The aim of this master thesis is to understand if a firm’s dependency on the over-all business cycle impacts the performance of accounting-based valuation mod-els. I evaluated stock- and flow-based valuation models and their performance across cyclical and acyclical industries and found that multi-period flow-based models perform well for cyclical firms, while acyclical firms are more accurately valued via market multiples (P/E-multiple). In extension of the empirical findings, an analysis of valuation models used in analyst reports points out that practitioners do not necessarily employ the theoretically superior valuation technique to generate their target prices - while practically used and empirically superior valuation models are consistent in acyclical industries, they significantly differ in cyclical industries. Furthermore, the analysis in this master thesis highlights that analysts show the tendency to adopt different valuation models dependent on the actual cycle phase.
TABLE OF CONTENTS
II List of Figures
III List of Tables
IV Abbreviations
1 Introduction
2 Review of Relevant Literature and Theory
2.1 Informational Content in Financial Statements
2.2 Discussion of Valuation Models
2.2.1 Fundamental Perspectives in Business-Valuation
2.2.2 Stock-Based Valuation Models (Market Multiples)
2.2.2.1 Identification of Comparable Firms
2.2.2.2 Computing the Benchmark Multiple
2.2.2.3 Selection of Relevant Value Drivers
2.2.3 Dividend Discount Model (DDM)
2.2.4 Residual Earnings Valuation Model (REVM)
2.2.5 Abnormal Earnings Growth Valuation Model (AEGM)
2.2.6 Option-Based Valuation Model
2.3 Discussion of Valuation Model Performance
2.4 Concluding Remarks
3 Large Sample Analysis
3.1 Research Question & Motivation
3.2 Original Data & Sample Selection
3.3 Research Methodology
3.3.1 Stock-Based Valuation
3.3.1.1 Selected Value Driver
3.3.1.2 Selection of Comparable Firms
3.3.1.3 Derivation of Benchmark Multiples
3.3.2 Flow-based Valuation
3.3.2.1 Cost of Capital ( )
3.3.2.2 Earnings Per Share (EPS) & forecasted EPS
3.3.2.3 Dividend Payout Rate
3.3.2.4 Long Term Growth-Rate
3.4 Empirical Results
3.4.1 Descriptive Statistics
3.4.2 Analysis of Valuation Errors
3.4.2.1 Intra-Sample Analysis of Valuation Errors
3.4.2.2 Cross-Sample Analysis of Valuation Errors
3.4.2.3 Difference in Valuation Errors between Valuation Models
3.4.2.4 Explanatory Power of Accounting-Based Valuation Models
3.5 Concluding Remarks on Empirical Results
4 Small Sample Analysis
4.1 Research Question & Aims
4.2 Data
4.3 Sample Selection
4.4 Methodology
4.5 Analysis of Primary Valuation Models
4.5.1 Models Employed in Valuing Acyclical Firms
4.5.2 Models Employed in Valuing Cyclical Firms
4.5.3 Summary of Analysts Ratings
4.5.4 Forecast Horizon in Reports on Acyclical and Cyclical Firms
4.5.5 Statements Focused by Analysts’ Forecasts
4.6 Concluding Remarks
5 Conclusion
V Reference List
VI Appendix
II LIST OF FIGURES
Figure 1 - Typical Business Cycle
Figure 2 - Identification of Cycle-Phases
Figure 3 - Forecast Horizons for Acyclical Firms
Figure 4 - Forecast Horizons for Cyclical Firms
III LIST OF TABLES
Table 1 - Summary of Industry Classification & Affiliation to Sub-Samples
Table 2 - Sample Selection Process
Table 3 - Sample Descriptive Statistics
Table 4 - Descriptive Statistics of Valuation Errors
Table 5 - Evaluation of Valuation Accuracy & Valuation Bias
Table 6 - Cross-Sample Analysis of Valuation Errors
Table 7 - Valuation Performance across Valuation Models
Table 8 - Explanatory Power of Accounting Based Valuation Models
Table 9 - Small Sample: Firms Representing Cyclical and Acyclical Industries
Table 10 - Valuation Models Employed in Valuation of Acyclical Firms
Table 11 - Valuation Models Employed in Valuation of Cyclical Firms
Table 12 - Summary of Investment Ratings
Table 13 - Mean Forecasting Horizon for Small Sample Firms
Table 14 - Forecasted Financial Statements
IV ABBREVIATIONS
illustration not visible in this excerpt
1 Introduction
Business cycles are predictable long-term pattern changes in national income and of common occurrence in developed economies. During these cycles economies undergo four stages: expansion, prosperity, contraction, and trough (Downs, Goodman, & Downes, 1998).
