Contents Cover Series Title Page Copyright Dedication Preface to the Third Edition Chapter 1: Introduction to Valuation A PHILOSOPHICAL BASIS FOR VALUATION GENERALITIES ABOUT VALUATION THE ROLE OF VALUATION CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 2: Approaches to Valuation DISCOUNTED CASH FLOW VALUATION RELATIVE VALUATION CONTINGENT CLAIM VALUATION CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 3: Understanding Financial Statements THE BASIC ACCOUNTING STATEMENTS ASSET MEASUREMENT AND VALUATION MEASURING FINANCING MIX MEASURING EARNINGS AND PROFITABILITY MEASURING RISK OTHER ISSUES IN ANALYZING FINANCIAL STATEMENTS CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 4: The Basics of Risk WHAT IS RISK? EQUITY RISK AND EXPECTED RETURN ALTERNATIVE MODELS FOR EQUITY RISK 1
A COMPARATIVE ANALYSIS OF EQUITY RISK MODELS MODELS OF DEFAULT RISK CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 5: Option Pricing Theory and Models BASICS OF OPTION PRICING DETERMINANTS OF OPTION VALUE OPTION PRICING MODELS EXTENSIONS OF OPTION PRICING CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 6: Market Efficiency—Definition, Tests, and Evidence MARKET EFFICIENCY AND INVESTMENT VALUATION WHAT IS AN EFFICIENT MARKET? IMPLICATIONS OF MARKET EFFICIENCY NECESSARY CONDITIONS FOR MARKET EFFICIENCY PROPOSITIONS ABOUT MARKET EFFICIENCY TESTING MARKET EFFICIENCY CARDINAL SINS IN TESTING MARKET EFFICIENCY SOME LESSER SINS THAT CAN BE A PROBLEM EVIDENCE ON MARKET EFFICIENCY TIME SERIES PROPERTIES OF PRICE CHANGES MARKET REACTION TO INFORMATION EVENTS MARKET ANOMALIES EVIDENCE ON INSIDERS AND INVESTMENT PROFESSIONALS CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 7: Riskless Rates and Risk Premiums THE RISK-FREE RATE EQUITY RISK PREMIUM DEFAULT SPREADS ON BONDS CONCLUSION 2
QUESTIONS AND SHORT PROBLEMS Chapter 8: Estimating Risk Parameters and Costs of Financing THE COST OF EQUITY AND CAPITAL COST OF EQUITY FROM COST OF EQUITY TO COST OF CAPITAL BEST PRACTICES AT FIRMS CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 9: Measuring Earnings ACCOUNTING VERSUS FINANCIAL BALANCE SHEETS ADJUSTING EARNINGS CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 10: From Earnings to Cash Flows THE TAX EFFECT REINVESTMENT NEEDS CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 11: Estimating Growth THE IMPORTANCE OF GROWTH HISTORICAL GROWTH ANALYST ESTIMATES OF GROWTH FUNDAMENTAL DETERMINANTS OF GROWTH QUALITATIVE ASPECTS OF GROWTH CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 12: Closure in Valuation: Estimating Terminal Value CLOSURE IN VALUATION THE SURVIVAL ISSUE CLOSING THOUGHTS ON TERMINAL VALUE CONCLUSION QUESTIONS AND SHORT PROBLEMS 3
Chapter 13: Dividend Discount Models THE GENERAL MODEL VERSIONS OF THE MODEL ISSUES IN USING THE DIVIDEND DISCOUNT MODEL TESTS OF THE DIVIDEND DISCOUNT MODEL CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 14: Free Cash Flow to Equity Discount Models MEASURING WHAT FIRMS CAN RETURN TO THEIR STOCKHOLDERS FCFE VALUATION MODELS FCFE VALUATION VERSUS DIVIDEND DISCOUNT MODEL VALUATION CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 15: Firm Valuation: Cost of Capital and Adjusted Present Value Approaches FREE CASH FLOW TO THE FIRM FIRM VALUATION: THE COST OF CAPITAL APPROACH FIRM VALUATION: THE ADJUSTED PRESENT VALUE APPROACH EFFECT OF LEVERAGE ON FIRM VALUE ADJUSTED PRESENT VALUE AND FINANCIAL LEVERAGE CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 16: Estimating Equity Value per Share VALUE OF NONOPERATING ASSETS FIRM VALUE AND EQUITY VALUE MANAGEMENT AND EMPLOYEE OPTIONS VALUE PER SHARE WHEN VOTING RIGHTS VARY CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 17: Fundamental Principles of Relative Valuation USE OF RELATIVE VALUATION STANDARDIZED VALUES AND MULTIPLES FOUR BASIC STEPS TO USING MULTIPLES 4
RECONCILING RELATIVE AND DISCOUNTED CASH FLOW VALUATIONS CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 18: Earnings Multiples PRICE-EARNINGS RATIO THE PEG RATIO OTHER VARIANTS ON THE PE RATIO ENTERPRISE VALUE TO EBITDA MULTIPLE CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 19: Book Value Multiples PRICE-TO-BOOK EQUITY APPLICATIONS OF PRICE–BOOK VALUE RATIOS USE IN INVESTMENT STRATEGIES VALUE-TO-BOOK RATIOS TOBIN’S Q: MARKET VALUE/REPLACEMENT COST CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 20: Revenue Multiples and Sector-Specific Multiples REVENUE MULTIPLES SECTOR-SPECIFIC MULTIPLES CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 21: Valuing Financial Service Firms CATEGORIES OF FINANCIAL SERVICE FIRMS WHAT IS UNIQUE ABOUT FINANCIAL SERVICE FIRMS? GENERAL FRAMEWORK FOR VALUATION DISCOUNTED CASH FLOW VALUATION ASSET-BASED VALUATION RELATIVE VALUATION ISSUES IN VALUING FINANCIAL SERVICE FIRMS CONCLUSION 5
QUESTIONS AND SHORT PROBLEMS Chapter 22: Valuing Firms with Negative or Abnormal Earnings NEGATIVE EARNINGS: CONSEQUENCES AND CAUSES VALUING NEGATIVE EARNINGS FIRMS CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 23: Valuing Young or Start-Up Firms INFORMATION CONSTRAINTS NEW PARADIGMS OR OLD PRINCIPLES: A LIFE CYCLE PERSPECTIVE VENTURE CAPITAL VALUATION GENERAL FRAMEWORK FOR ANALYSIS VALUE DRIVERS ESTIMATION NOISE IMPLICATIONS FOR INVESTORS IMPLICATIONS FOR MANAGERS THE EXPECTATIONS GAME CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 24: Valuing Private Firms WHAT MAKES PRIVATE FIRMS DIFFERENT? ESTIMATING VALUATION INPUTS AT PRIVATE FIRMS VALUATION MOTIVES AND VALUE ESTIMATES VALUING VENTURE CAPITAL AND PRIVATE EQUITY STAKES RELATIVE VALUATION OF PRIVATE BUSINESSES CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 25: Aquisitions and Takeovers BACKGROUND ON ACQUISITIONS EMPIRICAL EVIDENCE ON THE VALUE EFFECTS OF TAKEOVERS STEPS IN AN ACQUISITION TAKEOVER VALUATION: BIASES AND COMMON ERRORS STRUCTURING THE ACQUISITION 6
ANALYZING MANAGEMENT AND LEVERAGED BUYOUTS CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 26: Valuing Real Estate REAL VERSUS FINANCIAL ASSETS DISCOUNTED CASH FLOW VALUATION COMPARABLE/RELATIVE VALUATION VALUING REAL ESTATE BUSINESSES CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 27: Valuing Other Assets CASH-FLOW-PRODUCING ASSETS NON-CASH-FLOW-PRODUCING ASSETS ASSETS WITH OPTION CHARACTERISTICS CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 28: The Option to Delay and Valuation Implications THE OPTION TO DELAY A PROJECT VALUING A PATENT NATURAL RESOURCE OPTIONS OTHER APPLICATIONS CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 29: The Options to Expand and to Abandon: Valuation Implications THE OPTION TO EXPAND WHEN ARE EXPANSION OPTIONS VALUABLE? VALUING A FIRM WITH THE OPTION TO EXPAND VALUE OF FINANCIAL FLEXIBILITY THE OPTION TO ABANDON RECONCILING NET PRESENT VALUE AND REAL OPTION VALUATIONS CONCLUSION QUESTIONS AND SHORT PROBLEMS 7
Chapter 30: Valuing Equity in Distressed Firms EQUITY IN HIGHLY LEVERED DISTRESSED FIRMS IMPLICATIONS OF VIEWING EQUITY AS AN OPTION ESTIMATING THE VALUE OF EQUITY AS AN OPTION CONSEQUENCES FOR DECISION MAKING CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 31: Value Enhancement: A Discounted Cash Flow Valuation Framework VALUE-CREATING AND VALUE-NEUTRAL ACTIONS WAYS OF INCREASING VALUE VALUE ENHANCEMENT CHAIN CLOSING THOUGHTS ON VALUE ENHANCEMENT CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 32: Value Enhancement: Economic Value Added, Cash Flow Return on Investment, and Other Tools ECONOMIC VALUE ADDED CASH FLOW RETURN ON INVESTMENT A POSTSCRIPT ON VALUE ENHANCEMENT CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 33: Probabilistic Approaches in Valuation: Scenario Analysis, Decision Trees, and Simulations SCENARIO ANALYSIS DECISION TREES SIMULATIONS AN OVERALL ASSESSMENT OF PROBABILISTIC RISK-ASSESSMENT APPROACHES CONCLUSION QUESTIONS AND SHORT PROBLEMS Chapter 34: Overview and Conclusion CHOICES IN VALUATION MODELS WHICH APPROACH SHOULD YOU USE? CHOOSING THE RIGHT DISCOUNTED CASH FLOW MODEL 8
CHOOSING THE RIGHT RELATIVE VALUATION MODEL WHEN SHOULD YOU USE THE OPTION PRICING MODELS? CONCLUSION References Index
9
Founded in 1807, John Wiley & Sons is the oldest independent publishing company in the United States. With offices in North America, Europe, Australia, and Asia, Wiley is globally committed to developing and marketing print and electronic products and services for our customers' professional and personal knowledge and understanding. The Wiley Finance series contains books written specifically for finance and investment professionals as well as sophisticated individual investors and their financial advisors. Book topics range from portfolio management to ecommerce, risk management, financial engineering, valuation, and financial instrument analysis, as well as much more. For a list of available titles, please visit our Web site at www.WileyFinance.com.
10
11
Copyright © 2012 by Aswath Damodaran. All rights reserved. Published by John Wiley & Sons, Inc., Hoboken, New Jersey. Published simultaneously in Canada. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 646-8600, or on the Web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/ go/permissions. Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books. For more information about Wiley products, visit our web site at www.wiley.com.
Library of Congress Cataloging-in-Publication Data: Damodaran, Aswath. Investment valuation : tools and techniques for determining the value of any asset / Aswath Damodaran.—3rd ed. p. cm.—(Wiley finance series) Includes bibliographical references and index. ISBN 978-1-118-01152-2 (cloth); ISBN 978-1-118-20654-6 (ebk); ISBN 978-1-118-20655-3 (ebk); ISBN 978-1-118-20656-0 (ebk) ISBN 978-1-118-13073-5 (paper); ISBN 978-1-118-20657-7 (ebk); ISBN 978-1-118-20658-4 (ebk); ISBN 978-1-118-20659-1 (ebk) 1.Corporations—Valuation—Mathematical models. I. Title. HG4028.V3 D353 2012 658.15—dc23 2011052858
12
I would like to dedicate this book to Michele, whose patience and support made it possible, and to my four children —Ryan, Brendan, Kendra, and Kiran—who provided the inspiration.
13
Preface to the Third Edition This is a book about valuation—the valuation of stocks, bonds, options, futures and real assets. It is a fundamental precept of this book that any asset can be valued, albeit imprecisely in some cases. I have attempted to provide a sense of not only the differences between the models used to value different types of assets, but also the common elements in these models. The past decade has been an eventful one for those interested in valuation for several reasons. First, the growth of Asian and Latin American markets brought emerging market companies into the forefront, and you will see the increased focus on these companies in this edition. Second, we saw the havoc wreaked by macro-economic factors on company valuations during the bank crisis of 2008, and a blurring of the lines between developed and emerging markets. The lessons I learned about financial fundamentals during the crisis about risk-free rates, risk premiums and cash flow estimation are incorporated into the text. Third, the past year has seen the influx of social media companies, with small revenues and outsized market capitalizations, in an eerie replay of the dot-com boom from the late 1990s. More than ever, it made clear that the more things change, the more they stay the same. Finally, the entry of new players into equity markets (hedge funds, private equity investors and high-frequency traders) has changed markets and investing dramatically. With each shift, the perennial question arises: “Is valuation still relevant in this market?” and my answer remains unchanged, “Absolutely and more than ever.” As technology increasingly makes the printed page an anachronism, I have tried to adapt in many ways. First, this book will be available in e-book format, and hopefully will be just as useful as the print edition (if not more so). Second, every valuation in this book will be put on the web site that will accompany this book (www.damodaran.com), as will a significant number of datasets and spreadsheets. In fact, the valuations in the book will be updated online, allowing the book to have a much closer link to real-time valuations. In the process of presenting and discussing the various aspects of valuation, I have tried to adhere to four basic principles. First, I have attempted to be as comprehensive as possible in covering the range of valuation models that are available to an analyst doing a valuation, while presenting the common elements in these models and providing a framework that can be used to pick the right model for any valuation scenario. Second, the models are presented with real-world examples, warts and all, so as to capture some of the problems inherent in applying these models. There is the obvious danger that some of these valuations will appear to be hopelessly wrong in hindsight, but this cost is well worth the benefits. Third, in keeping with my belief that valuation models are universal and not market-specific, illustrations from markets outside the United States are interspersed throughout the book. Finally, I have tried to make the book as modular as possible, enabling a reader to pick and choose sections of the book to read, without a significant loss of continuity.
14
CHAPTER 1 Introduction to Valuation Every asset, financial as well as real, has a value. The key to successfully investing in and managing these assets lies in understanding not only what the value is, but the sources of the value. Every asset can be valued, but some assets are easier to value than others, and the details of valuation will vary from case to case. Thus, valuing of a real estate property will require different information and follow a different format than valuing a publicly traded stock. What is surprising, however, is not the differences in techniques across assets, but the degree of similarity in the basic principles of valuation. There is uncertainty associated with valuation. Often that uncertainty comes from the asset being valued, though the valuation model may add to that uncertainty. This chapter lays out a philosophical basis for valuation, together with a discussion of how valuation is or can be used in a variety of frameworks, from portfolio management to corporate finance.
A PHILOSOPHICAL BASIS FOR VALUATION It was Oscar Wilde who described a cynic as one who “knows the price of everything, but the value of nothing.” He could very well have been describing some analysts and many investors, a surprising number of whom subscribe to the “bigger fool” theory of investing, which argues that the value of an asset is irrelevant as long as there is a “bigger fool” around willing to buy the asset from them. While this may provide a basis for some profits, it is a dangerous game to play, since there is no guarantee that such an investor will still be around when the time to sell comes. A postulate of sound investing is that an investor does not pay more for an asset than it's worth. This statement may seem logical and obvious, but it is forgotten and rediscovered at some time in every generation and in every market. There are those who are disingenuous enough to argue that value is in the eye of the beholder, and that any price can be justified if there are other investors willing to pay that price. That is patently absurd. Perceptions may be all that matter when the asset is a painting or a sculpture, but investors do not (and should not) buy most assets for aesthetic or emotional reasons; financial assets are acquired for the cash flows expected on them. Consequently, perceptions of value have to be backed up by reality, which implies that the price that is paid for any asset should reflect the cash flows it is expected to generate. The models of valuation described in this book attempt to relate value to the level and expected growth of these cash flows. There are many areas in valuation where there is room for disagreement, including how to estimate true value and how long it will take for prices to adjust to true value. But there is one point on which there can be no disagreement: Asset prices cannot be justified by merely using the argument that there will be other investors around willing to pay those prices.
