bottom quintiles Q5–Q1
 Forward PEBook yieldFair PECFRIMROM
Annualised returns8.2%7.5%5.8%6.23%7.8%9.7%
t-statistic1.961.831.311.811.632.12
Annualised volatility14.7%14.6%17.5%12.5%18.2%16.1%
IR0.560.520.330.500.430.61
Maximum drawdown-12.6%-13.6%-13.1%-11.6%-17.6%-10.5%
Hit rate59.7%56.5%58.9%55.6%58.1%61.3%
Table 8. Rank ICs when combined with momentum: Descriptive statistics This table presents statistics of the rank IC, which is the cross-sectional correlation of the ranks of stocks upon portfolio formation and the next rebalancing in the subsequent month. We present the average IC, its standard deviation, and the t-statistic for a t-test of the null hypothesis that the mean IC is zero for stocks sorted on the basis of alternative valuation measures and momentum for the period January 2000 to April 2010 (total of 124 observations). These measures include the forward PE, which is the next 12 months’ (weighted between FY1 and FY2) I/B/E/S consensus EPS divided by the month-end closing price; book yield, which is the last 12 months’ (weighted between FY0 and FY-1) reported BPS divided by the month-end closing price; the fair PE, that is, the ratio of the forecast PE to the market PE, which is a theoretical PE estimated by a cross-sectional model (similar to a Gordon growth framework), using expected earnings growth and market cap (risk and liquidity proxy) as inputs; CF is the next 12 months’ (weighted between FY1 and FY2) I/B/E/S consensus cash flow per share forecast divided by the month-end closing price; RIM, which is obtained through the RIM (see Ohlson, 1991, 1995; Feltham and Ohlson, 1995); and ROM, which is obtained through the ROM (see Hwang and Sohn, 2010).
 AverageSt. Dev.t-statistic
Forward PE4.0%16.0%2.96
Book yield3.4%13.9%2.44
Fair PE4.3%14.1%2.53
CF3.5%14.5%1.96
RIM4.7%17.8%1.81
ROM4.4%15.9%2.18

Taken together, these figures suggest that the ROM, forward PE, book yield, and fair PE share similar correlation attributes between them, as well as with other quantitative factors. The RIM seems to be generally a little off the line. The differences are almost marginal with regard to among-value-factor correlations, but they are significant with regard to the ex–value-factor correlations. The highly negative correlation with the price momentum factor motivates further investigation of the diversification and/or other benefits of the combination of the two signals.

To test the implications of the latter observation, we conduct a simple test, creating an equal-weighted score between value and momentum and empirically testing this strategy. For momentum we simply use the 12-month price momentum (see the Appendix). Table 7 shows the results from this analysis. While this analysis may not be very detailed, it does show that the ROM, when combined with price momentum, has the best risk-adjusted results, and also some of the highest hit rates. Additionally, the average rank IC for the RIM is the highest, with the ROM's IC being second best relative to the other valuation factors (see Table 8).

In summary, our analysis of the signal uniqueness concludes that sophisticated models, that is, the RIM and ROM, are less vulnerable to sector biases. When sector bias is neutralised. the changes to the IRs of the portfolios obtained through simple valuation models are up to 78%, while those for sophisticated models are 20% at most. As expected, the correlations of the hedge portfolio returns are almost identical among value factors, but there are differences with regard to the ex–value-factor correlations. The ROM and, to a lesser extent, the RIM appear to be the most appropriate valuation measures to combine with momentum signals.

6. Implications

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Literature Review
  5. 3. Valuation Measures
  6. 4. Data
  7. 5. Empirical Analysis
  8. 6. Implications
  9. 7. Conclusion
  10. Appendix
  11. References

Our analysis has thus far provided important insights for equity portfolio managers engaging in valuation-related investment strategies. First, we find that, to the extent that our transaction cost hypothesis of fixed 50 bps is representative, the benefits of using the RIM or ROM over CF are marginal, that is, 0.4% or 0.1%, respectively, per annum in the hedge portfolio and 0.5% for the ROM in the long–only portfolio (Table 1). Hence, the benefits of the incremental predictability of sophisticated valuation models in the short term may deteriorate due to the aggressive rebalancing of the respective portfolios. Apparently, less conservative transaction cost assumptions favour sophisticated models even more. The comparison of the RIM and ROM with all other valuation measures, however, suggests that the incremental predictability of sophisticated models is economically significant.