Figure 1 displays the typical fluctuation in output levels representing a business cycle and its respective phases of expansion (recovery) and contraction (recession). The cycle’s peak and bottom mark the economies state of prosperity and trough.
illustration not visible in this excerpt
Source: Wall Street Survivor, 2009
Since World War 2, the US economy has experienced 11 recessions, indicating that an average economic cycle spans six years (National Bureau of Economic Research, 2010). Most recessions start in a single industry (e.g. the dot-com-bubble-burst in 2001, or the banking crisis in 2008) and spread rapidly to other economic sectors (Tremolizzo, 2009).
Consequently, industries do not remain unaffected by these changes in the overall macro-environment. Dependent on their sensitivity to changes in the economic environment, financial analysts categorize industries as cyclical or acyclical. The more macroeconomic conditions affect product demand, the more firms are considered to be cyclical. On the contrary, the less product demand is impacted by macro-economic changes the more a firm is considered to be acyclical1. While the latter are scarcely covered by academic literature, the former are frequently examined. A significant part of the related literature highlights considerable difficulties in the valuation of cyclical firms, predominantly attributed to the increased instability of earnings.
Koller et al. (2005) find that earnings volatility in cyclical companies introduces complexity to the valuation process and outline that the fluctuation of earnings is driven by changes in product prices. Additionally, they find that share prices of firms in cyclical industries are significantly more volatile than their acyclical counterparts. Consequently, Koller et al. (2005) suggest, that cyclical companies should be valued via a systematic flow-based valuation approach.
Complementing Koller et al. (2005), De Heer et al. (2001) argue that consensus earnings forecasts in cyclical industries appear to ignore business cyclicality, highlight- ing that these forecasts show the tendency to be overly optimistic regardless of the current cycle phase.
In this master thesis I examine cyclical and acyclical industries with particular regard to their valuation. A large sample analysis will assess the valuation of cyclical and acyclical companies using accounting-based valuation models to (1) outline the differences and similarities across industries and (2) highlight if there is a superior valuation model for each group, delivering more accurate / meaningful value estimates. Subsequently, a small sample analysis examines the valuation techniques employed by practitioners to value both types of firms. Additionally the small sample analysis aims to shed light on analyst forecasting behaviour and how it differs between both groups.
The remainder of the thesis is structured as follows: In the next section, I will outline the relevant literature about equity valuation using accounting-based valuation models. Section 3 delivers the results of the large sample analysis while section 4 turns the focus to a small sample of analyst reports to examine valuation techniques and forecast horizons employed by analysts. Section 5 concludes the thesis with a summary of main results and suggests fields for further research.
2 Review of Relevant Literature and Theory
Equity valuation can be considered to be the practical application of accounting and finance theory, representing a process of converting forecasts into a value estimate of a complete firm or one of its components (Palepu, Healy, & Bernard, 2007).
Given the significance of pricing for capital markets, it becomes evident why business school curricula devote substantial time to valuation and its concepts. Typi- cally all theoretically known and practically used valuation techniques have to accom- modate these concepts since ignoring them is the source of many misconceptions (Lee, 1999).
Starting with a reflection about informational content of financial statements, the following chapter will outline different accounting-based valuation models and introduce a body of literature to provide for the theoretical foundations of this thesis.
2.1 Informational Content in Financial Statements
The most apparent concept of valuation is its inherent prospective nature, involving the forecasting of uncertain future payoffs and other determining variables (Lee, 1999). Therefore, published financial statements build the basis for analysts’ forecasts and estimates, providing a comprehensive source of information on the company.