GENERALITIES ABOUT VALUATION Like all analytical disciplines, valuation has developed its own set of myths over time. This section examines and debunks some of these myths.
Myth 1: Since valuation models are quantitative, valuation is objective. Valuation is neither the science that some of its proponents make it out to be nor the objective search for true value that idealists would like it to become. The models that we use in valuation may be quantitative, but the inputs leave plenty of room for subjective judgments. Thus, the final value that we obtain from these models is colored by the bias that we bring into the process. In fact, in many valuations, the price gets set first and the valuation follows. The obvious solution is to eliminate all bias before starting on a valuation, but this is easier said than done. Given the exposure we have to external information, analyses, and opinions about a firm, it is unlikely that we embark on most valuations without some bias. There are two ways of reducing the bias in the process. The first is to avoid taking strong public positions on the value of a firm before the valuation is complete. In far too many cases, the decision on whether a firm is under- or overvalued precedes the actual valuation, 1 leading to seriously biased analyses. The second is to minimize, prior to the valuation, the stake we have in whether the firm is under- or overvalued. Institutional concerns also play a role in determining the extent of bias in valuation. For instance, it is an acknowledged fact that equity research analysts are more likely to issue buy rather than sell recommendations 2 (i.e., they are more likely to find firms to be undervalued than overvalued). This can be traced partly to the difficulties analysts face in obtaining access and collecting information on firms that they have issued sell recommendations on, and partly to pressure that they face from portfolio managers, some of whom might have large positions in the stock. In recent years, this trend has been exacerbated by the pressure on equity research 15
analysts to deliver investment banking business. When using a valuation done by a third party, the biases of the analyst(s) should be considered before decisions are made on its basis. For instance, a self-valuation done by a target firm in a takeover is likely to be positively biased. While this does not make the valuation worthless, it suggests that the analysis should be viewed with skepticism.
BIAS IN EQUITY RESEARCH The lines between equity research and salesmanship blur most in periods that are characterized by “irrational exuberance.” In the late 1990s, the extraordinary surge of market values in the companies that comprised the new economy saw a large number of equity research analysts, especially on the sell side, step out of their roles as analysts and become cheerleaders for these stocks. While these analysts might have been well-meaning in their recommendations, the fact that the investment banks that they worked for were leading the charge on initial public offerings from these firms exposed them to charges of bias and worse. In 2001, the crash in the market values of new economy stocks and the anguished cries of investors who had lost wealth in the crash created a firestorm of controversy. There were congressional hearings where legislators demanded to know what analysts knew about the companies they recommended and when the knew it, statements from the Securities and Exchange Commision (SEC) about the need for impartiality in equity research, and decisions taken by some investment banks to create at least the appearance of objectivity. Investment banks even created Chinese walls to separate their investment bankers from their equity research analysts. While that technical separation has helped, the real source of bias—the intermingling of banking business, trading, and investment advice—has not been touched. Should there be government regulation of equity research? It would not be wise, since regulation tends to be heavy-handed and creates side costs that seem quickly to exceed the benefits. A much more effective response can be delivered by portfolio managers and investors. Equity research that creates the potential for bias should be discounted or, in egregious cases, even ignored. Alternatively, new equity research firms that deliver only investment advice can meet a need for unbiased valuations.
Myth 2: A well-researched and well-done valuation is timeless. The value obtained from any valuation model is affected by firm-specific as well as marketwide information. As a consequence, the value will change as new information is revealed. Given the constant flow of information into financial markets, a valuation done on a firm ages quickly and has to be updated to reflect current information. This information may be specific to the firm, affect an entire sector, or alter expectations for all firms in the market. The most common example of firm-specific information is an earnings report that contains news not only about a firm's performance in the most recent time period but, even more importantly, about the business model that the firm has adopted. The dramatic drop in value of many new economy stocks from 1999 to 2001 can be traced, at least partially, to the realization that these firms had business models that might deliver customers but not earnings, even in the long term. We have seen social media companies like Linkedin and Zynga received enthusiastic market responses in 2010, and it will be interesting to see if history repeats itself. These companies offer tremendous promise because of their large member bases, but they are still in the nascent stages of commercializing that promise. In some cases, new information can affect the valuations of all firms in a sector. Thus, financial service companies that were valued highly in early 2008, on the assumption that the high growth and returns from the prior years would continue into the future, were valued much less in early 2009, as the banking crisis of 2008 laid bare the weaknesses and hidden risks in their businesses. Finally, information about the state of the economy and the level of interest rates affects all valuations in an economy. A weakening in the economy can lead to a reassessment of growth rates across the board, though the effect on earnings is likely to be largest at cyclical firms. Similarly, an increase in interest rates will affect all investments, though to varying degrees. When analysts change their valuations, they will undoubtedly be asked to justify them, and in some cases the fact that valuations change over time is viewed as a problem. The best response is the one that John Maynard Keynes gave when he was criticized for changing his position on a major economic issue: “When the facts change, I change my mind. And what do you do, sir?”
Myth 3: A good valuation provides a precise estimate of value. Even at the end of the most careful and detailed valuation, there will be uncertainty about the final numbers, colored as they are by assumptions that we make about the future of the company and the economy. It is unrealistic to expect or demand absolute certainty in valuation, since cash flows and discount rates are estimated. This also means that analysts have to give themselves a reasonable margin for error in making recommendations on the basis of valuations. The degree of precision in valuations is likely to vary widely across investments. The valuation of a large and mature company with a long financial history will usually be much more precise than the valuation of a young company in a sector in turmoil. If this latter company happens to operate in an emerging market, with additional disagreement about the future of the market thrown into the mix, the uncertainty is magnified. Later in this book, in Chapter 23, we argue that the difficulties associated with valuation can be related to where a firm is in the life cycle. Mature firms tend to be easier to value than growth firms, and young start-up companies are more difficult to 16
value than companies with established products and markets. The problems are not with the valuation models we use, though, but with the difficulties we run into in making estimates for the future. Many investors and analysts use the uncertainty about the future or the absence of information to justify not doing full-fledged valuations. In reality, though, the payoff to valuation is greatest in these firms.
Myth 4: The more quantitative a model, the better the valuation. It may seem obvious that making a model more complete and complex should yield better valuations; but it is not necessarily so. As models become more complex, the number of inputs needed to value a firm tends to increase, bringing with it the potential for input errors. These problems are compounded when models become so complex that they become “black boxes” where analysts feed in numbers at one end and valuations emerge from the other. All too often when a valuation fails, the blame gets attached to the model rather than the analyst. The refrain becomes “It was not my fault. The model did it.” There are three important points that need to be made about all valuation. The first is to adhere to the principle of parsimony, which essentially states that you do not use more inputs than you absolutely need to value an asset. The second is to recognize that there is a trade-off between the additional benefits of building in more detail and the estimation costs (and error) with providing the detail. The third is to understand that models don't value companies—you do. In a world where the problem that you often face in valuations is not too little information but too much, and separating the information that matters from the information that does not is almost as important as the valuation models and techniques that you use to value a firm.
Myth 5: To make money on valuation, you have to assume that markets are inefficient (but that they will become efficient). Implicit in the act of valuation is the assumption that markets make mistakes and that we can find these mistakes, often using information that tens of thousands of other investors have access to. Thus, it seems reasonable to say that those who believe that markets are inefficient should spend their time and resources on valuation whereas those who believe that markets are efficient should take the market price as the best estimate of value. This statement, though, does not reflect the internal contradictions in both positions. Those who believe that markets are efficient may still feel that valuation has something to contribute, especially when they are called on to value the effect of a change in the way a firm is run or to understand why market prices change over time. Furthermore, it is not clear how markets would become efficient in the first place if investors did not attempt to find under- and over-valued stocks and trade on these valuations. In other words, a precondition for market efficiency seems to be the existence of millions of investors who believe that markets are not efficient. On the other hand, those who believe that markets make mistakes and buy or sell stocks on that basis must believe that ultimately markets will correct these mistakes (i.e., become efficient), because that is how they make their money. This is therefore a fairly self-serving definition of inefficiency—markets are inefficient until you take a large position in the stock that you believe to be mispriced, but they become efficient after you take the position. It is best to approach the issue of market efficiency as a skeptic. Recognize that on the one hand markets make mistakes but, on the other, finding these mistakes requires a combination of skill and luck. This view of markets leads to the following conclusions: First, if something looks too good to be true—a stock looks obviously undervalued or overvalued—it is probably not true. Second, when the value from an analysis is significantly different from the market price, start off with the presumption that the market is correct; then you have to convince yourself that this is not the case before you conclude that something is over- or undervalued. This higher standard may lead you to be more cautious in following through on valuations, but given the difficulty of beating the market, this is not an undesirable outcome.
Myth 6: The product of valuation (i.e., the value) is what matters; the process of valuation is not important. As valuation models are introduced in this book, there is the risk of focusing exclusively on the outcome (i.e., the value of the company and whether it is under- or overvalued), and missing some valuable insights that can be obtained from the process of the valuation. The process can tell us a great deal about the determinants of value and help us answer some fundamental questions: What is the appropriate price to pay for high growth? What is a brand name worth? How important is it to improve returns on projects? What is the effect of profit margins on value? Since the process is so informative, even those who believe that markets are efficient (and that the market price is therefore the best estimate of value) should be able to find some use for valuation models.
THE ROLE OF VALUATION Valuation is useful in a wide range of tasks. The role it plays, however, is different in different arenas. The following section lays out the relevance of valuation in portfolio management, in acquisition analysis, and in corporate 17
finance.
Valuation in Portfolio Management The role that valuation plays in portfolio management is determined in large part by the investment philosophy of the investor. Valuation plays a minimal role in portfolio management for a passive investor, whereas it plays a larger role for an active investor. Even among active investors, the nature and the role of valuation are different for different types of active investment. Market timers should use valuation much less than investors who pick stocks for the long term, and their focus is on market valuation rather than on firm-specific valuation. Among stock pickers valuation plays a central role in portfolio management for fundamental analysts and a peripheral role for technical analysts.
Fundamental Analysts The underlying theme in fundamental analysis is that the true value of the firm can be related to its financial characteristics—its growth prospects, risk profile, and cash flows. Any deviation from this true value is a sign that a stock is under- or overvalued. It is a long-term investment strategy, and the assumptions underlying it are: The relationship between value and the underlying financial factors can be measured. The relationship is stable over time. Deviations from the relationship are corrected in a reasonable time period. Valuation is the central focus in fundamental analysis. Some analysts use discounted cash flow models to value firms, while others use multiples such as the price-earnings and price–book value ratios. Since investors using this approach hold a large number of undervalued stocks in their portfolios, their hope is that, on average, these portfolios will do better than the market.
Franchise Buyers The philosophy of a franchise buyer is best expressed by an investor who has been very successful at it—Warren Buffett. “We try to stick to businesses we believe we understand,” Mr. Buffett writes.3 “That means they must be relatively simple and stable in character. If a business is complex and subject to constant change, we're not smart enough to predict future cash flows.” Franchise buyers concentrate on a few businesses they understand well and attempt to acquire undervalued firms. Often, as in the case of Mr. Buffett, franchise buyers wield influence on the management of these firms and can change financial and investment policy. As a long-term strategy, the underlying assumptions are that: Investors who understand a business well are in a better position to value it correctly. These undervalued businesses can be acquired without driving the price above the true value and sometimes at a bargain. Valuation plays a key role in this philosophy, since franchise buyers are attracted to a particular business because they believe it is undervalued. They are also interested in how much additional value they can create by restructuring the business and running it right.
Chartists Chartists believe that prices are driven as much by investor psychology as by any underlying financial variables. The information available from trading—price movements, trading volume, short sales, and so forth—gives an indication of investor psychology and future price movements. The assumptions here are that prices move in predictable patterns, that there are not enough marginal investors taking advantage of these patterns to eliminate them, and that the average investor in the market is driven more by emotion than by rational analysis. While valuation does not play much of a role in charting, there are ways in which an enterprising chartist can incorporate it into analysis. For instance, valuation can be used to determine support and resistance lines4 on price charts.
Information Traders Prices move on information about the firm. Information traders attempt to trade in advance of new information or shortly after it is revealed to financial markets, buying on good news and selling on bad. The underlying assumption is that these traders can anticipate information announcements and gauge the market reaction to them better than the average investor in the market. For an information trader, the focus is on the relationship between information and changes in value, rather than on value per se. Thus an information trader may buy stock in even an overvalued firm if he or she believes that the next information announcement is going to cause the price to go up because it contains better than expected news. 18
If there is a relationship between how undervalued or overvalued a company is and how its stock price reacts to new information, then valuation could play a role in investing for an information trader.
Market Timers Market timers note, with some legitimacy, that the payoff to calling turns in markets is much greater than the returns from stock picking. They argue that it is easier to predict market movements than to select stocks and that these predictions can be based on factors that are observable. While valuation of individual stocks may not be of any use to a market timer, market timing strategies can use valuation in at least two ways: 1. The overall market itself can be valued and compared to the current level. 2. A valuation model can be used to value all stocks, and the results from the across all stocks be used to determine whether the market is over- or undervalued. For example, as the number of stocks that are overvalued, using a discounted cash flow model, increases relative to the number that are undervalued, there may be reason to believe that the market is overvalued.
Efficient Marketers Efficient marketers believe that the market price at any point in time represents the best estimate of the true value of the firm, and that any attempt to exploit perceived market efficiencies will cost more than it will make in excess profits. They assume that markets aggregate information quickly and accurately, that marginal investors promptly exploit any inefficiencies, and that any inefficiencies in the market are caused by friction, such as transaction costs, and cannot be arbitraged away. For efficient marketers, valuation is a useful exercise to determine why a stock sells for the price that it does. Since the underlying assumption is that the market price is the best estimate of the true value of the company, the objective becomes determining what assumptions about growth and risk are implied in this market price, rather than on finding under- or overvalued firms.
Valuation in Acquisition Analysis Valuation should play a central part in acquisition analysis. The bidding firm or individual has to decide on a fair value for the target firm before making a bid, and the target firm has to determine a reasonable value for itself before deciding to accept or reject the offer. There are also special factors to consider in takeover valuation. First, the effects of synergy on the combined value of the two firms (target plus bidding firm) have to be considered before a decision is made on the bid. Those who suggest that synergy is impossible to value and should not be considered in quantitative terms are wrong. Second, the effects on value of changing management and restructuring the target firm will have to be taken into account in deciding on a fair price. This is of particular concern in hostile takeovers. Finally, there is a significant problem with bias in takeover valuations. Target firms may be overly optimistic in estimating value, especially when the takeovers are hostile and they are trying to convince their stockholders that the offer prices are too low. Similarly, if the bidding firm has decided for strategic reasons to do an acquisition, there may be strong pressure on the analyst to come up with an estimate of value that backs up the acquisition.
Valuation in Corporate Finance If the objective in corporate finance is the maximization of firm value,5 the relationship between financial decisions, corporate strategy, and firm value has to be delineated. In recent years, management consulting firms have started offering companies advice on how to increase value.6 Their suggestions have often provided the basis for the restructuring of these firms. The value of a firm can be directly related to decisions that it makes—on which projects it takes, on how it finances them, and on its dividend policy. Understanding this relationship is key to making value-increasing decisions and to sensible financial restructuring.
CONCLUSION Valuation plays a key role in many areas of finance—in corporate finance, in mergers and acquisitions, and in portfolio management. The models presented in this book will provide a range of tools that analysts in each of these areas will find of use, but the cautionary note sounded in this chapter bears repeating. Valuation is not an objective exercise, and any preconceptions and biases that an analyst brings to the process will find their way into the value. And even the very best valuation will yield an estimate of the value, with a substantial likelihood of you 19
being wrong in your assessment.