To this point, our conclusions have not been significantly different from those presented in Dissanaike and Lim (2010). However, when we turn to the sector-neutral portfolio results (Table 5), we observe a dramatic change in the return of the majority of the strategies, and while the RIM and ROM rank first and second, respectively, CF ranks at the bottom, for both long–only and long–short investment strategies. Moreover, when we examine a popular investment strategy that combines valuation and momentum signals, we conclude that the ROM has the best risk-adjusted results.

Our analysis also provides important insights for the information content of the alternative valuation measures we consider as proxies for fundamental value. Our study sheds light on the issue of how the valuation measure's information content affected by the timing of the data used for its estimation. Book yield relies on reported book values. The forward PE and CF rely on weighted one-period-ahead forecasts of earnings and cash flow. The fair PE utilises two-period-ahead earnings forecasts, as well as current market capitalisation and current PE values. The RIM and ROM rely on reported and forecasted book values, forecasted earnings, and costs of equity estimated from historical data at the sector level. Hence the timing of the data required is not common across valuation measures.

Our analysis suggests that current information, that is, book values, is more reliable than information forecast one period ahead, that is, one-period-ahead forecasts of earnings or cash flow when either one is used in isolation. This is evident in the comparison of book yield with the forward PE and CF returns in Table 5. Dissanaike and Lim (2010) reach a similar conclusion when they compare measures that rely on current versus forecast earnings. Current information is also found to be very important in market turning points (see 9).

The fact that current information is important for short-term return predictability may, to some extent, explain part of the incremental predictability of the RIM and ROM. Dissanaike and Lim (2010) stress that the fact that the RIM relies on two-year-ahead earnings forecasts reduces the potential impact of noise and shocks in the one-year-ahead earnings (and cash flow) forecasts. This, we believe, can explain part of the superior information content of the RIM and ROM over the forward PE and CF (Table 5). If the reduced estimation error in the earnings forecasts were the only reason for the RIM and ROM's incremental predictability, we would expect the fair PE – which relies on two-year-ahead earnings forecasts and the current PE – to also be superior to book yield. This is not supported by our analysis for the sector-neutral portfolios after transaction costs, which suggests that current book value may contain important information too. Hence, we argue that the RIM and ROM's superior information content should be attributed, as Richardson et al. (2010) also state, to the intuitive functional form of the models linking fundamental value with expected profitability (positively) and asset or book growth (negatively).

Our study also contributes to the literature studying the impact of analysts’ earnings forecasts on firm valuations. Cheng (2005), for example, finds that analysts do not fully incorporate industry characteristics in their forecasts, which potentially has important implications for accounting-based valuations. Kim et al. (2009) provide theoretical motivation and empirical evidence on why an implementation of the RIM that accounts more explicitly for industry-wide valuations improves stock-specific valuations. We do not explicitly address such issues in our analysis. However, we observe that the superior performance of the RIM- or ROM-based portfolios becomes more pronounced than that of other models when we control for sector biases (compare Table 1 and Table 5). We argue that this evidence supports the conjecture that the RIM and ROM are more capable than all other models that use earnings forecasts of correctly identifying stock-specific value and efficiently incorporating sector-wide risk in valuations.

7. Conclusion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Literature Review
  5. 3. Valuation Measures
  6. 4. Data
  7. 5. Empirical Analysis
  8. 6. Implications
  9. 7. Conclusion
  10. Appendix
  11. References

Valuation signals have been among the most popular equity selection signals among equity portfolio managers and have recently attracted significant interest from cross-asset managers. Given the large variation in techniques and theories with regard to how value is measured and in light of recent developments in the academic literature, this article carefully examines the efficacy of alternative valuation measures.

Our approach to the problem takes a more practical stance than that of the current literature and, while we carry out an empirical investigation within an academically rigorous framework, we set up our experiments in a way that is more appealing for an equity portfolio manager. In particular, we empirically investigate the predictive ability of sophisticated and simple valuation measures over the cross section of future equity returns, but we a) focus on the constituents of a broadly used benchmark index, that is, the MSCI Europe Index, to replicate what a typical manager would do; b) ‘shorten’ the investment horizon to comply with standard practice; c) consider a pan-European investment space; d) use a large number of ‘goodness’ metrics to extend our insights and judge performance within a standard industry setting; e) carry out a comprehensive robustness analysis for different market regimes, as well as for the impact of sector or risk attributes; and f) investigate the correlations and interactions between several value and non-value factors.