Within these published and audited figures, earnings are believed to be the premier source of information, thus most valuation models used by practitioners and academics share the common element of employing forecasted earnings as explana- tory variable for their value estimates (Lev, 1989). Following this notion, earnings pro- vide an adequate measure of value, reflecting the current firm-specific results of em- ploying its resources and represent a relatively more important determinant of value if the firm’s current activities are successful (Burgstahler & Dichev, 1997). Regardless of which valuation concept is applied, it is highly dependant on reported earnings, earn- ings forecasts and the way earnings are measured and reported in financial state- ments. Owed to the importance of earnings for the forecasting and valuation process, investors directly look at and react on earnings releases (Beaver, 1968).
Beaver (1968) finds that both an individual investor’s and the overall market’s expectations are altered by earnings announcements, thus earnings-figures enclose significant information. Ball & Brown (1968) complement Beaver (1968) by empirically testing how stock prices reflect the flow of information, emphasising informational con- tent of income numbers. According to their study, information contained in annual in- come figures appear to be particularly useful to investors if reported income differs from expected income. They argue that share prices move consistent with the fact that if reported earnings meet or fall short market expectations, thus information contained in earnings is useful and highly related to stock prices. However, this notion is limited by the fact that annual reports as primary source of information are not regarded to be a timely medium, since their content is captured by more prompt media. In fact, investors have turned to more timely sources, gaining the opportunity to act more promptly upon recent changes in a firm’s profitability (Ball & Brown, 1968).
Contradicting Beaver (1968) and Ball & Brown (1968), who consider earnings as the most accurate proxy for value creation, Lev (1989) argues that their usefulness to investors as forecasting basis is limited by their low correlation to stock returns and the considerable instability of the returns / earnings relation. He attributes the low re- turn / earnings correlation to the poor informational content of earnings since they are subject to the limitations of accounting measurement, accounting valuation principles and the possible manipulation of accounting data by corporate management (Lev, 1989).
In extension of the delineation of informational content in financial statements, Aboody et al. (1999) examine the relationship between asset revaluations and a firm’s future performance. They prove that upward revaluations are significantly correlated with a firm’s future sound performance, measured by operating income and cash from operations. Furthermore, the study highlights a positive correlation between revaluations and annual returns. However, they argue that the association of asset revaluations and future performance to share prices is weaker for high debt-to-equity firms than for firms burdened with less debt.
2.2 Discussion of Valuation Models
Practitioners and academics employ a variety of accounting-based valuation models, which can be separated into stock-based and flow-based approaches. The main difference between both approaches is that the latter depends greatly on a range of forecasted inputs while the former does not.
In the following sub-section I will introduce the fundamental perspectives towards business valuation and outline five accounting-based valuation models and their implementation issues.
2.2.1 Fundamental Perspectives in Business-Valuation
Prior to the discussion of valuation models, the distinction between equity and entity perspective in business valuation needs to be presented, since all valuation models can be expressed in either perspective. The equity perspective2 distinguishes between the sources of capital, while the entity perspective3 ignores them entirely. Generally, the valuation perspective depends on the likelihood a valuation itself is impacted by firm specific accounting or financing policies (Citigroup Global Markets - Equity Re- search, 2005).
The entity perspective is less affected by financing differences focussing on the firm as a single operating unit. By valuing all outstanding financial claims on the firm, it enables a more meaningful comparison of value estimates because these are uninflu- enced by management’s financing decisions (Citigroup Global Markets - Equity Re- search, 2005).
The equity perspective solely focuses on firm-value attributable to equity hold- ers, using operating statistics affected by capital structure (Deutsche Bank Global Banking, 2008). The equity perspective is preferred if entity valuations are affected by accounting differences or are simply less meaningful4. A downside of equity-based value estimates is their affection by firm-specific financing decisions, reducing useful- ness of a comparison across firms (Deutsche Bank Global Banking, 2008).
2.2.2 Stock-Based Valuation Models (Market Multiples)
Stock-based valuation models use a ratio of the current stock price to summary statis- tics from financial statements of comparable firms to generate a value estimate (Pen- man, 2008). Generally, these market multiples are considered to be the most common technique in equity valuation, widely used by analysts and investment bankers in re- ports, sell-side recommendations, fairness opinions, IPOs, LBOs, SEOs and other merger & acquisition transactions (Bhojraj & Lee, 2002). They are particularly important to the valuation of IPO’s, since most complex valuation models appear to be imprecise due to the difficulties in estimating valuation parameters (Kim & Ritter, 1999).