QUESTIONS AND SHO RT PROBLEMS In the problems following, use an equity risk premium of 5.5 percent if none is specified. 1. The value of an investment is: a. The present value of the cash flows on the investment. b. Determined by investor perceptions about it. c. Determined by demand and supply. d. Often a subjective estimate, colored by the bias of the analyst. e. All of the above. 2. There are many who claim that value is based on investor perceptions, and perceptions alone, and that cash flows and earnings do not matter. This argument is flawed because: a. Value is determined by earnings and cash flows, and investor perceptions do not matter. b. Perceptions do matter, but they can change. Value must be based on something more substantial. c. Investors are irrational. Therefore, their perceptions should not determine value. d. Value is determined by investor perceptions, but it is also determined by the underlying earnings and cash flows. Perceptions must be based on reality. 3. You use a valuation model to arrive at a value of $15 for a stock. The market price of the stock is $25. The difference may be explained by: a. A market inefficiency; the market is overvaluing the stock. b. The use of the wrong valuation model to value the stock. c. Errors in the inputs to the valuation model. d. All of the above. This is most visible in takeovers, where the decision to acquire a firm often seems to precede the valuation of the firm. It should come as no surprise, therefore, that the analysis almost invariably supports the decision. 1
In most years buy recommendations outnumber sell recommendations by a margin of 10 to 1. In recent years this trend has become even stronger. 2
3
This is extracted from Mr. Buffett's letter to stockholders in Berkshire Hathaway for 1993.
On a chart, the support line usually refers to a lower bound below which prices are unlikely to move, and the resistance line refers to the upper bound above which prices are unlikely to venture. While these levels are usually estimated using past prices, the range of values obtained from a valuation model can be used to determine these levels (i.e., the maximum value will become the resistance line and the minimum value will become the support line). 4
5
Most corporate financial theory is constructed on this premise.
The motivation for this has been the fear of hostile takeovers. Companies have increasingly turned to “value consultants” to tell them how to restructure, increase value, and avoid being taken over. 6
20
CHAPTER 2 Approaches to Valuation Analysts use a wide range of models in practice, ranging from the simple to the sophisticated. These models often make very different assumptions, but they do share some common characteristics and can be classified in broader terms. There are several advantages to such a classification: It makes it easier to understand where individual models fit into the big picture, why they provide different results, and when they have fundamental errors in logic. In general terms, there are three approaches to valuation. The first, discounted cash flow (DCF) valuation, relates the value of an asset to the present value (PV) of expected future cash flows on that asset. The second, relative valuation, estimates the value of an asset by looking at the pricing of comparable assets relative to a common variable such as earnings, cash flows, book value, or sales. The third, contingent claim valuation, uses option pricing models to measure the value of assets that share option characteristics. Some of these assets are traded financial assets like warrants, and some of these options are not traded and are based on real assets, (projects, patents, and oil reserves are examples). The latter are often called real options. There can be significant differences in outcomes, depending on which approach is used. One of the objectives in this book is to explain the reasons for such differences in value across different models, and to help in choosing the right model to use for a specific task.
DISCOUNTED CASH FLOW VALUATION While discounted cash flow valuation is only one of the three ways of approaching valuation and most valuations done in the real world are relative valuations, it is the foundation on which all other valuation approaches are built. To do relative valuation correctly, we need to understand the fundamentals of discounted cash flow valuation. To apply option pricing models to value assets, we often have to begin with a discounted cash flow valuation. This is why so much of this book focuses on discounted cash flow valuation. Anyone who understands its fundamentals will be able to analyze and use the other approaches. This section considers the basis of this approach, a philosophical rationale for discounted cash flow valuation, and an examination of the different subapproaches to discounted cash flow valuation.
Basis for Discounted Cash Flow Valuation This approach has its foundation in the present value rule, where the value of any asset is the present value of expected future cash flows on it.
where n = Life of the asset CFt = Cash flow in period t r = Discount rate reflecting the riskiness of the estimated cash flows The cash flows will vary from asset to asset—dividends for stocks, coupons (interest) and the face value for bonds, and after-tax cash flows for a real project. The discount rate will be a function of the riskiness of the estimated cash flows, with higher rates for riskier assets and lower rates for safer projects. You can in fact think of discounted cash flow valuation on a continuum. At one end of the spectrum you have the default-free zero coupon bond, with a guaranteed cash flow in the future. Discounting this cash flow at the riskless rate should yield the value of the bond. A little further up the risk spectrum are corporate bonds where the cash flows take the form of coupons and there is default risk. These bonds can be valued by discounting the cash flows at an interest rate that reflects the default risk. Moving up the risk ladder, we get to equities, where there are expected cash flows with substantial uncertainty around the expectations. The value here should be the present value of the expected cash flows at a discount rate that reflects the uncertainty.
Underpinnings of Discounted Cash Flow Valuation In discounted cash flow valuation, we try to estimate the intrinsic value of an asset based on its fundamentals. What is intrinsic value? For lack of a better definition, consider it the value that would be attached to the firm by an unbiased analyst, who not only estimates the expected cash flows for the firm correctly, given the information available at the time, but also attaches the right discount rate to value these cash flows. Hopeless though the task of estimating intrinsic value may seem to be, especially when valuing young companies with substantial uncertainty about the future, making the best estimates that you can and persevering to estimate value can still pay off because markets make mistakes. While market prices can deviate from intrinsic value (estimated based on fundamentals), you are hoping that the two will converge sooner rather than later. 21
Categorizing Discounted Cash Flow Models There are literally thousands of discounted cash flow models in existence. Investment banks or consulting firms often claim that their valuation models are better or more sophisticated than those used by their contemporaries. Ultimately, however, discounted cash flow models can vary only a couple of dimensions.
Equity Valuation and Firm Valuation There are two paths to valuation in a business: The first is to value just the equity stake in the business, while the second is to value the entire business, which includes, besides equity, the other claimholders in the firm (bondholders, preferred stockholders). While both approaches discount expected cash flows, the relevant cash flows and discount rates are different under each. Figure 2.1 capture s the essence of the two approaches. Figure 2.1 Equity versus Firm Valuation
The value of equity is obtained by discounting expected cash flows to equity (i.e., the residual cash flows after meeting all expenses, reinvestment needs, tax obligations, and interest and principal payments) at the cost of equity (i.e., the rate of return required by equity investors in the firm).
where n = Life of the asset CF to equityt = Expected cash flow to equity in period t ke = Cost of equity The dividend discount model is a special case of equity valuation, where the value of equity is the present value of expected future dividends. The value of the firm is obtained by discounting expected cash flows to the firm (i.e., the residual cash flows after meeting all operating expenses, reinvestment needs, and taxes, but prior to any payments to either debt or equity holders) at the weighted average cost of capital (WACC), which is the cost of the different components of financing used by the firm, weighted by their market value proportions.
where n = Life of the asset CF to firmt = Expected cash flow to firm in period t WACC = Weighted average cost of capital 22
While these approaches use different definitions of cash flow and discount rates, they will yield consistent estimates of value for equity as long as you are consistent in your assumptions in valuation. The key error to avoid is mismatching cash flows and discount rates, since discounting cash flows to equity at the cost of capital will lead to an upwardly biased estimate of the value of equity, while discounting cash flows to the firm at the cost of equity will yield a downwardly biased estimate of the value of the firm. Illustration 2.1 shows the equivalence of equity and firm valuation.
ILLUSTRATION 2.1: Effects of Mismatching Cash Flows and Discount Rates Assume that you are analyzing a company with the following cash flows for the next five years. Assume also that the cost of equity is 13.625% and the firm can borrow long term at 10%. (The tax rate for the firm is 50%.) The current market value of equity is $1,073, and the value of debt outstanding is $800.
The cost of equity is given as an input and is 13.625%, and the after-tax cost of debt is 5%.
Given the market values of equity and debt, we can estimate the cost of capital.
METHOD 1: DISCOUNT CASH FLOWS TO EQUITY AT COST OF EQUITY TO GET VALUE OF EQUITY We discount cash flows to equity at the cost of equity:
METHOD 2: DISCOUNT CASH FLOWS TO FIRM AT COST OF CAPITAL TO GET VALUE OF FIRM
Note that the value of equity is $1,073 under both approaches. It is easy to make the mistake of discounting cash flows to equity at the cost of capital or the cash flows to the firm at the cost of equity. ERROR 1: DISCOUNT CASH FLOWS TO EQUITY AT COST OF CAPITAL TO GET TOO HIGH A VALUE FOR EQUITY
ERROR 2: DISCOUNT CASH FLOWS TO FIRM AT COST OF EQUITY TO GET TOO LOW A VALUE FOR THE FIRM
The effects of using the wrong discount rate are clearly visible in the last two calculations (Error 1 and Error 2). When the cost of capital is mistakenly used to discount the cash flows to equity, the value of equity increases by $175 over its true value ($1,073). When the cash flows to the firm are erroneously discounted at the cost of equity, the value of the firm is understated by $260. It must be pointed out, though, that getting the values of equity to agree with the firm and equity valuation approaches can be much more difficult in practice than in this example. We return to this subject in Chapters 14 and 15 and consider the assumptions that we need to make to arrive at this result.
Cost of Capital versus APV Approaches In Figure 2.1, we noted that a firm can finance its assets, using either equity or debt. What are the effects of using debt on value? On the plus side, the tax deductibility of interest expenses provides a tax subsidy or benefit to the firm, which increases with the tax rate faced by the firm on its income. On the minus side, debt does increase the likelihood that the firm will default on its commitments and be forced into bankruptcy. The net effect can be positive, neutral or negative. In the cost of capital approach, we capture the effects of debt in the discount rate: 23
Cost of capital = Cost of equity(Proportion of equity used to fund business) + Pretax cost of debt (1 – Tax rate) (Proportion of debt used to fund business)
The cash flows discounted are predebt cash flows and do not include any of the tax benefits of debt (since that would be double counting). In a variation, called the adjusted present value (APV) approach, we separate the effects on value of debt financing from the value of the assets of a business. Thus, we start by valuing the business as if it were all equity funded and assess the effect of debt separately, by first valuing the tax benefits from the debt and then subtracting out the expected bankruptcy costs. Value of business = Value of business with 100% equity financing + Present value of expected tax benefits of debt - Expected bankruptcy costs
While the two approaches take different tacks to evaluating the value added or destroyed by debt, they will provide the same estimate of value, if we are consistent in our assumptions about cash flows and risk. In chapter 15, we will return to examine these approaches in more detail.
Total Cash Flow versus Excess Cash Flow Models The conventional discounted cash flow model values an asset by estimating the present value of all cash flows generated by that asset at the appropriate discount rate. In excess return (and excess cash flow) models, only cash flows earned in excess of the required return are viewed as value creating, and the present value of these excess cash flows can be added to the amount invested in the asset to estimate its value. To illustrate, assume that you have an asset in which you invested $100 million and that you expect to generate $12 million in after-tax cash flows in perpetuity. Assume further that the cost of capital on this investment is 10 percent. With a total cash flow model, the value of this asset can be estimated as follows: Value of asset = $12 million/.1 = $120 million With an excess return model, we would first compute the excess return made on this asset: Excess return = Cash flow earned – Cost of capital × Capital invested in asset = $12 million – .10 × $100 million = $2 million
A SIMPLE TEST OF CASH FLOWS There is a simple test that can be employed to determine whether the cash flows being used in a valuation are cash flows to equity or cash flows to the firm. If the cash flows that are being discounted are after interest expenses (and principal payments), they are cash flows to equity and the discount rate used should be the cost of equity. If the cash flows that are discounted are before interest expenses and principal payments, they are usually cash flows to the firm. Needless to say, there are other items that need to be considered when estimating these cash flows, and they are considered in extensive detail in the coming chapters.
We then add the present value of these excess returns to the investment in the asset: Value of asset = Present value of excess return + Investment in the asset = $2 million/.1 + $100 million = $120 million
Note that the answers in the two approaches are equivalent. Why, then, would we want to use an excess return model? By focusing on excess returns, this model brings home the point that it is not earnings per se that create value, but earnings in excess of a required return. Chapter 32 considers special versions of these excess return models. As in this simple example, with consistent assumptions, total cash flow and excess return models are equivalent.
Applicability and Limitations of Discounted Cash Flow Valuation Discounted cash flow valuation is based on expected future cash flows and discount rates. Given these estimation requirements, this approach is easiest to use for assets (firms) whose cash flows are currently positive and can be estimated with some reliability for future periods, and where a proxy for risk that can be used to obtain discount rates is available. The further we get from this idealized setting, the more difficult (and more useful) discounted cash flow valuation becomes. Here are some scenarios where discounted cash flow valuation might run into trouble and need to be adapted.
Firms in Trouble A distressed firm generally has negative earnings and cash flows, and expects to lose money for some time in the future. For these firms, estimating future cash flows is difficult to do, since there is a strong probability of 24
bankruptcy. For firms that are expected to fail, discounted cash flow valuation does not work very well, since the method values the firm as a going concern providing positive cash flows to its investors. Even for firms that are expected to survive, cash flows will have to be estimated until they turn positive, since obtaining a present value of negative cash flows will yield a negative value for equity1 or for th e firm. We will examine these firms in more detail in chapters 22 and 30.
Cyclical Firms The earnings and cash flows of cyclical firms tend to follow the economy—rising during economic booms and falling during recessions. If discounted cash flow valuation is used on these firms, expected future cash flows are usually smoothed out, unless the analyst wants to undertake the onerous task of predicting the timing and duration of economic recessions and recoveries. In the depths of a recession many cyclical firms look like troubled firms, with negative earnings and cash flows. Estimating future cash flows then becomes entangled with analyst predictions about when the economy will turn and how strong the upturn will be, with more optimistic analysts arriving at higher estimates of value. This is unavoidable, but the economic biases of the analysts have to be taken into account before using these valuations.
Firms with Unutilized Assets Discounted cash flow valuation reflects the value of all assets that produce cash flows. If a firm has assets that are unutilized (and hence do not produce any cash flows), the value of these assets will not be reflected in the value obtained from discounting expected future cash flows. The same caveat applies, in lesser degree, to underutilized assets, since their value will be understated in discounted cash flow valuation. While this is a problem, it is not insurmountable. The value of these assets can always be obtained externally2 and added t o the value obtained from discounted cash flow valuation. Alternatively, the assets can be valued as though they are used optimally.
Firms with Patents or Product Options Firms sometimes have unutilized patents or licenses that do not produce any current cash flows and are not expected to produce cash flows in the near future, but are valuable nevertheless. If this is the case, the value obtained from discounting expected cash flows to the firm will understate the true value of the firm. Again, the problem can be overcome, by valuing these assets in the open market or by using option pricing models, and then adding the value obtained from discounted cash flow valuation. Chapter 28 examines the use of option pricing models to value patents.
Firms in the Process of Restructuring Firms in the process of restructuring often sell some of their assets, acquire other assets, and change their capital structure and dividend policy. Some of them also change their ownership structure (going from publicly traded to private status and vice versa) and management compensation schemes. Each of these changes makes estimating future cash flows more difficult and affects the riskiness of the firm. Using historical data for such firms can give a misleading picture of the firm's value. However, these firms can be valued, even in the light of the major changes in investment and financing policy, if future cash flows reflect the expected effects of these changes and the discount rate is adjusted to reflect the new business and financial risk in the firm. Chapter 31 takes a closer look at how value can be altered by changing the way a business is run.
Firms Involved in Acquisitions There are at least two specific issues relating to acquisitions that need to be taken into account when using discounted cash flow valuation models to value target firms. The first is the thorny one of whether there is synergy in the merger and how its value can be estimated. To do so will require assumptions about the form the synergy will take and its effect on cash flows. The second, especially in hostile takeovers, is the effect of changing management on cash flows and risk. Again, the effect of the change can and should be incorporated into the estimates of future cash flows and discount rates and hence into value. Chapter 25 looks at the value of synergy and control in acquisitions.
Private Firms The biggest problem in using discounted cash flow valuation models to value private firms is the measurement of risk (to use in estimating discount rates), since most risk/return models require that risk parameters be estimated from historical prices on the asset being analyzed and make assumptions about the profiles of investors in the firm that may not fit private businesses. One solution is to look at the riskiness of comparable firms that are publicly traded. The other is to relate the measure of risk to accounting variables, which are available for the private firm. Chapter 24 looks at adaptations to valuation models that are needed to value private businesses. The point is not that discounted cash flow valuation cannot be done in these cases, but that we have to be flexible enough to adapt our models. The fact is that valuation is simple for firms with well-defined assets that generate 25
cash flows that can be easily forecasted. The real challenge in valuation is to extend the valuation framework to cover firms that vary to some extent or the other from this idealized framework. Much of this book is spent considering how to value such firms.