We consider a comprehensive cross section of alternative measures that includes the forward PE, book yield, fair PE, CF, RIM, and ROM. Equity portfolios obtained through the RIM or ROM stock rankings outperform all others on a long-only and a long–short basis, even after accounting for transaction costs. They performed rather consistently following the massive value sell-off in August 2007 on and, in particular, during fall 2008, and stood out in the value rally starting in March 2009. More importantly, we find that RIM- and ROM-based portfolios are less affected by sector bias, suggesting that these measures are able to correctly identify stock-specific value and more efficiently incorporate sector-wide risk in their valuations. Overall, we find that sophisticated models such as the RIM and ROM within an investment strategy have merit over their simple peers.

However, our study has a limitation. Sophisticated models need extensive data for practical implementation. In case the investment universe is broader than a widely used benchmark such as the MSCI Europe Index, practical implementation may suffer from survivorship bias, since we consider only firms that satisfy the criteria of inclusion in the MSCI Europe Index. Therefore, to the extent that portfolio managers include short-lived stocks in their portfolios, the reported superiority of the sophisticated valuations can be diminished.

Appendix

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Literature Review
  5. 3. Valuation Measures
  6. 4. Data
  7. 5. Empirical Analysis
  8. 6. Implications
  9. 7. Conclusion
  10. Appendix
  11. References

A1. Cost of equity

We calculate the cost of equity following Fama and French (1997), who derive cost of equity forecasts through the equation

  • display math

where bi, si, and hi are obtained from the regression

  • display math

In this regression Ri is the return on industry i. We estimate this regression on the basis of a five-year rolling window of monthly returns. Fama and French (1997) highlight that ordinary least squares estimates of the regression coefficients can, in some cases, be imprecise, thus delivering unreasonable estimates of an industry's cost of equity. The authors advocate that a Bayesian shrinkage method can mitigate such problems. To deal with this issue, we use robust linear regression methods that are controlled for biases and inefficiencies that can arise due to outliers and, generally, inconsistencies in the theoretical and empirical distribution of the data. For the estimates of (RM-RF), SMB, and HML, that is market risk, value, and size premiums respectively, we use their average monthly values. To calculate averages, we use an expanding window.

One other issue that arises when the analysis is carried out cross country is what the relevant risk-free rate should be. We used an industry-specific risk-free rate. This is computed by market cap weighting the risk-free rates of a country in line with the contribution of each firm (that comes from a certain country) to the industry's total market cap.

A2. Fair PE model

The theoretical PE is estimated by a cross-sectional model, using expected earnings growth and market capitalisation (risk and liquidity proxy) as inputs. The form of the model is

  • display math

where PE denotes the price-to-earnings ratio, EG is the annualised two-year-forward I/B/E/S EPS growth rate, and MC is the standardised market capitalisation, capped at 3. In addition to these factors, we also include country dummy variables (which, e.g., equal one if they belongs to a particular sector, and zero otherwise) to capture the differences in PE ratios within each European country.

Every month we re-estimate the model using the previous month's data for the independent variables and the current month's data for the dependent variables. We then use this model to forecast the theoretical or fair PE for the next month.

A3. Price momentum (t-statistic methodology)

Our one-year price momentum measure considers the underlying price trend adjusted for volatility. The specific measure is the t-statistic of a trend line slope fitted to logged stock prices, using 260 days of data. Figure 1 shows the price history of company XYZ and a price trend line.

image

Figure 1. Price momentum t-statistic methodology

This figure shows the price history of company XYZ and a line fitted for a period of 260 trading days.

Download figure to PowerPoint

For stocks with a trading history spanning fewer than 260 days, we employ the median price momentum of the universe as a substitute.

  1. 1

    The IR is defined as the portfolio average excess return (over a benchmark) or, in the case of a long–short portfolio, the average hedge return divided by its standard deviation.

  2. 2

    The rank IC is defined as the cross-sectional correlation of the ranks of the stocks with respect to their valuation upon portfolio formation and the ranks of their realised returns between portfolio formation and the next rebalance, upon the next rebalance.

  3. 3

    Dissanaike and Lim (2010) include a number of variants of Ohlson's (1995) model in their list of sophisticated models. Given their conclusion, that the difference with RIM is marginal, we do not include Ohlson's (1995) model or any of its variants in our analysis for the sake of brevity.

  4. 4

    Hwang and Sohn (2010) interpret this option as a call option to facilitate the comparison of the ROM with the RIM or the Burgstahler–Dichev (1997) model. The authors also provide an alternative interpretation of the shareholder option: a put option to sell the capitalised future dividends at the NAV price should they expect the dividends to be below the net asset value; that is, liquidating the firm's assets.