The immense popularity of multiples in valuation is mainly grounded on their simple derivation and application. Like more sophisticated valuation techniques, multiples rely on the assumption that value is an increasing function of future payoffs and a decreasing function of risk (Liu, Nissim, & Thomas, 2002). However, analysts embed both techniques in their reports using multiple-based valuations as implicit support of more complex models (Demirakos, Strong, & Walker, 2004).
By construction, multiples refer to market values, thus creating an indirect, mar- ket-based value estimate. Since all value estimates refer to a market variable, they reflect the mood of the market and need to be regarded as relative rather than intrinsic value estimates. This particular feature of multiple valuations assists investors to get a feeling for market values of privately held entities5 and plays a key role in the process of determining appropriate prices and price ranges for transactions (Penman, 2008).
A typical multiple valuation involves four steps: The first two steps involve the selection of value drivers and the identification of the comparable firms. In the third step, all derived multiples of comparable firms are aggregated into a single number yielding a benchmark multiple. The last step is the actual valuation by applying the benchmark multiple to the value driver of the target firm (Penman, 2008). Algebraically the multiple-valuation can be written as: (1)
illustration not visible in this excerpt
Where Vi is the value estimate, represents the value driver (where VDi>0) and reflects the set of comparable firms6.
As outlined in section 2.2.1, multiples may reflect firm value at the equity level, where is an equity value driver (e.g. Net Income), or at the entity level, where represents an entity value driver (e.g. NOPAT). The most common multiple used by practitioners and academics is the price-earnings multiple (P/E-multiple) which uses earnings due to their unique characteristics discussed in section 2.1.
Despite their popularity, multiple-based valuations incorporate several concep- tual problems mainly attributable to their nature as a valuation heuristic. These imple- mentation issues will be outlined and dealt with in the subsections 2.2.2.1 - 2.2.2.3.
2.2.2.1 Identification of Comparable Firms
The identification of comparable firms has a significant impact on a multiple’s valuation performance. Liu et al. (2002) underline the materiality of the choice of comparable firms by showing that valuation performance deteriorates substantially if all firms of a sample are used as comparable firms. In practise, comparable firms are selected by industry affiliation, since firms of the same industry are expected to be similar in terms of earnings growth and operating risks.
This practise was empirically investigated by Alford (1992) who finds that valua- tions using comparables chosen by their 3-digit SIC code appear to be a good surro- gate for industry specific characteristics. He finds that valuation accuracy gradually increases if the distinction of industry definition used to identify comparables is nar- rowed down from 1-digit to 3-digit SIC codes. However, no further improvement is achieved if comparable firms are identified by their 4-digit SIC codes. Multiple value estimates based on Alfords (1992) industry matching approach often show the ten- dency to undervalue (Lie & Lie, 2002).
An alternative to Alford’s identification of comparable firms using industry matching was introduced by Bhojraj & Lee (2002). They find that relative to the identification by industry matching, the explanatory power of multiple valuations is substantially improved if valuation multiples incorporate firm specific characteristics like growth and profitability. Value estimates following this approach generally showed a higher valuation accuracy compared to those generated by industry matching7.
2.2.2.2 Computing the Benchmark Multiple
Combined with the identification of comparable firms is the problem that most common value drivers in multiples8 are more volatile than equity prices and therefore compara- ble firm multiples are broadly dispersed (Fernández, 2002). The problem of dispersion of comparable firm multiples is often encountered in accounting research and demands a statistic estimator to synthesise a benchmark-multiple. The following formulae outline the common estimation techniques next to the median9:
illustration not visible in this excerpt
Where VDj denotes the value driver, Pj the observed price for the jth comparator firm. In their studies, Alford (1992) and Bhojraj & Lee (2002) derive benchmark-multiples by applying the median of comparable firm multiples.
Baker & Ruback (1999) argue that the application of the harmonic mean (4) delivers superior valuation performance out of the four previously outlined estimators. Furthermore they highlight that benchmark-multiples derived by the simple mean will show a constant tendency to overvalue, due to the mathematical fact that it always yields a higher number than a harmonic mean. Additionally, they show that benchmark- multiples-based on the simple mean are considerably influenced by outliers.