RELATIVE VALUATION While we tend to focus most on discounted cash flow valuation when discussing valuation, the reality is that most valuations are relative valuations. The values of most assets, from the house you buy to the stocks you invest in, are based on how similar assets are priced in the marketplace. This section begins with a basis for relative valuation, moves on to consider the underpinnings of the model, and then considers common variants within relative valuation.
Basis for Relative Valuation In relative valuation, the value of an asset is derived from the pricing of comparable assets, standardized using a common variable such as earnings, cash flows, book value, or revenues. One illustration of this approach is the use of an industry-average price-earnings ratio to value a firm, the assumption being that the other firms in the industry are comparable to the firm being valued and that the market, on average, prices these firms correctly. Another multiple in wide use is the price–book value ratio, with firms selling at a discount on book value relative to comparable firms being considered undervalued. Revenue multiple are also used to value firms, with the average price-sales ratios of firms with similar characteristics being used for comparison. While these three multiples are among the most widely used, there are others that also play a role in analysis—EV to EBITDA, EV to invested capital, and market value to replacement value (Tobin's Q), to name a few.
Underpinnings of Relative Valuation Unlike discounted cash flow valuation, which is a search for intrinsic value, relative valuation relies much more on the market being right. In other words, we assume that the market is correct in the way it prices stocks on average, but that it makes errors on the pricing of individual stocks. We also assume that a comparison of multiples will allow us to identify these errors, and that these errors will be corrected over time. The assumption that markets correct their mistakes over time is common to both discounted cash flow and relative valuation, but those who use multiples and comparables to pick stocks argue, with some basis, that errors made in pricing individual stocks in a sector are more noticeable and more likely to be corrected quickly. For instance, they would argue that a software firm that trades at a price-earnings ratio of 10 when the rest of the sector trades at 25 times earnings is clearly undervalued and that the correction toward the sector average should occur sooner rather than later. Proponents of discounted cash flow valuation would counter that this is small consolation if the entire sector is overpriced by 50 percent.
Categorizing Relative Valuation Models Analysts and investors are endlessly inventive when it comes to using relative valuation. Some compare multiples across companies, while other compare the multiple of a company to the multiples it used to trade at in the past. While most relative valuations are based on the pricing of comparable assets at the same time, there are some relative valuations that are based on fundamentals.
Fundamentals versus Comparables In discounted cash flow valuation, the value of a firm is determined by its expected cash flows. Other things remaining equal, higher cash flows, lower risk, and higher growth should yield higher value. Some analysts who use multiples go back to these discounted cash flow models to extract multiples. Other analysts compare multiples across firms or time and make explicit or implicit assumptions about how firms are similar or vary on fundamentals.
Using Fundamentals The first approach relates multiples to fundamentals about the firm being valued—growth rates in earnings and cash flows, reinvestment and risk. This approach to estimating multiples is equivalent to using discounted cash flow models, requiring the same information and yielding the same results. Its primary advantage is that it shows the relationship between multiples and firm characteristics, and allows us to explore how multiples change as these characteristics change. For instance, what will be the effect of changing profit margins on the price-sales ratio? What will happen to price-earnings ratios as growth rates decrease? What is the relationship between price–book value ratios and return on equity?
Using Comparables The more common approach to using multiples is to compare how a firm is valued with how similar firms are priced by the market or, in some cases, with how the firm was valued in prior periods. As we see in the later chapters, 26
finding similar and comparable firms is often a challenge, and frequently we have to accept firms that are different from the firm being valued on one dimension or the other. When this is the case, we have to either explicitly or implicitly control for differences across firms on growth, risk, and cash flow measures. In practice, controlling for these variables can range from the naive (using industry averages) to the sophisticated (multivariate regression models where the relevant variables are identified and controlled for).
Cross-Sectional versus Time Series Comparisons In most cases, analysts price stocks on a relative basis, by comparing the multiples they are trading at to the multiples at which other firms in the same business are trading at contemporaneously. In some cases, however, especially for mature firms with long histories, the comparison is done across time.
Cross-Sectional Comparisons When we compare the price-earnings ratio of a software firm to the average price-earnings ratio of other software firms, we are doing relative valuation and we are making cross-sectional comparisons. The conclusions can vary depending on our assumptions about the firm being valued and the comparable firms. For instance, if we assume that the firm we are valuing is similar to the average firm in the industry, we would conclude that it is cheap if it trades at a multiple that is lower than the average multiple. If, however, we assume that the firm being valued is riskier than the average firm in the industry, we might conclude that the firm should trade at a lower multiple than other firms in the business. In short, you cannot compare firms without making assumptions about their fundamentals.
Comparisons across Time If you have a mature firm with a long history, you can compare the multiple it trades at today to the multiple it used to trade at in the past. Thus, Ford Motor Company may be viewed as cheap because it trades at six times earnings, if it has historically traded at 10 times earnings. To make this comparison, however, you have to assume that your firm's fundamentals have not changed over time. For instance, you would expect a high-growth firm's priceearnings ratio to drop over time and its expected growth rate to decrease as it becomes larger. Comparing multiples across time can also be complicated by changes in interest rates and the behavior of the overall market. For instance, as interest rates fall below historical norms and the overall market increases in value, you would expect most companies to trade at much higher multiples of earnings and book value than they have historically.
Applicability and Limitations of Multiples The allure of multiples is that they are simple and easy to relate to. They can be used to obtain estimates of value quickly for firms and assets, and are particularly useful when a large number of comparable firms are traded on financial markets, and the market is, on average, pricing these firms correctly. They tend to be more difficult to use to value unique firms with no obvious comparables, with little or no revenues, and with negative earnings. By the same token, multiples are also easy to misuse and manipulate, especially when comparable firms are used. Given that no two firms are exactly alike in terms of risk and growth, the definition of comparable firms is a subjective one. Consequently, a biased analyst can choose a group of comparable firms to confirm his or her biases about a firm's value. Illustration 2.2 shows an example. While this potential for bias exists with discounted cash flow valuation as well, the analyst in DCF valuation is forced to be much more explicit about the assumptions that determine the final value. With multiples, these assumptions are often left unstated.
ASSET-BASED VALUATION MODELS There are some analysts who add a fourth approach to valuation to the three described in this chapter. They argue that you can value the individual assets owned by a firm and aggregate them to arrive at a firm value—asset-based valuation models. In fact, there are several variants on asset-based valuation models. The first is liquidation value, which is obtained by aggregating the estimated sale proceeds of the assets owned by a firm. The second is replacement cost, where you estimate what it would cost you to replace all of the assets that a firm has today. The third is the simplest: use accounting book value as the measure of the value of the assets, with adjustments to the book value made where necessary. While analysts may use asset-based valuation approaches to estimate value, they are not alternatives to discounted cash flow, relative, or option pricing models since both replacement and liquidation values have to be obtained using one or another of these approaches. Ultimately, all valuation models attempt to value assets; the differences arise in how we identify the assets and how we attach value to each asset. In liquidation valuation, we look only at assets in place and estimate their value based on what similar assets are priced at in the market. In traditional discounted cash flow valuation, we consider all assets and include expected growth potential to arrive at value. The two approaches may, in fact, yield the same values if you have a firm that has no growth potential and the market assessments of value reflect expected cash flows.
ILLUSTRATION 2.2: The Potential for Misuse with Comparable Firms Assume that an analyst is valuing an initial public offering (IPO) of a firm that manufactures computer software. At the same time,3 the priceearnings mult iples of other publicly traded firms manufacturing software are:
27
Firm
Multiple
Adobe Systems
23.2
Autodesk
20.4
Broderbund
32.8
Computer Associates 18.0 Lotus Development
24.1
Microsoft
27.4
Novell
30.0
Oracle
37.8
Software Publishing
10.6
System Software
15.7
Average PE ratio
24.0
While the average PE ratio using the entire sample is 24, it can be changed markedly by removing a couple of firms from the group. For instance, if the two firms with the lowest PE ratios in the group (Software Publishing and System Software) are eliminated from the sample, the average PE ratio increases to 27. If the two firms with the highest PE ratios in the group (Broderbund and Oracle) are removed from the group, the average PE ratio drops to 21.
The other problem with using multiples based on comparable firms is that it builds in errors (overvaluation or undervaluation) that the market might be making in valuing these firms. In Illustration 2.2, for instance, if the market has overvalued all computer software firms, using the average PE ratio of these firms to value an initial public offering will lead to an overvaluation of the IPO stock. In contrast, discounted cash flow valuation is based on firm-specific growth rates and cash flows, so it is less likely to be influenced by market errors in valuation.
CONTINGENT CLAIM VALUATION Perhaps the most revolutionary development in valuation is the acceptance, at least in some cases, that the value of an asset may be greater than the present value of expected cash flows if the cash flows are contingent on the occurrence or nonoccurrence of an event. This acceptance has largely come about because of the development of option pricing models. While these models were initially used to value traded options, there has been an attempt in recent years to extend the reach of these models into more traditional valuation. There are many who argue that assets such as patents or undeveloped reserves are really options and should be valued as such, rather than with traditional discounted cash flow models.
Basis for Approach A contingent claim or option is a claim that pays off only under certain contingencies—if the value of the underlying asset exceeds a prespecified value for a call option or is less than a prespecified value for a put option. Much work has been done in the past 20 years in developing models that value options, and these option pricing models can be used to value any assets that have optionlike features. Figure 2.2 illustrates the payoffs on call and put options as a function of the value of the underlying asset. An option can be valued as a function of the following variables: the current value and the variance in value of the underlying asset, the strike price and the time to expiration of the option, and the riskless interest rate. This was first established by Fischer Black and Myron Scholes in 1972 and has been extended and refined subsequently in numerous variants. While the Black-Scholes option pricing model ignore dividends and assumes that options will not be exercised early, it can be modified to allow for both. A discrete-time variant, the binomial option pricing model, has also been developed to price options. Figure 2.2 Payoff Diagram on Call and Put Options
28
An asset can be valued as an option if the payoffs are a function of the value of an underlying asset. It can be valued as a call option if when that value exceeds a prespecified level the asset is worth the difference. It can be valued as a put option if it gains value as the value of the underlying asset drops below a prespecified level, and if it is worth nothing when the underlying asset's value exceeds that specified level.
Underpinnings of Contingent Claim Valuation The fundamental premise behind the use of option pricing models is that discounted cash flow models tend to understate the value of assets that provide payoffs that are contingent on the occurrence of an event. As a simple example, consider an undeveloped oil reserve belonging to Petrobras. You could value this reserve based on expectations of oil prices in the future, but this estimate would miss the fact that the oil company will develop this reserve only if oil prices go up and will not if oil prices decline. An option pricing model would yield a value that incorporates this right. When we use option pricing models to value assets such as patents and undeveloped natural resource reserves, we are assuming that markets are sophisticated enough to recognize such options and incorporate them into the market price. If the markets do not do so right now, we assume that they will eventually; the payoff to using such models comes about when this correction occurs.
Categorizing Option Pricing Models The first categorization of options is based on whether the underlying asset is a financial asset or a real asset. Most listed options, whether they be options listed on the Chicago Board Options Exchange or callable fixed income securities, are on financial assets such as stocks and bonds. In contrast, options can be on real assets such as commodities, real estate, or even investment projects; such options are often called real options. A second and overlapping categorization is based on whether the underlying asset is traded. The overlap occurs because most financial assets are traded, whereas relatively few real assets are traded. Options on traded assets are generally easier to value, and the inputs to the option pricing models can be obtained from financial markets. Options on nontraded assets are much more difficult to value, since there are no market inputs available on the underlying assets.
Applicability and Limitations of Option Pricing Models There are several direct examples of securities that are options—LEAPS, which are long-term equity options on traded stocks; contingent value rights, which provide protection to stockholders in companies against stock price declines; and warrants, which are long-term call options issued by firms. There are other assets that generally are not viewed as options but still share several option characteristics. Equity, for instance, can be viewed as a call option on the value of the underlying firm, with the face value of debt representing the strike price and the term of the debt measuring the life of the option. A patent can be analyzed as a call option on a product, with the investment outlay needed to get the project going considered the strike price and the patent life becoming the time to expiration of the option. There are limitations in using option pricing models to value long-term options on nontraded assets. The assumptions made about constant variance and dividend yields, which are not seriously contested for short-term options, are much more difficult to defend when options have long lifetimes. When the underlying asset is not traded, the inputs for the value of the underlying asset and the variance in that value cannot be extracted from financial markets and have to be estimated. Thus the final values obtained from these applications of option pricing models have much more estimation error associated with them than the values obtained in their more standard applications (to value short-term traded options).
CONCLUSION There are three basic, though not mutually exclusive, approaches to valuation. The first is discounted cash flow valuation, where cash flows are discounted at a risk-adjusted discount rate to arrive at an estimate of value. The analysis can be done purely from the perspective of equity investors by discounting expected cash flows to equity at the cost of equity, or it can be done from the viewpoint of all claimholders in the firm, by discounting expected cash flows to the firm at the weighted average cost of capital. The second is relative valuation, where the value of an asset is based on the pricing of similar assets. The third is contingent claim valuation, where an asset with the characteristics of an option is valued using an option pricing model. There should be a place for each among the tools available to any analyst interested in valuation.
QUESTIONS AND SHORT PROBLEMS In the problems following, use an equity risk premium of 5.5 percent if none is specified. 29
1. Discounted cash flow valuation is based on the notion that the value of an asset is the present value of the expected cash flows on that asset, discounted at a rate that reflects the riskiness of those cash flows. Specify whether the following statements about discounted cash flow valuation are true or false, assuming that all variables are constant except for the one mentioned: a. As the discount rate increases, the value of an asset increases. True____False____ b. As the expected growth rate in cash flows increases, the value of an asset increases. True____False____ c. As the life of an asset is lengthened, the value of that asset increases. True____False____ d. As the uncertainty about the expected cash flow increases, the value of an asset increases. True____False____ e. An asset with an infinite life (i.e., it is expected to last forever) will have an infinite value. True____False____ 2. Why might discounted cash flow valuation be difficult to do for the following types of firms? a. A private firm, where the owner is planning to sell the firm. b. A biotechnology firm with no current products or sales, but with several promising product patents in the pipeline. c. A cyclical firm during a recession. d. A troubled firm that has made significant losses and is not expected to get out of trouble for a few years. e. A firm that is in the process of restructuring, where it is selling some of its assets and changing its financial mix. f. A firm that owns a lot of valuable land that is currently unutilized. 3. The following are the projected cash flows to equity and to the firm over the next five years:
The firm has a cost of equity of 12% and a cost of capital of 9.94%. Answer the following questions: a. What is the value of the equity in this firm? b. What is the value of the firm? 4. You are estimating the price-earnings multiple to use to value Paramount Corporation by looking at the average price-earnings multiple of comparable firms. The following are the price-earnings ratios of firms in the entertainment business. Firm
PE Ratio
Disney (Walt)
22.09
Time Warner
36.00
King World Productions 14.10 New Line Cinema
26.70
a. What is the average PE ratio? b. Would you use all the comparable firms in calculating the average? Why or why not? c. What assumptions are you making when you use the industry-average PE ratio to value Paramount Corporation? 30
The protec tion of limited liability should ensure that no stock will sell for less than zero. The price of such a stock can never be negative. 1
If these a ssets are traded on external markets, the market prices of these assets can be used in the valuation. If not, the cash flows can be projected, assuming full utilization of assets, and the value can be estimated. 2
3
These were the PE ratios for these firms at the end of 1992.