  5. 5

    The ROM and RIM rank first and second, in terms of Q5 returns, respectively, that is, 8.72% and 8.30% annualised. These are buy-and-hold returns over three-month periods annualised/averaged with the rebalancing method. The RIM ranks higher than the ROM in terms of hedge returns, that is, 8.54% versus 7.73% annualised, respectively. The forward PE ranks worst in terms of hedge returns, with 5.53%, while the fair PE ranks second worst, with a hedge return of 5.81%. The IR metric also favours the RIM, with a value of 0.54 annualised, followed by the ROM and CF, with values of 0.46 and 0.42, respectively, for the hedge portfolio. For the Q5 portfolio the RIM's IR is 0.69, followed by the ROM's and BV's IR, which equal 0.63 and 0.58, respectively.

  6. 6

    Total buys and sells divided by the twice the number of stocks in a portfolio. For example, if all the stocks were traded, the turnover would be 100%. However, for high–low strategies, this can be up to 200%.

  7. 7

    We acknowledge that the method we use for dealing with trading costs may be considered a simple one. We believe, however, that our approach is realistic enough to provide reasonable insights on the relative impact of transaction costs on the alternative investment strategies. An exercise that incorporates transaction costs in a more explicit way is beyond the scope of this article. We refer the reader to Agyei-Ampomah (2007) and De Groot et al. (2010) for a detailed investigation along these lines.

  8. 8

    A more detailed view of the pre-/post-crisis performance may be obtained if we alternatively compare several alternative and perhaps shorter time windows. However, we think that an investigation with shorter time windows, although possibly more informative, would also be more challenging in terms of statistical inference, given the limited number of observations. Moreover, from an ex post perspective, the cut-off of 2007 is a reasonable threshold, given the performance of valuation-based strategies before and after that point of time.