Supporting Baker & Ruback (1999), Beatty et al. (1999) and Liu et al. (2001) argue that valuation accuracy is significantly increased if benchmark multiples are cal- culated using the harmonic mean of comparable firm multiples rather than median or mean.
2.2.2.3 Selection of Relevant Value Drivers
The choice of value driver is essential to multiple valuations and should generally reflect an adequate proxy for the company’s performance. Since accounting-based value drivers can turn negative, a value estimate based on these negative statistics would not be meaningful. The most frequently used value driver in multiples is accounting earnings, due to their high informational content. However the value driver selection differs dependent on the company and the industry it is assigned to.
Liuetal. (2002) and Fernández (2002) find that multiples using forecasted earn- ings as value driver perform best and valuation performance is further increased by an extension of their forecast horizon. Additionally, Liu et al. (2002) underline the applica- tion of forward earnings showing that across different GAAP jurisdictions, multiples using forecasted earnings as value drivers outperform multiples using reported earn- ings. The popularity of P/E-multiples is further underlined by Liu et al. (2007) who show that valuations based on earnings multiples are, for a substantial majority of compa- nies, remarkably more accurate than value estimates generated by cash flow multiples.
Fernández (2002), argues that book-value-based multiples generate accurate valuations for companies whose assets are precisely measured by accounting rules. He highlights that multiples using non-GAAP earnings10 outperform multiples-based on GAAP earnings in R&D-intensive industries, since they exclude one-off- and non- current-expenses. The study underlines that operating earnings are a better proxy for value generation than GAAP earnings and concludes that valuation performance de- pends on the explanatory power of GAAP figures within industries. Furthermore the study shows that valuation performance of multiple value estimates varies significantly with a company’s size, profitability, and the amount of intangible assets held. In particu- lar value estimates in intangible-rich industries lack accuracy, suggesting that they do not capture the growth potential and intangibles associated with the companies and their industry (Fernández, 2002).
This issue is further highlighted by Demirakos et al. (2004) who argue that the pervasiveness of comparative valuation models results in no significant difference in the use across sectors, but are more popular in stable sectors where conventional ac- counting does a better job of capturing firm value. Although multiples are subject to several conceptual restrictions, practitioners prefer them to the more complex valuation models because they are easier to process and more efficient to communicate (Demirakos et al., 2004).
Hence, investment professionals tend to prefer certain multiples following some principles: First, for firms whose assets are well measured by current accounting rules, multiples include book values. Secondly, multiples for capital-intensive industries include cash flows in order to ignore most of the non-cash charges. Thirdly, industries in which many firms show losses due to accounting principles are often valued by revenue-based multiples (Liu, Nissim, & Thomas, 2007).
The choice of preferred multiples can be viewed as consistent with the logic related to the reliability and relevance of the numbers produced by the accounting system. Where accounting rules measure key assets well, analysts use audited summary statistics from financial statements, but if the accounting system fails to measure key assets well, value drivers are chosen from higher up the income statement to reverse effects of adulterant accounting procedures (Tasker, 1998).
2.2.3 Dividend Discount Model (DDM)
The DDM can be regarded as the fundamental flow-based equity valuation model, since any other model is reconcilable with it, making it a reference point for many other valuation techniques (Barker, 2001). The model itself is based on the assumption that the market value of equity capital is defined as the sum of discounted future net cash flows. From the investor’s point of view, the equity value is determined by the present value of dividends received in future periods and capital gains.
illustration not visible in this excerpt
Were dt denotes the dividends and ptE the costs of equity-capital. Using dividends as main input variable, the DDM appeals to investors because forecasting is considered to be easy given usually stable dividend policies (Brealey, Myers & Franklin, 2005; Penman, 2008). However, the dividend discount model requires dividend forecasts to infinity, which in practise is impossible.
To deal with this shortcoming, Copeland et al. (2008) suggest separating business value into two periods, the first during and the second after the explicit forecasting period. The value beyond the forecast horizon should represent a “terminal” or “continuing” value featuring a continuous growth rate.