31
CHAPTER 3 Understanding Financial Statements Financial statements provide the fundamental information that we use to analyze and answer valuation questions. It is important, therefore, that we understand the principles governing these statements by looking at four questions: 1. How valuable are the assets of a firm? The assets of a firm can come in several forms—assets with long lives such as land and buildings, assets with shorter lives such as inventory, and intangible assets that nevertheless produce revenues for the firm such as patents and trademarks. 2. How did the firm raise the funds to finance these assets? In acquiring assets, firms can use the funds of the owners (equity) or borrowed money (debt), and the mix is likely to change as the assets age. 3. How profitable are these assets? A good investment is one that makes a return greater than the cost of funding it. To evaluate whether the investments that a firm has already made are good investments, we need to estimate what returns these investments are producing. 4. How much uncertainty (or risk) is embedded in these assets? While we have not yet directly confronted the issue of risk, estimating how much uncertainty there is in existing investments, and the implications for a firm, is clearly a first step. This chapter looks at the way accountants would answer these questions, and why the answers might be different when doing valuation. Some of these differences can be traced to the differences in objectives: Accountants try to measure the current standing and immediate past performance of a firm, whereas valuation is much more forwardlooking.
THE BASIC ACCOUNTING STATEMENTS There are three basic accounting statements that summarize information about a firm. The first is the balance sheet, shown in Figure 3.1, which summarizes the assets owned by a firm, the value of these assets, and the mix of financing (debt and equity) used to finance these assets at a point in time. Figure 3.1 The Balance Sheet
The next is the income statement, shown in Figure 3.2, which provides information on the reve nues and expenses of the firm, and the resulting income made by the firm, during a period. The period can be a quarter (if it is a quarterly income statement) or a year (if it is an annual report). Figure 3.2 Income Statement
32
Finally, there is the statement of cash flows, shown in Figure 3.3, which specifies the sources and uses o f cash to the firm from operating, investing, and financing activities during a period. The statement of cash flows can be viewed as an attempt to explain what the cash flows during a period were, and why the cash balance changed during the period. Figure 3.3 Statement of Cash Flows
ASSET MEASUREMENT AND VALUATION When analyzing any firm, we want to know the types of assets that it owns, the value of these assets, and the degree of uncertainty about this value. Accounting statements do a reasonably good job of categorizing the assets owned by a firm, a partial job of assessing the value of these assets, and a poor job of reporting uncertainty about asset value. This section begins by looking at the accounting principles underlying asset categorization and measurement, and the limitations of financial statements in providing relevant information about assets.
Accounting Principles Underlying Asset Measurement An asset is any resource that has the potential either to generate future cash inflows or to reduce future cash outflows. While that is a general definition broad enough to cover almost any kind of asset, accountants add a caveat that for a resource to be an asset a firm has to have acquired it in a prior transaction and be able to quantify future benefits with reasonable precision. The accounting view of asset value is to a great extent grounded in the notion of historical cost, which is the original cost of the asset, adjusted upward for improvements made to the 33
asset since purchase and downward for the loss in value associated with the aging of the asset. This historical cost is called the book value. While the generally accepted accounting principles (GAAP) for valuing an asset vary across different kinds of assets, three principles underlie the way assets are valued in accounting statements: 1. An abiding belief in book value as the best estimate of value. Accounting estimates of asset value begin with the book value, and unless a substantial reason is given to do otherwise, accountants view the historical cost as the best estimate of the value of an asset. 2. A distrust of market or estimated value. When a current market value exists for an asset that is different from the book value, accounting convention seems to view this market value with suspicion. The market price of an asset is often viewed as both much too volatile and too easily manipulated to be used as an estimate of value for an asset. This suspicion runs even deeper when a value is estimated for an asset based on expected future cash flows. 3. A preference for underestimating value rather than overestimating it. When there is more than one approach to valuing an asset, accounting convention takes the view that the more conservative (lower) estimate of value should be used rather than the less conservative (higher) estimate of value. Thus, when both market and book value are available for an asset, accounting rules often require that you use the lesser of the two numbers.
Measuring Asset Value The financial statement in which accountants summarize and report asset value is the balance sheet. To examine how asset value is measured, let us begin with the way assets are categorized in the balance sheet. First there are the fixed assets, which include the long-term assets of the firm, such as plant, equipment, land, and buildings. Next, we have the short-term assets of the firm, including inventory (raw materials, work in progress, and finished goods, receivables (summarizing moneys owed to the firm), and cash; these are categorized as current assets. We then have investments in the assets and securities of other firms, which are generally categorized as financial investments. Finally, we have what is loosely categorized as intangible assets. These include not only assets such as patents and trademarks that presumably will create future earnings and cash flows, but also uniquely accounting assets such as goodwill that arise because of acquisitions made by the firm.
Fixed Assets Generally accepted accounting principles (GAAP) in the United States require the valuation of fixed assets at historical cost, adjusted for any estimated loss in value from the aging of these assets. While in theory the adjustments for aging should reflect the loss of earning power of the asset as it ages, in practice they are much more a product of accounting rules and convention, and these adjustments are called depreciation. Depreciation methods can very broadly be categorized into straight line (where the loss in asset value is assumed to be the same every year over its lifetime) and accelerated (where the asset loses more value in the earlier years and less in the later years). While tax rules, at least in the United States, have restricted the freedom that firms have on their choices of asset life and depreciation methods, firms continue to have a significant amount of flexibility on these decisions for reporting purposes. Thus, the depreciation that is reported in the annual reports may not be, and generally is not, the same depreciation that is used in the tax statements. Since fixed assets are valued at book value and are adjusted for depreciation provisions, the value of a fixed asset is strongly influenced by both its depreciable life and the depreciation method used. Many firms in the United States use straight-line depreciation for financial reporting while using accelerated depreciation for tax purposes, since firms can report better earnings with the former, at least in the years right after the asset is acquired.1 In contrast, firms in other cou ntries often use accelerated depreciation for both tax and financial reporting purposes, leading to reported income that is understated relative to that of their U.S. counterparts.
Current Assets Current assets include inventory, cash, and accounts receivable. It is in this category that accountants are most amenable to the use of market value, especially in valuing marketable securities.
Accounts Receivable Accounts receivable represent money owed by entities to the firm on the sale of products on credit. When the Home Depot sells products to building contractors and gives them a few weeks to make their payments, it is creating accounts receivable. The accounting convention is for accounts receivable to be recorded as the amount owed to the firm based on the billing at the time of the credit sale. The only major valuation and accounting issue is when the firm has to recognize accounts receivable that are not collectible. Firms can set aside a portion of their income to cover expected bad debts from credit sales, and accounts receivable will be reduced by this reserve. Alternatively, the bad debts can be recognized as they occur, and the firm can reduce the accounts receivable accordingly. There is the danger, however, that absent a decisive declaration of a bad debt, firms may continue to show as accounts receivable amounts that they know are unlikely ever to be collected. 34
Cash Cash is one of the few assets for which accountants and financial analysts should agree on value. The value of a cash balance should not be open to estimation error. Having said this, we note that fewer and fewer companies actually hold cash in the conventional sense (as currency or as demand deposits in banks). Firms often invest the cash in interest-bearing accounts, commercial paper or in Treasuries so as to earn a return on their investments. In either case, market value can sometimes deviate from book value. While there is minimal default risk in either of these investments, interest rate movements can affect their value. The valuation of marketable securities is examined later in this section.
Inventory Three basic approaches to valuing inventory are allowed by GAAP: first in, first out (FIFO); last in, first out (LIFO); and weighted average. 1. First in, first out (FIFO). Under FIFO, the cost of goods sold is based on the cost of material bought earliest in the period, while the cost of inventory is based on the cost of material bought later in the year. This results in inventory being valued close to current replacement cost. During periods of inflation, the use of FIFO will result in the lowest estimate of cost of goods sold among the three valuation approaches, and the highest net income. 2. Last in, first out (LIFO). Under LIFO, the cost of goods sold is based on the cost of material bought toward the end of the period, resulting in costs that closely approximate current costs. The inventory, however, is valued on the basis of the cost of materials bought earlier in the year. During periods of inflation, the use of LIFO will result in the highest estimate of cost of goods sold among the three approaches, and the lowest net income. 3. Weighted average. Under the weighted average approach, both inventory and the cost of goods sold are based on the average cost of all material bought during the period. When inventory turns over rapidly, this approach will more closely resemble FIFO than LIFO. Firms often adopt the LIFO approach for its tax benefits during periods of high inflation. The cost of goods sold is then higher because it is based on prices paid toward to the end of the accounting period. This, in turn, will reduce the reported taxable income and net income while increasing cash flows. Studies indicate that larger firms with rising prices for raw materials and labor, more variable inventory growth, and an absence of other tax loss carryforwards are much more likely to adopt the LIFO approach. Given the income and cash flow effects of inventory valuation methods, it is often difficult to compare the profitability of firms that use different methods. There is, however, one way of adjusting for these differences. Firms that choose the LIFO approach to value inventories have to specify in a footnote the difference in inventory valuation between FIFO and LIFO, and this difference is termed the LIFO reserve. It can be used to adjust the beginning and ending inventories, and consequently the cost of goods sold, and to restate income based on FIFO valuation.
Investments (Financial) and Marketable Securities In the category of investments and marketable securities, accountants consider investments made by firms in the securities or assets of other firms, as well as other marketable securities, including Treasury bills or bonds. The way in which these assets are valued depends on the way the investment is categorized and the motive behind the investment. In general, an investment in the securities of another firm can be categorized as a minority passive investment, a minority active investment, or a majority active investment, and the accounting rules vary depending on the categorization.
Minority Passive Investments If the securities or assets owned in another firm represent less than 20 percent of the overall ownership of that firm, an investment is treated as a minority passive investment. These investments have an acquisition value, which represents what the firm originally paid for the securities, and often a market value. Accounting principles require that these assets be subcategorized into one of three groups—investments that will be held to maturity, investments that are available for sale, and trading investments. The valuation principles vary for each. For an investment that will be held to maturity, the valuation is at historical cost or book value, and interest or dividends from this investment are shown in the income statement. For an investment that is available for sale, the valuation is at market value, but the unrealized gains or losses are shown as part of the equity in the balance sheet and not in the income statement. Thus, unrealized losses reduce the book value of the equity in the firm, and unrealized gains increase the book value of equity. For a trading investment, the valuation is at market value, and the unrealized gains and losses are shown in the income statement. 35
Firms are allowed an element of discretion in the way they classify investments and, subsequently, in the way they value these assets. This classification ensures that firms such as investment banks, whose assets are primarily securities held in other firms for purposes of trading, revalue the bulk of these assets at market levels each period. This is called marking to market, and provides one of the few instances in which market value trumps book value in accounting statements. Note, however, that this mark-to-market ethos did not provide any advance warning in 2008 to investors in financial service firms of the overvaluation of subprime and mortgage-backed securities.
Minority Active Investments If the securities or assets owned in another firm represent between 20 percent and 50 percent of the overall ownership of that firm, an investment is treated as a minority active investment. While these investments have an initial acquisition value, a proportional share (based on ownership proportion) of the net income and losses made by the firm in which the investment was made is used to adjust the acquisition cost. In addition, the dividends received from the investment reduce the acquisition cost. This approach to valuing investments is called the equity approach. The market value of these investments is not considered until the investment is liquidated, at which point the gain or loss from the sale relative to the adjusted acquisition cost is shown as part of the earnings in that period.
Majority Active Investments If the securities or assets owned in another firm represent more than 50 percent of the overall ownership of that firm, an investment is treated as a majority active investment. In this case, the investment is no longer shown as a financial investment but is instead replaced by the assets and liabilities of the firm in which the investment was made. This approach leads to a consolidation of the balance sheets of the two firms, where the assets and liabilities of the two firms are merged and presented as one balance sheet.2 The share of the equity in the su bsidiary that is owned by other investors is shown as a minority interest on the liability side of the balance sheet. To provide an illustration, assume that Firm A owns 60% of Firm B. Firm A will be required to consolidate 100% of Fir m B's revenues, earnings, and assets into its own financial statements and then show a liability (minority interest) reflecting the accounting estimate of value of the 40% of Firm B's equity that does not belong to it. A similar consolidation occurs in the other financial statements of the firm as well, with the statement of cash flows reflecting the cumulated cash inflows and outflows of the combined firm. This is in contrast to the equity approach used for minority active investments, in which only the dividends received on the investment are shown as a cash inflow in the cash flow statement. Here again, the market value of this investment is not considered until the ownership stake is liquidated. At that point, the difference between the market price and the net value of the equity stake in the firm is treated as a gain or loss for the period.
Intangible Assets Intangible assets include a wide array of assets, ranging from patents and trademarks to goodwill. The accounting standards vary across intangible assets.
Patents and Trademarks Patents and trademarks are valued differently depending on whether they are generated internally or acquired. When patents and trademarks are generated from internal research, the costs incurred in developing the asset are expensed in that period, even though the asset might have a life of several accounting periods. Thus, the intangible asset is not valued in the balance sheet of the firm. In contrast, when an intangible asset is acquired from an external party, it is treated as an asset. Intangible assets have to be amortized over their expected lives, with a maximum amortization period of 40 years. The standard practice is to use straight-line amortization. For tax purposes, however, firms are generally not allowed to amortize goodwill or other intangible assets with no specific lifetime, though recent changes in the tax law allow for some flexibility in this regard.
Goodwill Goodwill is the by-product of acquisitions. When a firm acquires another firm, the purchase price is first allocated to tangible assets, and the excess price is then allocated to any intangible assets such as patents or trade names. Any residual becomes goodwill. While accounting principles suggest that goodwill captures the value of any intangibles that are not specifically identifiable, it is really a reflection of the difference between the book value of assets of the acquired firm and the market value paid in the acquisition. This approach is called purchase accounting, and goodwill is amortized over time. Until 2000, firms that did not want to see this charge against their earnings often used an alternative approach called pooling accounting, in which the purchase price never shows up in the balance sheet. Instead, the book values of the two companies involved in the merger were aggregated to create the consolidated balance of the combined firm. The rules on acquisition accounting have changed substantially in the 36
past decade both in the United States and internationally. Not only is purchase accounting required on all acquisitions, but firms are no longer allowed to automatically amortize goodwill over long periods (as they were used to doing). Instead, acquiring firms are required to reassess the values of the acquired entities every year; if the values have dropped since the acquisition, the value of goodwill must be reduced (impaired) to reflect the decline in value. If the acquired firm's values have gone up, though, the goodwill cannot be increased to reflect this change.3
ILLUSTRATION 3.1: Asset Values for Boeing and the Home Depot in 1998 The following table summarizes asset values, as measured in the balance sheets of Boeing, the aerospace giant, and the Home Depot, a building supplies retailer, at the end of the 1998 financial year (in millions of dollars):
Boeing
Home Depot
Net fixed assets
$ 8,589
$ 8,160
Goodwill
$ 2,312
$
Investments and notes receivable
$
Deferred income taxes
$
411
$
Prepaid pension expense
$ 3,513
$
Customer financing
$ 4,930
$
Other assets
$
542
$
Cash
$ 2,183
$
Short-term marketable investments
$
279
$
Accounts receivables
$ 3,288
$
140
41 $
191
Current Assets
Current portion of customer financing $
62 0 469
781
$
Deferred income taxes
$ 1,495
$
Inventories
$ 8,349
$ 4,293
Other current assets
$
$
Total current assets
$16,375
$ 4,933
Total assets
$36,672
$13,465
109
There are five points worth noting about these asset values: 1. Goodwill. Boeing, which acquired Rockwell in 1996 and McDonnell Douglas in 1997, used purchase accounting for the Rockwell acquisition and pooling for McDonnell Douglas. The goodwill on the balance sheet reflects the excess of acquisition value over book value for Rockwell and is being amortized over 30 years (which Boeing would not be able to do under current rules). With McDonnell Douglas, there is no recording of the premium paid on the acquisition among the assets, suggesting that the acquisition was structured to qualify for pooling, which would also not be allowed under current rules. 2. Customer financing and accounts receivable. Boeing often either provides financing to its customers to acquire its planes or acts as the lessor on the planes. Since these contracts tend to run over several years, the present value of the payments due in future years on the financing and the lease payments is shown as customer financing. The current portion of these payments is shown as accounts receivable. The Home Depot provides credit to its customers as well, but all these payments due are shown as accounts receivable, since they are all short-term. 3. Inventories. Boeing values inventories using the weighted average cost method, while the Home Depot uses the FIFO approach for valuing inventories. 4. Marketable securities. Boeing classifies its short-term investments as trading investments and records them at market value. The Home Depot has a mix of trading, available-for-sale, and held-to-maturity investments and therefore uses a mix of book and market value to value these investments. 5. Prepaid pension expense. Boeing records the excess of its pension fund assets over its expected pension fund liabilities as an asset on the balance sheet. Finally, the balance sheet for Boeing fails to report the value of a very significant asset, which is the effect of past research and development (R&D) expenses. Since accounting convention requires that these be expensed in the year that they occur and not be capitalized, the research asset does not show up in the balance sheet. Chapter 9 considers how to capitalize research and development expenses and the effects on balance sheets.