  9. 9

    An analysis of the time series of portfolio returns available on request finds that the RIM and ROM performed the best in the period following the large value sell-off from August 2007 on and, in particular, during the third quarter of 2008. This result favours the explanation that typical value factors have been overcrowded and thus hampered during liquidity squeezes. Less attended factors, on the other hand, such as the RIM and ROM, presented some resistance. A second observation is that the RIM, ROM, and book yield were strong performers in the value rally starting in March 2009. The ROM's performance, in particular, implies that it was able to identify companies with the highest price appreciation. Our results suggest that (current) book values – as opposed to earnings expectations – primarily drove prices up in the recent value rally. All three factors, although to varying extents, were able to capture this potential, which can, to some degree, be explained by the financials sector. Moderate earnings (and earnings growth) expectations – published under uncertain global economic conditions in early 2009 – appear not to have sufficed to distinguish the RIM and book yield valuations. The volatility in past residual income, however, seems to have determined significant option values for the ROM, which delivered different and more correct, as it turned out, valuations than the RIM. However, should the RIM and ROM deviate in their valuations? In theory, the top quintiles of the RIM and ROM should present with significant overlap in the firms they comprise, since they contain profitable firms with a low probability of liquidation. In other words, the value of the option should be a smaller fraction of the overall value of the firm; thus, the RIM and ROM valuations should be almost identical in theory. It appears, however, that for the top basket in two instances, that is, March 9 onwards, as mentioned earlier, and (about) January 3 to (about) August 2007, the ROM outperforms the RIM.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Literature Review
  5. 3. Valuation Measures
  6. 4. Data
  7. 5. Empirical Analysis
  8. 6. Implications
  9. 7. Conclusion
  10. Appendix
  11. References
  • Agyei-Ampomah, S., ‘The post-cost profitability of momentum trading strategies: further evidence from the UK’, European Financial Management, Vol. 13 (4), 2007, pp. 776802.
  • Ali, A.L.-S. Hwang, and Trombley, M.A., ‘Residual-income-based valuation predicts future stock returns: evidence on mispricing vs. risk explanations’, Accounting Review, Vol. 78(2), 2003, pp. 37796.
  • Ball, R., ‘Anomalies in relationships between securities’ yields and yield-surrogates’, Journal of Financial Economics, 1978, Vol. 6, pp. 10326.
  • Basu, S., ‘Investment performance of common stocks in relation to their price-earnings ratios: a test of the efficient market hypothesis’, Journal of Finance, Vol. 32(3), 1977, pp. 66382.
  • Bhandari, L.C., ‘Debt/equity ratio and expected common stock returns: empirical evidence’, Journal of Finance, Vol. 43(2), 1988, pp. 50728.
  • Black, F., and Scholes, M., ‘The pricing of options and corporate liabilities’, Journal of Political Economy, Vol. 81, 1973, pp. 63754.
  • Burgstahler, D.C. and Dichev, I.D., ‘Earnings, adaptation and equity value’, Accounting Review, Vol. 72(2), 1997, pp. 187215.
  • Carhart, M.M., ‘On the persistence in mutual fund performance’, Journal of Finance, 52, 1997, pp. 5782.
  • Chalmers, J.M.R., Edelen R.M. and Kadlec, G.B., ‘Transaction-cost expenditures and relative performance of mutual funds’, WFIC Working Paper No. 00–02, (1999).
  • Chan, L.K.C., Hamao Y. and Lakonishok, J., ‘Fundamentals and stock returns in Japan’, Journal of Finance, Vol. 46, 1991, pp. 173964.
  • Cheng, Q., ‘The role of analysts’ forecasts in accounting-based valuation: a critical evaluation’, Review of Accounting Studies, Vol. 10, 2005, pp. 531.
  • Conrad, J., Cooper M. and Kaul, G., ‘Value versus glamour’, Journal of Finance, Vol. 58, 2003, pp. 196997.
  • De Groot, W., Huij J. and Zhou, W., ‘Another look at trading costs and short-term reversal profits’, available at SSRN: http://ssrn.com/abstract=1605049, 2010.
  • Dissanaike, G., ‘On the computation of returns in tests of the stock market overreaction hypothesis’, Journal of Banking and Finance, Vol. 18(6), 1994, pp. 108394.
  • Dissanaike, G. and Lim, K.-H., ‘The sophisticated and the simple: the profitability of contrarian strategies’, European Financial Management, Vol. 16(2), 2010, pp. 22955.
  • Fama, E.F., ‘Stock returns, expected returns, and real activity’, Journal of Finance, Vol. 45(4), 1990, pp. 1089108.
  • Fama, E.F. and French, K.R., ‘Common risk factors in the returns of stocks and bonds’, Journal of Financial Economics, Vol. 33, 1993, pp. 356.
  • Fama, E.F. and French, K.R., ‘Industry costs of equity’, Journal of Financial Economics, Vol. 43, 1997, pp. 153193.
  • Feltham, A.G. and Ohlson, J.A., ‘Valuation and clean surplus accounting for operating and financial activities’, Contemporary Accounting Research, Vol. 11, 1995, pp. 689731.
  • Francis, J., Olsson J.P. and Oswald, D.R., ‘Comparing the accuracy and explainability of dividend, free cash flow, and abnormal earnings equity value estimates’, Journal of Accounting Research, Vol. 38(1), 2000, pp. 4570.
  • Frankel, R. and Lee, C.M., ‘Accounting valuation, market expectation, and cross-sectional stock returns’, Journal of Accounting and Economics, Vol. 25(3), 1998, pp. 283319.
  • Hwang, L.-S. and Sohn, B.-C., ‘Return predictability and shareholders’ real options’, Review of Accounting Studies, Vol. 15, 2010, pp. 367402.
  • Hwang, L.-S. and Sohn, B.-C., ‘A new value-to-price anomaly and idiosyncratic risk’, available at SSRN: http://ssrn.com/abstract=1497974, 2009.
  • Jiang, X. and Lee, B.-S., ‘An empirical test of the accounting-based residual income model and the traditional dividend discount model’, Journal of Business, Vol. 78(4), 2005, pp. 14651504.
  • Jones, R.C., ‘Maybe it really is different this time’, Journal of Portfolio Management, Vol. 36(2), 2010, pp. 6072.
  • Kim, K.-J., Lee C. and Tiras, S.L., ‘Residual income valuation: a new approach based on the value-to-book multiple’, available at SSRN: http://ssrn.com/abstract=1465855, 2009.
  • Lee, C.M.C, Myers J. and Swaminathan, B., ‘What is the intrinsic value of the Dow?Journal of Finance, Vol. 54(5), 1999, pp. 1693741.
  • Ohlson, A.J., ‘The theory of value and earnings, and an introduction to the Ball–Brown analysis’, Contemporary Accounting Research, Vol. 8, 1991, pp. 119.
  • Ohlson, A.J., ‘Earnings, book values, and dividends in equity valuation’, Contemporary Accounting Research, Vol. 11, 1995, pp. 66187.
  • Richardson, S., Tuna I. and Wysocki, P., ‘Accounting anomalies and fundamental analysis: a review of recent research advances’, Journal of Accounting and Economics, Vol. 50, 2010, pp. 41054.