Penman (2008) highlights several conceptual restrictions of DDM since there is no explicit direction given to the terminal value calculation. Even more immanent is the DDM’s dependency on Modigliani and Miller’s (1961) dividend irrelevance proposition, highlighting that future dividends are value-irrelevant since they are unrelated to value creation and ignore inter-period capital gains. To overcome Modigliani & Miller’s dividend irrelevancy paradigm Barker (2001) suggests applying the DDM only with dividends tied to the value generation within the company.
The models which will be introduced in section 2.2.4 and 2.2.5 aim to overcome the shortcomings of the DDM by extending the model itself, its assumptions and its principles.
2.2.4 Residual Earnings Valuation Model (REVM)
Initially proposed by Edwards & Bell (1961) and promoted again by Ohlson (1995), the REVM estimates equity value as the book value of equity capital plus the present value of future clean surplus residual earnings. Ohlson (1995) defines residual earnings[11] as accounting profit less a capital charge based on the opening book value of equity.
Ohlson’s definition is comparable to the economic value added (EVA) paradigm where EVA is the product of total capital and the difference between return on assets (ROA) and weighted average cost of capital (WACC). Thus residual earnings and EVA can be interpreted as the difference between earnings and a charge for capital (Anand & Faseruk, 2008).
The link between the REVM and DDM is the clean surplus relationship, suggesting that all equity effects[12] are carried through the income statement. Hence the existing book value of equity equals opening book value plus expected net income less expected dividends. The formal derivation of the REVM grounds on the present value of future dividends.
illustration not visible in this excerpt
Essential for the REVM’s application is the assumption of a valid clean surplus relationship, where earnings equal the change in book value of equity plus dividends net of capital transactions[12] (O'Hanlon, 2009). The clean surplus relation can be written as:
Еагщ = B t + d t - B t-i (10)
Hence, residual earnings can be expressed as
REt = Bt + dt — Bt-i — g°e — l)Bt-i ([11])
REt = Bt + dt — oEBt-l ([12])
Substituting (12) into (8) yields the REVM-formula:
illustration not visible in this excerpt
The REVM claims that equity value always equals its book value, plus a premium resulting from the present value of future residual income.
O’Hanlon & Peasnell (2002) confirm this idea by demonstrating that regardless of the accounting procedure, intrinsic value of the firm can be calculated by the combination of income statement and balance sheet figures. Contrasting the DDM, the REVM captures the Modigliani & Miller setting of dividend policy irrelevance effectively since the model doesn’t use dividends as primary input variable.
Furthermore, Bryan et al. (2001) consider the REVM to adequately address conceptual problems of multiples, because it provides a simple and theoretical model for the computation of fundamental equity values by referring to observed book values, return on equity (ROE) and cost of equity capital (Bryan, Hwang, & Lilien, 2001). By construction, the REVM shows that the deviation of price and book value depends on the firm’s ability to generate “abnormal earnings” (Palepu et al., 2007).
One major advantage of the residual income valuation model is its reliance on easily available data inputs, since forecasted EPS, long term growth in EPS, forecasted dividends, opening book value and cost of capital receive enormous attention and are readily available from various databases (Bryan et al., 2001).
In several empirical studies, residual-earnings-based value estimates perform significantly better than dividend discount or cash flow-based value estimates, due to (1) the model’s anchor in book values, which explains a larger proportion of intrinsic value and (2) the use of more accurate earnings forecasts (Courteau, Gray, Kao, O'Keefe, & Richardson, 2007). Neither accounting discretion nor accounting conservatism, two factors often quoted as limitations of accounting-based valuation models, show significant impact on the reliability of residual earnings value estimates (Francis, Olsson, & Oswald, 2000).
Francis et al. (2000) highlight this by showing no empirical basis for concerns that accounting practices result in inferior residual earnings value estimates for high / low R&D firms. They suggest that the articulation of the clean surplus relationship ensures that value estimates are unaffected by accounting conservatism or accrual practices.
Frankel and Lee (1998) support Francis et al. (2000), showing that residual earnings value-estimates using analysts’ earnings forecasts are highly correlated with stock prices and explain more than 70 percent of their cross-sectional variation.