MEASURING FINANCING MIX The second set of questions that we would like to answer, and would l ike accounting statements to shed some light on, relate to the mix of debt and equity used by the firm, and the current values of each. The bulk of the information about these questions is provided on the liabilities side of the balance sheet and the footnotes to it.
Accounting Principles Underlying Liability and Equity Measurement 37
Just as with the measurement of asset value, the accounting categorization of liabilities and equity is governed by a set of fairly rigid principles. The first is a strict categorization of financing into either a liability or an equity based on the nature of the obligation. For an obligation to be recognized as a liability, it must meet three requirements: 1. The obligation must be expected to lead to a future cash outflow or the loss of a future cash inflow at some specified or determinable date. 2. The firm cannot avoid the obligation. 3. The transaction giving rise to the obligation has to have already happened. In keeping with the earlier principle of conservatism in estimating asset value, accountants recognize as liabilities only cash flow obligations that cannot be avoided. The second principle is that the values of both liabilities and equity in a firm are better estimated using historical costs with accounting adjustments, rather than with expected future cash flows or market value. The process by which accountants measure the value of liabilities and equities is inextricably linked to the way they value assets. Since assets are primarily valued at historical cost or at book value, both debt and equity also get measured primarily at book value. The next section examines the accounting measurement of both liabilities and equity.
Measuring the Value of Liabilities and Equities Accountants categorize liabilities into current liabilities, long-term debt, and long-term liabilities that are not debt or equity. Next, we will examine the way they measure each of these.
Current Liabilities Under current liabilities are categorized all obligations that the firm has coming due in the next year. These generally include:
Accounts payable, representing credit received from suppliers and other vendors to the firm. The value of accounts payable represents the amounts due to these creditors. For this item, book and market values should be similar. Short-term borrowing, representing short-term loans (due in less than a year) taken to finance the operations or current asset needs of the business. Here again, the value shown represents the amounts due on such loans, and the book and market values should be similar, unless the default risk of the firm has changed dramatically since it borrowed the money. Short-term portion of long-term borrowing, representing the portion of the long-term debt or bonds that is coming due in the next year. Here again, the value shown is the actual amount due on these loans, and market and book values should converge as the due date approaches. Other short-term liabilities, which is a catchall component for any other short-term liabilities that the firm might have, including wages due to its employees and taxes due to the government. Of all the items in the balance sheet, absent outright fraud, current liabilities should be the one for which the accounting estimates of book value and financial estimates of market value are closest.
Long-Term Debt Long-term debt for firms can take one of two forms. It can be a long-term loan from a bank or other financial institution, or it can be a long-term bond issued to financial markets, in which case the creditors are the investors in the bond. Accountants measure the value of long-term debt by looking at the present value of payments due on the loan or bond at the time of the borrowing. For bank loans, this will be equal to the nominal value of the loan. With bonds, however, there are three possibilities: When bonds are issued at par value, for instance, the value of the long-term debt is generally measured in terms of the nominal obligation created (i.e., principal due on the borrowing). When bonds are issued at a premium or a discount on par value, the bonds are recorded at the issue price, but the premium or discount is amortized over the life of the bond. As an extreme example, companies that issue zero coupon debt have to record the debt at the issue price, which will be significantly below the principal (face value) due at maturity. The difference between the issue price and the face value is amortized each period and is treated as a noncash interest expense that is tax deductible. In all these cases, the value of debt is unaffected by changes in interest rates during the life of the loan or bond. Note that as market interest rates rise or fall, the present value of the loan obligations should decrease or increase. This updated market value for debt is not shown on the balance sheet. If debt is retired prior to maturity, the difference between book value and the amount paid at retirement is treated as an extraordinary gain or loss in the income statement. Finally, companies that have long-term debt denominated in nondomestic currencies have to adjust the book value of debt for changes in exchange rates. Since exchange rate changes reflect underlying changes in interest rates, it 38
does imply that this debt is likely to be valued much nearer to market value than is debt in the domestic currency.
Other Long-Term Liabilities Firms often have long-term obligations that are not captured in the long-term debt item. These include obligations to lessors on assets that firms have leased, to employees in the form of pension fund and health care benefits yet to be paid, and to the government in the form of taxes deferred. In the past two decades accountants have increasingly moved toward quantifying these liabilities and showing them as long-term liabilities.
Leases Firms often choose to lease long-term assets rather than buy them. Lease payments create the same kind of obligation that interest payments on debt create, and they must be viewed in a similar light. If a firm is allowed to lease a significant portion of its assets and keep it off its financial statements, a perusal of the statements will give a very misleading view of the company's financial strength. Consequently, accounting rules have been devised to force firms to reveal the extent of their lease obligations on their books. There are two ways of accounting for leases. In an operating lease, the lessor (or owner) transfers only the right to use the property to the lessee. At the end of the lease period, the lessee returns the property to the lessor. Since the lessee does not assume the risk of ownership, the lease expense is treated as an operating expense in the income statement and the lease does not affect the balance sheet. In a capital lease, the lessee assumes some of the risks of ownership and enjoys some of the benefits. Consequently, the lease, when signed, is recognized both as an asset and as a liability (for the lease payments) on the balance sheet. The firm gets to claim depreciation each year on the asset and also deducts the interest expense component of the lease payment each year. In general, capital leases recognize expenses sooner than equivalent operating leases. Since firms prefer to keep leases off the books and sometimes to defer expenses, they have a strong incentive to report all leases as operating leases. Consequently the Financial Accounting Standards Board has ruled that a lease should be treated as a capital lease if it meets any one of the following four conditions: 1. The lease life exceeds 75 percent of the life of the asset. 2. There is a transfer of ownership to the lessee at the end of the lease term. 3. There is an option to purchase the asset at a bargain price at the end of the lease term. 4. The present value of the lease payments, discounted at an appropriate discount rate, exceeds 90 percent of the fair market value of the asset. The lessor uses the same criteria for determining whether the lease is a capital or operating lease and accounts for it accordingly. If it is a capital lease, the lessor records the present value of future cash flows as revenue and recognizes expenses. The lease receivable is also shown as an asset on the balance sheet, and the interest revenue is recognized over the term of the lease as paid. From a tax standpoint, the lessor can claim the tax benefits of the leased asset only if it is an operating lease, though the tax code uses slightly different criteria for determining whether the lease is an operating lease.4
Employee Benefits Employers can provide pension and health care benefits to their employees. In many cases, the obligations created by these benefits are extensive, and a failure by the firm to adequately fund these obligations needs to be revealed in financial statements.
Pension Plans In a pension plan, the firm agrees to provide certain benefits to its employees, either by specifying a defined contribution (wherein a fixed contribution is made to the plan each year by the employer, without any promises as to the benefits that will be delivered in the plan) or a defined benefit (wherein the employer promises to pay a certain benefit to the employee). In the latter case, the employer has to put sufficient money into the plan each period to meet the defined benefits. Under a defined contribution plan, the firm meets its obligation once it has made the prespecified contribution to the plan. Under a defined benefit plan, the firm's obligations are much more difficult to estimate, since they will be determined by a number of variables, including the benefits that employees are entitled to, the prior contributions made by the employer and the returns they have earned, and the rate of return that the employer expects to make on current contributions. As these variables change, the value of the pension fund assets can be greater than, less than, or equal to pension fund liabilities (which include the present value of promised benefits). A pension fund whose assets exceed its liabilities is an overfunded plan, whereas one whose assets are less than its liabilities is an 39
underfunded plan, and disclosures to that effect have to be included in financial statements, generally in the footnotes. When a pension fund is overfunded, the firm has several options. It can withdraw the excess assets from the fund, it can discontinue contributions to the plan, or it can continue to make contributions on the assumption that the overfunding is a transitory phenomenon that could well disappear by the next period. When a fund is underfunded, the firm has a liability, though accounting standards require that firms reveal only the excess of accumulated pension fund liability5 over pension fund assets on the balance sheet.
Health Care Ben efits A firm can provide health care benefits in either of two ways—by making a fixed contribution to a health care plan without promising specific benefits (analogous to a defined contribution plan) or by promising specific health benefits and setting aside the funds to provide these benefits (analogous to a defined benefit plan). The accounting for health care benefits is very similar to the accounting for pension obligations.
Deferred Taxes Firms often use different methods of accounting for tax and financial reporting purposes, leading to a question of how tax liabilities should be reported. Since accelerated depreciation and favorable inventory valuation methods for tax accounting purposes lead to a deferral of taxes, the taxes on the income reported in the financial statements will generally be much greater than the actual tax paid. The same principles of matching expenses to income that underlie accrual accounting suggest that the deferred income tax be recognized in the financial statements. Thus a company that pays taxes of $55,000 on its taxable income based on its tax accounting, and that would have paid taxes of $75,000 on the income reported in its financial statements, will be forced to recognize the difference ($20,000) as deferred taxes. Since the deferred taxes will be paid in later years, they will be recognized when paid. It is worth noting that companies that actually pay more in taxes than the taxes they report in the financial statements create an asset called a deferred tax asset. This reflects the fact that the firm's earnings in future periods will be greater as the firm is given credit for the deferred taxes. The question of whether the deferred tax liability is really a liability is an interesting one. On one hand, the firm does not owe the amount categorized as deferred taxes to any entity, and treating it as a liability makes the firm look more risky than it really is. On the other hand, the firm will eventually have to pay its deferred taxes, and treating the amount as a liability seems to be the conservative thing to do.
Preferred Stock When a company issues preferred stock, it generally creates an obligation to pay a fixed dividend on the stock. Accounting rules have conventionally not viewed preferred stock as debt because the failure to meet preferred dividends does not result in bankruptcy. At the same time, the fact the preferred dividends are cumulative makes them more onerous than common equity. Thus, preferred stock is a hybrid security, sharing some characteristics with equity and some with debt. Preferred stock is valued on the balance sheet at its original issue price, with any cumulated unpaid dividends added on. Convertible preferred stock is treated similarly, but it is treated as equity on conversion.
Equity The accounting measure of equity is a historical cost measure. The value of equity shown on the balance sheet reflects the original proceeds received by the firm when it issued the equity, augmented by any earnings made since (or reduced by losses, if any) and reduced by any dividends paid out during the period. While these three items go into what we can call the book value of equity, three other points need to be made about this estimate: 1. When companies buy back stock for short periods, with the intent of reissuing the stock or using it to cover option exercises, they are allowed to show the repurchased stock as treasury stock, which reduces the book value of equity. Firms are not allowed to keep treasury stock on the books for extended periods, and have to reduce their book value of equity by the value of repurchased stock in the case of stock buybacks. Since these buybacks occur at the current market price, they can result in significant reductions in the book value of equity. 2. Firms that have significant losses over extended periods or carry out massive stock buybacks can end up with negative book values of equity. 3. Relating back to the discussion of marketable securities, any unrealized gain or loss in marketable securities that are classified as available for sale is shown as an increase or a decrease in the book value of equity in the balance sheet.
40
As part of their financial statements, firms provide a summary of changes in shareholders' equity during the period, where all the changes that occurred to the accounting measure of equity value are summarized. As a final point on equity, accounting rules still seem to consider preferred stock, with its fixed dividend, as equity or near-equity, largely because of the fact that preferred dividends can be deferred or cumulated without the risk of default. To the extent that there can still be a loss of control in the firm (as opposed to bankruptcy), we have already argued that preferred stock shares almost as many characteristics with unsecured debt as it does with equity.
ILLUSTRATION 3.2: Measuring Liabilities and Equity in Boeing and the Home Depot in 1998 The following table summarizes the accounting estimates of liabilities and equity at Boeing and the Home Depot for the 1998 financial year in millions of dollars:
Boeing Home Depot Accounts payable and other liabilities
$10,733 $ 1,586
Accrued salaries and expenses
Advances in excess of costs
$ 1,251 $ 0
Taxes payable
$ 569
Short-term debt and current long-term debt $ 869
$ 1,010
$ 247 $ 14
Total current liabilities
$13,422 $ 2,857
Accrued health care benefits
$ 4,831 0
Other long-term liabilities
$ 210
Deferred income taxes
$ 83
Long-term debt
$ 6,103 $ 1,566
Minority interests
$9
$0
Shareholders' Equity Par value
$ 5,059 $ 37
Additional paid-in capital
$0
Retained earnings
$ 7,257 $ 5,812
Total shareholders' equity
$12,316 $ 8,740
Total liabilities
$36,672 $13,465
$ 2,891
The most significant difference between the companies is the accrued health care liability shown by Boeing, representing the present value of expected health care obligations promised to employees in excess of health care assets. The shareholders' equity for both firms represents the book value of equity and is significantly different from the market value of equity. The follwing table summarizes the difference at the end of 1998 (in millions of dollars):
Boeing
Home Depot
Book value of equity
$12,316
$ 8,740
Market value of equity $32,595
$85,668
One final point needs to be made about the Home Depot's liabilities. The Home Depot has substantial operating leases. Because these leases are treated as operating expenses, they do not show up in the balance sheet. Since they represent commitments to make payments in the future, we would argue that operating leases should be capitalized and treated as part of the liabilities of the firm. How best to do this is considered in Chapter 9.
MEASURING EARNINGS AND PROFITABILITY How profitable is a firm? What did it earn on the assets th at it invested in? These are fundamental questions we would like financial statements to answer. Accountants use the income statement to provide information about a firm's operating activities over a specific time period. The income statement is designed to measure the earnings from assets in place. This section examines the principles underlying earnings and return measurement in accounting, and the way they are put into practice.
Accounting Principles Underlying Measurement of Earnings and Profitability Two primary principles underlie the measurement of accounting earnings and profitability. The first is the principle of accrual accounting. In accrual accounting, the revenue from selling a good or service is recognized in the period in which the good is sold or the service is performed (in whole or substantially). A corresponding effort is made on the expense side to match expenses to revenues.6 This is in contrast to a cash-based system of accounting, where revenues are recognized when payment is received and expenses are recorded when paid. 41
The second principle is the categorization of expenses into operating, financing, and capital expenses. Operating expenses are expenses that, at least in theory, provide benefits only for the current period; the cost of labor and materials expended to create products that are sold in the current period is a good example. Financing expenses are expenses arising from the nonequity financing used to raise capital for the business; the most common example is interest expenses. Capital expenses are expenses that are expected to generate benefits over multiple periods; for instance, the cost of buying land and buildings is treated as a capital expense. Operating expenses are subtracted from revenues in the current period to arrive at a measure of operating earnings of the firm. Financing expenses are subtracted from operating earnings to estimate earnings to equity investors or net income. Capital expenses are written off over their useful lives (in terms of generating benefits) as depreciation or amortization.