Supplementing the use of “core earnings” by Frankel & Lee (1998), Courteau et al. (2007) show that valuation performance of the REVM is increased with the number of forecasted input variables. By incorporating a broader set of variables out of different informational sources like I/B/E/S[13] and ValueLine[14] the REVM’s performance was increased the more forecasted variables were entered. Additionally, Lee et al. (1999) suggest that the inclusion of a time-varying discount rate increases accuracy and explanatory power for the REVM.
In theory the DDM, the discounted cash flow (DCF) model and the REVM yield identical value estimates, but they differ in practice if forecasted attributes, growth rates or discount rates are inconsistent. There was a controversy among researchers whether REVM or DCF represents a superior valuation technique. While Penman & Sougiannis (1998), Francis et al. (2000) and Courtreau et al. (2001) regard the REVM as superior technique for valuation within finite horizons, Lundholm & O’Keefe (2001) contradict that REVM and DCF express equal value estimates if a full set of pro-forma statements is available. The discussion was brought to an end by Richardson & Tinai- kar (2004) who conclude that Penman and Lundholm & O’Keefe are correct (Anand & Faseruk, 2008).
Next to the aspect of the model’s valuation accuracy, Richardson & Tinaikar (2004) point out that the REVM, based on its linear information dynamics, is able to handle relevant but omitted variables from other information, thus providing an intellectual foundation for forecasting. Furthermore they highlight the REVM’s academic contribution to the understanding of the properties of accounting-based valuation models using truncated horizons. Following this notion, Bryan et al. (2001) suggest that the REVM should be used to standardise and benchmark analysts’ assumptions enabling an assessment of the reasonableness of target prices.
Ohlson (2005) highlights the REVM’s deficiencies by relating to its dependency on the clean surplus relationship and the anchorage in book values. Since the model uses the present value of above or below benchmark increments in expected book values, adjusted for dividends, to explain the market minus book value premium, there is no implicit need to refer to earnings or residual earnings, because they enter the model only via the clean surplus relationship and in an underhanded way (Ohlson, 2005). Additionally Ohlson (2005) highlights the REVM’s inability to capture effects of capital transactions if it is applied on a per-share-basis, since it is dependent on expected dividends and a valid clean surplus relationship on a per-share-basis. All together, Ohlson (2005) concludes that the REVM is not able to generate per-share value estimates unless Modigliani & Miller-type restrictions are re-introduced.15
As a consequence of these problems, Ohlson & Juettner-Nauroth (2005) expanded the valuation model, by relating a firm’s share price to its capitalised next period earnings, its short and long term earnings growth and cost of equity capital - creating the abnormal earnings growth valuation model.
2.2.5 Abnormal Earnings Growth Valuation Model (AEGM)
Like the REVM, the AEGM conceptually grounds on the DDM. In contrast to the REVM, the AEGM expresses intrinsic value of equity as capitalised, next-period earnings plus the present value of capitalised, forecasted abnormal earnings growth in subsequent periods. Abnormal earnings growth is defined as the difference between periodic earnings change and a normal return on previous-period earnings (Ohlson & Juetter- Nauroth, 2005). The AEGM uses a similar mathematical structure like the REVM, also starting from the present value of future dividends (Penman, 2008; O'Hanlon, 2009).
illustration not visible in this excerpt
Extension of the present value of future dividends by the following zero sum expression
illustration not visible in this excerpt
The AEGM can be derived by defining ytyt as the expectation at t = Oî = Û of future earnings at t = If +■ 1 capitalised as perpetuity (Ohlson, 2005):
illustration not visible in this excerpt
Substituting yt with capitalised future earnings yields:
illustration not visible in this excerpt
Since abnormal earnings growth can be algebraically defined as AEGt+1 _ (Earnt+1 _ Earnt) _ (pE _ 1 )(Earnt _ dt)
The AEGM is represented by the following formula:
illustration not visible in this excerpt
The AEGM overcomes the problems of the REVM by anchoring on capitalised, next- period earnings, expressing its premium in terms of subsequent increments in expected earnings adjusted for dividends (O'Hanlon, 2009). It neither demands an anchor in book values nor relies on the theoretical concept of the clean surplus relationship enabling a valuation on both a total dollar as well as a per-share basis, ignoring adverse effects of capital transactions (Ohlson, 2005).