Measuring Accounting Earnings and Profitability Since income can be generated from a number of different sources, generally accepted accounting principles (GAAP) require that income statements be classified into four sections—income from continuing operations, income from discontinued operations, extraordinary gains or losses, and adjustments for changes in accounting principles. Generally accepted accounting principles require the recognition of revenues when the service for which the firm is getting paid has been performed in full or substantially, and the firm has received in return either cash or a receivable that is both observable and measurable. Expenses linked directly to the production of revenues (like labor and materials) are recognized in the same period in which revenues are recognized. Any expenses that are not directly linked to the production of revenues are recognized in the period in which the firm consumes the services. Accounting has resolved one inconsistency that bedeviled it for years, with a change in the way it treats employee options. Unlike the old rules, these option grants were not treated as expenses when granted but only when exercised, the new rules require that employee options be valued and expensed, when granted (with allowances for amortization over periods). Since employee options are part of compensation, which is an operating expense, the new rules make more sense. While accrual accounting is straightforward in firms that produce goods and sell them, there are special cases where accrual accounting can be complicated by the nature of the product or service being offered. For instance, firms that enter into long-term contracts with their customers are allowed to recognize revenue on the basis of the percentage of the contract that is completed. As the revenue is recognized on a percentage-of-completion basis, a corresponding proportion of the expense is also recognized. When there is considerable uncertainty about the capacity of the buyer of a good or service to pay for it, the firm providing the good or service may recognize the income only when it collects portions of the selling price under the installment method. Reverting back to the discussion of the difference between capital and operating expenses, operating expenses should reflect only those expenses that create revenues in the current period. In practice, however, a number of expenses are classified as operating expenses that do not seem to meet this test. The first is depreciation and amortization. While the notion that capital expenditures should be written off over multiple periods is reasonable, the accounting depreciation that is computed on the original historical cost often bears little resemblance to the actual economic depreciation. The second expense is research and development expenses, which accounting standards classify as operating expenses, but which clearly provide benefits over multiple periods. The rationale used for this classification is that the benefits cannot be counted on or easily quantified. Much of financial analysis is built around the expected future earnings of a firm, and many of these forecasts start with the current earnings. It is therefore important to know how much of these earnings comes from the ongoing operations of the firm and how much can be attributed to unusual or extraordinary events that are unlikely to recur on a regular basis. From that standpoint, it is useful that firms categorize expenses into operating and nonrecurring expenses, since it is the earnings prior to extraordinary items that should be used in forecasting. Nonrecurring items include:
Unusual or infrequent items, such as gains or losses from the divestiture of an asset or division, and writeoffs or restructuring costs. Companies sometimes include such items as part of operating expenses. As an example, Boeing in 1997 took a write-off of $1,400 million to adjust the value of assets it acquired in its acquisition of McDonnell Douglas, and it showed this as part of operating expenses. Extraordinary items, which are defined as events that are unusual in nature, infrequent in occurrence, and material in impact. Examples include the accounting gain associated with refinancing high-coupon debt with lower-coupon debt, and gains or losses from marketable securities that are held by the firm. Losses associated with discontinued operations, which measure both the loss from the phaseout period and any estimated loss on sale of the operations. To qualify, however, the operations have to be separable from the firm. Gains or losses associated with accounting changes, which measure earnings changes created by both accounting changes made voluntarily by the firm (such as a change in inventory valuation) and accounting changes mandated by new accounting standards. 42
ILLUSTRATION 3.3: Measures of Earnings—Boeing and the Home Depot in 1998 The following table summarizes the income statements of Boeing and the Home Depot for the 1998 financial year:
Boeing (in $ millions) Home Depot (in $ millons) Sales and other operating revenues
$56,154
$30,219
– Operating costs and expenses
$51,022
$27,185
– Depreciation
$ 1,517
$ 373
– Research and development expenses
$ 1,895
$0
Operating income
$ 1,720
$ 2,661
+ Other income (includes interest income) $ 130
$ 30
– Interest expenses
$ 453
$ 37
Earnings before taxes
$ 1,397
$ 2,654
– Income taxes
$ 277
$ 1,040
Net earnings (Loss)
$ 1,120
$ 1,614
Boeing's operating income is reduced by the research and development expense, which is treated as an operating expense by accountants. The Home Depot's operating expenses include operating leases. As noted earlier, the treatment of both these items skews earnings, and how best to adjust earnings when such expenses exist is considered in Chapter 9.
Measures of Profitability While the income statement allows us to estimate how profitable a firm is in absolute terms, it is just as important that we gauge the profitability of the firm in terms of percentage returns. Two basic ratios measure profitability. One examines the profitability relative to the capital employed to get a rate of return on investment. This can be done either from the viewpoint of just the equity investors or by looking at the entire firm. Another examines profitability relative to sales, by estimating a profit margin.
Return on Assets and Return on Capital The return on assets (ROA) of a firm measures its operating efficiency in generating profits from its assets, prior to the effects of financing. Return on assets = Earnings before interest and taxes(1 – Tax rate)/Total assets Earnings before interest and taxes (EBIT) is the accounting measure of operating income from the income statement, and total assets refers to the assets as measured using accounting rules—that is, using book value (BV) for most assets. Alternatively, return on assets can be written as: Return on assets = [Net income + Interest expenses(1 – Tax rate)]/Total assets By separating the financing effects from the operating effects, the return on assets provides a cleaner measure of the true return on these assets. By dividing by total assets, the return on assets does understate the profitability of firms that have substantial current assets. ROA can also be computed on a pretax basis with no loss of generality, by using the earnings before interest and taxes and not adjusting for taxes: Pretax ROA = Earnings before interest and taxes/Total assets This measure is useful if the firm or division is being evaluated for purchase by an acquirer with a different tax rate. A more useful measure of return relates the operating income to the capital invested in the firm, where capital is defined as the sum of the book value of debt and equity, net of cash. This is the return on invested capital (ROC or ROIC), and provides not only a truer measure of return but one that can be compared to the cost of capital, to measure the quality of a firm's investments.
The denominator is generally termed invested capital and measures the book value of operating assets. For both measures, the book value can be measured at the beginning of the period or as an average of beginning and ending values.
ILLUSTRATION 3.4: Estimating Return on Capital—Boeing and the Home Depot in 1998 43
The following table summarizes the after-tax return on assets and return on capital estimates for Boeing and the Home Depot, using both average and beginning measures of capital in 1998:
Boeing
Home Depot
(in $millions) (in $millions) After-tax operating income
$ 1,118
$ 1,730
Book value of capital—beginning
$19,807
$ 8,525
Book value of capital—ending
$19,288
$10,320
Book value of capital—average
$19,548
$ 9,423
Return on capital (based on average)
5.72%
18.36%
Return on capital (based on beginning) 5.64%
20.29%
Boeing had a terrible year, in 1998, in terms of after-tax returns. The Home Depot had a much better year in terms of those same returns.
Decomposing Return on Capital The return on capital of a firm can be written as a function of the operating profit margin it has on its sales, and its capital turnover ratio.
Thus, a firm can arrive at a high ROC by either increasing its profit margin or utilizing its capital more efficiently to increase sales. There are likely to be competitive constraints and technological constraints on both variables, but a firm still has some freedom within these constraints to choose the mix of profit margin and capital turnover that maximizes its ROC. The return on capital varies widely across firms in different businesses, largely as a consequence of differences in profit margins and capital turnover ratios.
mgnroc.xls: This is a dataset on the Web that summarizes the operating margins, turnover ratios, and returns on capital of firms in the United States, classified by industry.
Return on Equity While the return on capital measures the profitability of the overall firm, the return on equity (ROE) examines profitability from the perspective of the equity investor, by relating the equity investor's profits (net profit after taxes and interest expenses) to the book value of the equity investment.
Since preferred stockholders have a different type of claim on the firm than do common stockholders, the net income should be estimated after preferred dividends, and the book value should be that of only common equity.
Determinants of Noncash ROE Since the ROE is based on earnings after interest payments, it is affected by the financing mix the firm uses to fund its projects. In general, a firm that borrows money to finance projects and that earns a ROC on those projects that exceeds the after-tax interest rate it pays on its debt will be able to increase its ROE by borrowing. The return on equity, not including cash, can be written as follows:7
where ROC = EBIT(1 – t)/(BV of debt + BV of equity – Cash) D/E = BV of debt/BV of equity i = Interest expense on debt/BV of debt 44
t = Tax rate on ordinary income The second term captures the benefit of financial leverage.
ILLUSTRATION 3.5: Return on Equity Computations: Boeing and the Home Depot in 1998 The following table summarizes the return on equity for Boeing and the Home Depot in 1998:
Boeing
Home Depot
Return Ratios
(in $millions) (in $millions)
Net income
$ 1,120
$1,614
Book value of equity—beginning
$12,953
$7,214
Book value of equity—ending
$12,316
$8,740
Book value of equity—average
$12,635
$7,977
Return on equity (based on average)
8.86%
20.23%
Return on equity (based on beginning) 8.65%
22.37%
The results again indicate that Boeing had a substandard year in 1998, while the Home Depot reported healthier returns on equity. The returns on equity can also be estimated by decomposing into the components just specified (using the adjusted beginning-of-the-year numbers):
Boeing
Home Depot
(in $millions) (in $millions) After-tax return on capital
5.82%
16.37%
Debt-equity ratio
35.18%
48.37%
Book interest rate(1 – Tax rate) 4.22%
4.06%
Return on equity
22.33%
6.38%
Note that a tax rate of 35% is used on both the return on capital and the book interest rate. This approach results in a return on equity that is different from the one estimated using the net income and the book value of equity.
rocroe.xls: This is a dataset on the Web that summarizes the return on capital, debt equity ratios, book interest rates, and returns on equity of firms in the United States, classified by industry.
MEASURING RISK How risky are the investments the firm has made over time? How much risk do equity investors in a firm face? These are tw o more questions that we would like to find the answers to in the course of an investment analysis. Accounting statements do not really claim to measure or quantify risk in a systematic way, other than to provide footnotes and disclosures where there might be risk embedded in the firm. This section examines some of the ways in which accountants try to assess risk.
Accounting Principles Underlying Risk Measurement To the extent that accounting statements and ratios do attempt to measure risk, there seem to be two common themes. The first is that the risk being measured is the risk of default—that is, the risk that a fixed obligation, such as interest or principal due on outstanding debt, will not be met. The broader equity notion of risk, which measures the variance of actual returns around expected returns, does not seem to receive much attention. Thus, an allequity-financed firm with positive earnings and few or no fixed obligations will generally emerge as a low-risk firm from an accounting standpoint, in spite of the fact that its earnings are unpredictable. The second theme is that accounting risk measures generally take a static view of risk, by looking at the capacity of a firm at a point in time to meet its obligations. For instance, when ratios are used to assess a firm's risk, the ratios are almost always based on one period's income statement and balance sheet.
Accounting Measures of Risk Accounting measures of risk can be broadly categorized into two groups. The first is disclosures about potential obligations or losses in values that show up as footnotes on balance sheets, which are designed to alert potential or 45
current investors to the possibility of significant losses. The second measure is ratios that are designed to measure both liquidity and default risk.
Disclosures in Financial Statements In recent years, the disclosures that firms have to make about future obligations have proliferated. Consider, for instance, the case of contingent liabilities. These refer to potential liabilities that will be incurred under certain contingencies, as is the case, for instance, when a firm is the defendant in a lawsuit. The general rule that has been followed is to ignore contingent liabilities that hedge against risk, since the obligations on the contingent claim will be offset by benefits elsewhere.8 In recent periods, however, significant losses borne by firms from supposedly hedged derivatives positions (such as options and futures) have led to FASB requirements that these derivatives be discl osed as part of a financial statement. In fact, pension fund and health care obligations have moved from mere footnotes to actual liabilities for firms.
Financial Ratios Financial statements have long been used as the basis for estimating financial ratios that measure profitability, risk, and leverage. Earlier, the section on earnings looked at two of the profitability ratios—return on equity and return on capital. This section looks at some of the financial ratios that are often used to measure the financial risk in a firm.
Short-Term Liquidity Risk Short-term liquidity risk arises primarily from the need to finance current operations. To the extent that the firm has to make payments to its suppliers before it gets paid for the goods and services it provides, there is a cash shortfall that has to be met, usually through short-term borrowing. Though this financing of working capital needs is done routinely in most firms, financial ratios have been devised to keep track of the extent of the firm's exposure to the risk that it will not be able to meet its short-term obligations. The two ratios most frequently used to measure shortterm liquidity risk are the current ratio and the quick ratio.
Current Ratios The current ratio is the ratio of the firm's current assets (cash, inventory, accounts receivable) to its current liabilities (obligations coming due within the next period).
A current ratio below 1, for instance, would indicate that the firm has more obligations coming due in the next year than assets it can expect to turn into cash. That would be an indication of liquidity risk. While traditional analysis suggests that firms maintain a current ratio of 2 or greater, there is a trade-off here between minimizing liquidity risk and tying up more and more cash in net working capital (Net working capital = Current assets – Current liabilities). In fact, it can be reasonably argued that a very high current ratio is indicative of an unhealthy firm that is having problems reducing its inventory. In recent years firms have worked at reducing their current ratios and managing their net working capital better. Reliance on current ratios has to be tempered by a few concerns. First, the ratio can be easily manipulated by firms around the time of financial reporting dates to give the illusion of safety; second, current assets and current liabilities can change by an equal amount, but the effect on the current ratio will depend on its level before the change.9
Quick or Acid Test Ratios The quick or acid test ratio is a variant of the current ratio. It distinguishes current assets that can be converted quickly into cash (cash, marketable securities) from those that cannot (inventory, accounts receivable).
The exclusion of accounts receivable and inventory is not a hard-and-fast rule. If there is evidence that either can be converted into cash quickly, it can, in fact, be included as part of the quick ratio.
Turnover Ratios Turnover ratios measure the efficiency of working capital management by looking at the relationship of accounts receivable and inventory to sales and to the cost of goods sold: 46
These statistics can be interpreted as measuring the speed with which the firm turns accounts receivable into cash or inventory into sales. These ratios are often expressed in terms of the number of days outstanding:
A similar pair of statistics can be computed for accounts payable, relative to puchases:
Since accounts receivable and inventory are assets, and accounts payable is a liability, these three statistics (standardized in terms of days outstanding) can be combined to get an estimate of how much financing the firm needs to raise to fund working capital needs.
The greater the financing period for a firm, the greater is its short-term liquidity risk.
wcdata.xls: This is a dataset on the Web that summarizes working capital ratios for firms in the United States, classified by industry.
finratio.xls: This spreadsheet allows you to compute the working capital ratios for a firm, based upon financial statement data.
Long-Term Solvency and Default Risk Measures of long-term solvency attempt to examine a firm's capacity to meet interest and principal payments in the long term. Clearly, the profitability ratios discussed earlier in the section are a critical component of this analysis. The ratios specifically designed to measure long-term solvency try to relate profitability to the level of debt payments in order to identify the degree of comfort with which the firm can meet these payments.
Interest Coverage Ratios The interest coverage ratio measures the capacity of the firm to meet interest payments from predebt, pretax earnings.
The higher the interest coverage ratio, the more secure is the firm's capacity to make interest payments from earnings. This argument, however, has to be tempered by the recognition that the amount of earnings before interest and taxes is volatile and can drop significantly if the economy enters a recession. Consequently, two firms can have the same interest coverage ratio but be viewed very differently in terms of risk. The denominator in the interest coverage ratio can be easily extended to cover other fixed obligations such as lease payments. If this is done, the ratio is called a fixed charges coverage ratio:
Finally, this ratio, while stated in terms of earnings, can be restated in terms of cash flows by using earnings before interest, taxes, depreciation, and amortization (EBITDA) in the numerator and cash fixed charges in the denominator. 47
Both interest coverage and fixed charges coverage ratios are open to the criticism that they do not consider capital expenditures, a cash flow that may be discretionary in the very short term, but not in the long term if the firm wants to maintain growth. One way of capturing the extent of this cash flow, relative to operating cash flows, is to compute a ratio of the two:
While there are a number of different definitions of cash flows from operations, the most reasonable way of defining it is to measure the cash flows from continuing operations, before interest but after taxes and after meeting working capital needs.
covratio.xls: This is a dataset on the Web that summarizes the interest coverage and fixed charges coverage ratios for firms in the United States, classified by industry.
ILLUSTRATION 3.6: Interest and Fixed Charges Coverage Ratios: Boeing and the Home Depot in 1998 The following table summarizes interest and fixed charges coverage ratios for Boeing and the Home Depot in 1998:
Boeing Home Depot EBIT
$1,720 $2,661
Interest expense
$ 453
$ 37
Interest coverage ratio
3.80
71.92
EBIT
$1,720 $2,661
Operating lease expenses
$ 215
$ 290
Interest expenses
$ 453
$ 37
Fixed charges coverage ratio
2.90
9.02
EBITDA
$3,341 $3,034
Cash fixed charges
$ 668
Cash fixed charges coverage ratio 5.00
$ 327 9.28
Cash flows from operations
$2,161 $1,662
Capital expenditures
$1,584 $2,059
Cash flows/Capital expenditures
1.36
0.81
Boeing, based on its operating income in 1998, looks riskier than the Home Depot on both the interest coverage ratio basis and fixed charges coverage ratio basis. On a cash flow basis, however, Boeing does look much better. In fact, when capital expenditures are considered, the Home Depot has a lower ratio. For Boeing, the other consideration is the fact that operating income in 1998 was depressed relative to income in earlier years, and this does have an impact on the ratios across the board. It might make more sense when computing these ratios to look at the average operating income over time.
finratio.xls: This spreadsheet allows you to compute the interest coverage and fixed charges coverage ratios for a firm based on financial statement data.