From the practitioner's side, the conceptual advantage of the AEGM is its focus on earnings, which are considered the main catalysts for value creation. Furthermore the model underlines the idea that ex-ante capitalised earnings approximate market value more closely than book values (Ohlson, 2005).
Finally, Ohlson (2005), who developed the residual earnings valuation framework, predicts that the AEGM will replace the REVM, due to its alignment with the earnings focus of the investment industry.
2.2.6 Option-Based Valuation Model
Burgstahler & Dichev (1997) explain equity value as a function of expected earnings and book values. Their model grounds on a convex relationship between expected earnings and book values, contradicting all common valuation approaches, which assume that earnings and book values are additive. Burgstahler & Dichev (1997) assume equity value to be a function of earnings and book values depending on their relative values, forming an option-type valuation model. They define “recursion value” as the present value of a stream of earnings assuming that the business technology remains to be applied. As complement to recursion value, they define “adaption value” as the value of resources adaptable to alternative external or internal purposes. In essence earnings and net-assets are combined in an option, where the likelihood of exercising the “adaption option” depends on the relative value of earnings and book values while value relevance is a function of their relative levels. The model is the first valuation model suggesting that valuation should incorporate both earnings and balance sheet figures as measures of value.
2.3 Discussion of Valuation Model Performance
When it comes to evaluating valuation models, the gap between practitioners and academics widens, since the theoretical superiority of multi-period valuation models contrasts their application by practitioners (Demirakos et al., 2004).
Frankel & Lee (1998) find that REVM-value estimates are highly correlated to stock prices since their multiple consisting of a REVM-value estimate and the current price represents a sound predictor of long-term cross-sectional returns. Lee et al. (1999) extend Frankel & Lee (1998) by comparing the performance of traditional market multiples (B/P, E/P, and D/P ratio) and find that these ratios lag behind the predictive power of a V/P ratio, where V is based on a residual earnings valuation. The line of argumentation in favour of formal valuation models is further extended by Gleason et al. (2008) arguing that the predictive performance of price targets is significantly improved if analysts apply flow-based valuation models rather than multiples, highlighting that the investment value of analysts’ target prices declines significantly if these ground on valuation heuristics combined with inaccurate earnings forecasting.
Although academics prove the superiority of formal valuation models, they play a limited role in practical valuation, as practitioners tend towards the application of stock-based models. Valuation heuristic
[...]
1 Often referred to as defensive (both terms are used interchangeably in this master thesis).
2 or proprietary perspective
3 or enterprise perspective
4 For example in the valuation of financial service providers.
5 e.g. private firms, subsidiaries or single business units of publicly traded firms
6 See section 2.2.2.2 for derivation of the benchmark multiple.
7 using 2-digit SIC codes
8 e.g. earnings, EBITDA and profit after tax
9 A measure of tendency, represented by the middle number of values arranges in ascending (descending) order (Newbold, Thorne, & Carlson, 2007)
10 often referred to as operating earnings or street earnings
11 also referred to as abnormal earnings
12 excluding capital transactions
13 This earnings definition is often referred to as comprehensive income, indicating that all prior-year adjustments, extraordinary items and asset revaluations pass through the income statement.
14 only providing earnings forecast
15 providing forecasts for earnings, book values, cost of capital as well as individual terminal values
-
-
-
-
Upload your own papers! Earn money and win an iPhone X. -
-
Upload your own papers! Earn money and win an iPhone X. -
-
Upload your own papers! Earn money and win an iPhone X. -
-
Upload your own papers! Earn money and win an iPhone X. -
-
Upload your own papers! Earn money and win an iPhone X. -
-
Upload your own papers! Earn money and win an iPhone X. -
-
Upload your own papers! Earn money and win an iPhone X. -
-
Upload your own papers! Earn money and win an iPhone X. -
-
Upload your own papers! Earn money and win an iPhone X. -
-
Upload your own papers! Earn money and win an iPhone X. -
-
Upload your own papers! Earn money and win an iPhone X. -
-
Upload your own papers! Earn money and win an iPhone X. -
-
Upload your own papers! Earn money and win an iPhone X. -
-
Upload your own papers! Earn money and win an iPhone X. -
-
Upload your own papers! Earn money and win an iPhone X.