Debt Ratios Interest coverage ratios measure the capacity of the firm to meet interest payments, but do not examine whether it can pay back the principal on outstanding debt. Debt ratios attempt to do this, by relating debt to total capital or to equity:
48
The first ratio measures debt as a proportion of the total capital of the firm and cannot exceed 100 percent. The second measures debt as a proportion of the book value of equity in the firm and can be easily derived from the first, since:
While these ratios presume that capital is raised from only debt and equity, they can be easily adapted to include other sources of financing, such as preferred stock. Although preferred stock is sometimes combined with common stock under the equity label, it is better to keep the two sources of financing separate and to compute the ratio of preferred stock to capital (which will include debt, equity, and preferred stock). There are two close variants of debt ratios. In the first, only long-term debt is used rather than total debt, with the rationale that short-term debt is transitory and will not affect the long-term solvency of the firm.
Given the ease with which some firms can roll over short-term debt and the willingness of many firms to use shortterm financing to fund long-term projects, these variants can provide a misleading picture of the firm's financial leverage risk. The second variant of debt ratios uses market value (MV) instead of book value, primarily to reflect the fact that some firms have a significantly greater capacity to borrow than their book values indicate.
Many analysts disavow the use of market value in their calculations, contending that market values, in addition to being difficult to get for debt, are volatile and hence unreliable. These contentions are open to debate. It is true that the market value of debt is difficult to get for firms that do not have publicly traded bonds, but the market value of equity not only is easy to obtain, but it also is constantly updated to reflect marketwide and firm-specific changes. Furthermore, using the book value of debt as a proxy for market value in those cases where bonds are not traded does not significantly shift most market value-based debt ratios.10
ILLUSTRATION 3.7: Book Value Debt Ratios and Variants—Boeing and the Home Depot The following table summarizes different estimates of the debt ratio for Boeing and the Home Depot, in 2008, using book values of debt and equity for both firms:
Boeing
Home Depot
(in $millions) (in $millions) Long-term debt
$ 6,103
$1,566
Short-term debt
$ 869
$ 14
Book value of equity
$12,316
$8,740
Long-term debt/Equity
49.55%
17.92%
Long-term debt/(Long-term debt + Equity) 33.13%
15.20%
Debt/Equity
56.61%
18.08%
Debt/(Debt + Equity)
36.15%
15.31%
In 2008, Boeing has a much higher book value debt ratio, considering either long-term or total debt, than the Home Depot.
dbtfund.xls: This is a dataset on the Web that summarizes the book value debt ratios and market value debt ratios for firms in the United States, classified by industry.
OTHER ISSUES IN ANALYZING FINANCIAL STATEMENTS There are significant differences in accounting standards and practices across countries and thes e differences may color comparisons across companies.
Differences in Accounting Standards and Practices 49
Differences in accounting standards across countries affect the measurement of earnings. These differences, however, are not so great as they are made out to be by some analysts, and they cannot explain away radical departures from fundamental principles of valuation. Choi and Levich, in a 1990 survey of accounting standards across developed markets, note that most countries subscribe to basic accounting notions of consistency, realization, and historical cost principles in preparing accounting statements. As countries increasingly move toward international financial reporting standards (IFRS), it is worth noting that IFRS and U.S. GAAP are more similar than dissimilar on many issues. It is true that there are areas of differences that still remain, and we note some of them in Table 3.1. Table 3.1 Key Differences between IFRS and GAAP
Most of these differences can be accounted and adjusted for when comparisons are made between companies in the United States and companies in other financial markets. Statistics such as price-earnings ratios, which use stated and unadjusted earnings, can be misleading when accounting standards vary widely across the companies being compared.
CONCLUSION Financial statements remain the primary source of information for most investors and analysts. There are differences, however, betwee n how accounting and financial analysts approach answering a number of key questions about the firm. The first question relates to the nature and the value of the assets owned by a firm. Assets can be categorized into investments already made (assets in place) and investments yet to be made (growth assets); accounting statements provide a substantial amount of historical information about the former and very little about the latter. The focus on the original price of assets in place (book value) in accounting statements can lead to significant differences between the stated value of these assets and their market value. With growth assets, accounting rules result in low or no values for assets generated by internal research. The second issue is the measurement of profitability. The two principles that govern how profits are measured are accrual accounting—in which revenues and expenses are shown in the period in which transactions occur rather than when the cash is received or paid—and the categorization of expenses into operating, financing, and capital expenses. While operating and financing expenses are shown in income statements, capital expenditures are spread over several time periods and take the form of depreciation and amortization. Accounting standards miscategorize operating leases and research and development expenses as operating expenses (when the former should be categorized as financing expenses and the latter as capital expenses). Financial statements also deal with short-term liquidity risk and long-term default risk. While the emphasis in accounting statements is on examining the risk that firms may be unable to make payments that they have committed to make, there is very little focus on risk to equity investors. 50
QUESTIONS AND SHORT PROBLEMS In the problems following, use an equity risk premium of 5.5 percent if none is specified. Coca-Cola's balance sheet for December 1998 is summarized (in millions of dollars) for problems 1 through 9:
1. Consider the assets on Coca-Cola's balance sheet and answer the following questions: a. Which assets are likely to be assessed closest to market value? Explain. b. Coca-Cola has net fixed assets of $3,669 million. Can you estimate how much Coca-Cola paid for these assets? Is there any way to know the age of these assets? c. Coca-Cola seems to have far more invested in current assets than in fixed assets. Is this significant? Explain. d. In the early 1980s, Coca-Cola sold off its bottling operations, and the bottlers became independent companies. How would this action have impacted the assets on Coca-Cola's balance sheet? (The manufacturing plants are most likely to be part of the bottling operations.) 2. Examine the liabilities on Coca-Cola's balance sheet. a. How much interest-bearing debt does Coca-Cola have outstanding? (You can assume that other short-term liabilities represent sundry payables, and other long-term liabilities represent health care and pension obligations.) b. How much did Coca-Cola obtain in equity capital when it issued stock originally to the financial markets? c. Is there any significance to the fact that the retained earnings amount is much larger than the original paid-in capital? d. The market value of Coca-Cola's equity is $140 billion. What is the book value of equity in Coca-Cola? Why is there such a large difference between the market value of equity and the book value of equity? 3. Coca-Cola's most valuable asset is its brand name. Where in the balance sheet do you see its value? Is there any way to adjust the balance sheet to reflect the value of this asset? 4. Assume that you have been asked to analyze Coca-Cola's working capital management. a. Estimate the net working capital and noncash working capital for Coca-Cola. b. Estimate the firm's current ratio. c. Estimate the firm's quick ratio. d. Would you draw any conclusions about the riskiness of Coca-Cola as a firm by looking at these numbers? Why or why not? Coca-Cola's income statements for 1997 and 1998 are summarized (in millions of dollars) for problems 5 through 9: 1997
1998
Net revenues
$18,868 $18,813
Cost of goods sold
6,015
5,562
Selling, general, and administrative expenses 7,852
8,284
Earnings before interest and taxes
5,001
4,967
Interest expenses
258
277
Nonoperating gains
1,312
508
51
Income tax expenses
1,926
1,665
Net income
4,129
3,533
Dividends
1,387
1,480
The following questions relate to Coca-Cola's income statements. 5. How much operating income did Coca-Cola earn, before taxes, in 1998? How does this compare to how much Coca-Cola earned in 1997? What are the reasons for the difference? 6. The biggest expense for Coca-Cola is advertising, which is part of the selling, generals and administrative (G&A) expenses. A large portion of these expenses is designed to build up Coca-Cola's brand name. Should advertising expenses be treated as operating expenses or are they really capital expenses? If they are to be treated as capital expenses, how would you capitalize them? (Use the capitalization of R&D as a guide.) 7. What effective tax rate did Coca-Cola have in 1998? How does it compare with what the company paid in 1997 as an effective tax rate? What might account for the difference? 8. You have been asked to assess the profitability of Coca-Cola as a firm. To that end, estimate the pretax operating and net margins in 1997 and 1998 for the firm. Are there any conclusions you would draw from the comparisons across the two years? 9. The book value of equity at Coca-Cola in 1997 was $7,274 million. The book value of interest-bearing debt was $3,875 million. Estimate: a. The return on equity (beginning of the year) in 1998. b. The pretax return on capital (beginning of the year) in 1998. c. The after-tax return on capital (beginning of the year) in 1998, using the effective tax rate in 1998. 10. SeeSaw Toys reported that it had a book value of equity of $1.5 billion at the end of 1998 and 100 million shares outstanding. During 1999, it bought back 10 million shares at a market price of $40 per share. The firm also reported a net income of $150 million for 1999, and paid dividends of $50 million. Estimate: a. The book value of equity at the end of 1999. b. The return on equity, using beginning book value of equity. c. The return on equity, using the average book value of equity. Depreciation is treated as an accounting expense. Hence, the use of straight-line depreciation (which is lower than accelerated depreciation in the first few years after an asset is acquire d) will result in lower expenses and higher income. 1
Firms have evaded the requirements of consolidation by keeping their share of ownership in other firms below 50 percent. 2
Once an acquisition is complete, the difference between market value and book value for the target firm does not automatically become goodwill. Existing assets can be reappraised first to f air value and the difference becomes goodwill. 3
The requirements for an operating lease in the tax code are: (1) The property can be used by someone other than the lessee at the end of the lease term, (2) the lessee cannot buy the asset using a bargain purchase option, (3) the lessor has at least 20 percent of its capital at risk, (4) the lessor has a positive cash flow from the lease independent of tax benefits, and (5) the lessee does not have an investment in the lease. 4
The accumulated pension fund liability does not take into account the projected benefit obligation, where actuarial estimates of future benefits are made. Consequently, it is much smaller t han the total pension liabilities. 5
If a cost (such as an administrative cost) cannot easily be linked with particular revenues, it is usually recognized as an expense in the period in which it is consumed. 6
7
52
8 This assumes that the hedge is set up competently. It is entirely possible that a hedge, if sloppily set up, can end up costing the firm money.
If the current assets and current liabilities increase by an equal amount, the current ratio will go down if it was greater than 1 before the increase, and go up if it was less than 1. 9
Deviations in the market value of equity from book value are likely to be much larger than deviations for debt, and are likely to dominate in most debt ratio calculations. 10
53
CHAPTER 4 The Basics of Risk When valuing assets and firms, we need to use discount rates that reflect the riskiness of the cash flows. In particular, the cost of debt has to incorporate a default spread for the default risk in the debt, and the cost of equity has to include a risk premium for equity risk. But how do we measure default and equity risk? More importantly, how do we come up with the default and equity risk premiums? This chapter lays the foundations for analyzing risk in valuation. It presents alternative models for measuring risk and converting these risk measures into acceptable hurdle rates. It begins with a discussion of equity risk and presents the analysis in three steps. In the first step, risk is defined in statistical terms to be the variance in actual returns around an expected return. The greater this variance, the more risky an investment is perceived to be. The next step, the central one, is to decompose this risk into risk that can be diversified away by investors and risk that cannot. The third step looks at how different risk and return models in finance attempt to measure this nondiversifiable risk. It compares the most widely used model, the capital asset pricing model (CAPM), with other models and explains how and why they diverge in their measures of risk and the implications for the equity risk premium. The final part of this chapter considers default risk and how it is measured by ratings agencies. By the end of the chapter, we should have a way of estimating the equity risk and default risk for any firm.
WHAT IS RISK? Risk, for most of us, refers to the likelihood that in life’s games of chance we will receive an outcome that we will not like. For in stance, the risk of driving a car too fast is getting a speeding ticket or, worse still, getting into an accident. Merriam-Webster’s Collegiate Dictionary, in fact, defines the verb to risk as “to expose to hazard or danger.” Thus risk is perceived almost entirely in negative terms. In finance, our definition of risk is both different and broader. Risk, as we see it, refers to the likelihood that we will receive a return on an investment that is different from the return we expect to make. Thus, risk includes not only the bad outcomes (returns that are lower than expected), but also good outcomes (returns that are higher than expected). In fact, we can refer to the former as downside risk and the latter as upside risk, but we consider both when measuring risk. In fact, the spirit of our definition of risk in finance is captured best by the Chinese symbols for risk:
Loosely defined, the first symbol is the symbol for “danger,” while the second is the symbol for “opportunity,” making risk a mix of danger and opportunity. It illustrates very clearly the trade-off that every investor and business has to make—between the higher rewards that come with the opportunity and the higher risk that has to be borne as a consequence of the danger. Much of this chapter can be viewed as an attempt to come up with a model that best measures the danger in any investment, and then attempts to convert this into the opportunity that we would need to compensate for the danger. In finance terms, we term the danger to be “risk” and the opportunity to be “expected return.” What makes the measurement of risk and expected return so challenging is that it can vary depending on whose perspective we adopt. When analyzing the risk of a firm, for instance, we can measure it from the viewpoint of the firm’s managers. Alternatively, we can argue that the firm’s equity is owned by its stockholders, and that it is their perspective on risk that should matter. A firm’s stockholders, many of whom hold the stock as one investment in a larger portfolio, might perceive the risk in the firm very differently from the firm’s managers, who might have the bulk of their capital, human and financial, invested in the firm. We argue that risk in an investment has to be perceived through the eyes of investors in the firm. Since firms often have thousands of investors, often with very different perspectives, it can be asserted that risk has to be measured from the perspective of not just any investor in the stock, but of the marginal investor, defined to be the investor most likely to be trading on the stock at any given point in time. The objective in valuation is to measure the value of an asset to those who will be pricing it. If we want to stay true to this objective, we have to consider the viewpoint of those who set the stock prices, and they are the marginal investors.
EQUITY RISK AND EXPECTED RETURN 54
To demonstrate how risk is viewed in finance, risk analysis is presented here in three steps: first, defining risk in terms of the distribution of act ual returns around an expected return; second, differentiating between risk that is specific to one or a few investments and risk that affects a much wider cross section of investments (in a market where the marginal investor is well diversified, it is only the latter risk, called market risk, that will be rewarded); and third, alternative models for measuring this market risk and the expected returns that go with it.
Defining Risk Investors who buy an asset expect to earn returns over the time horizon that they hold the asset. Their actual returns over this holding period may be very different from the expected returns, and it is this difference between actual and expected returns that is a source of risk. For example, assume that you are an investor with a one-year time horizon buying a one-year Treasury bill (or any other default-free one-year bond) with a 5 percent expected return. At the end of the one-year holding period, the actual return on this investment will be 5 percent, which is equal to the expected return. The return distribution for this investment is shown in Figure 4.1. This is a riskless investment. Figure 4.1 Probability Distribution of Returns on a Risk-Free Investment
To provide a contrast to the riskless investment, consider an investor who buys stock in a firm, say Boeing. This investor, having done her research, may conclude that she can make an expected return of 30 percent on Boeing over her one-year holding period. The actual return over this period will almost certainly not be equal to 30 percent; it might be much greater or much lower. The distribution of returns on this investment is illustrated in Figure 4.2. Figure 4.2 Return Distribution for Risky Investment
In addition to the expected return, an investor now has to consider the following. First, note that the actual returns, in this case, are different from the expected return. The spread of the actual returns around the expected return is measured by the variance or standard deviation of the distribution; the greater the deviation of the actual returns from the expected return, the greater the variance. Second, the bias toward positive or negative returns is represented by the skewness of the distribution. The distribution in Figure 4.2 is positively skewed, since there is a higher probability of large positive returns than large negative returns. Third, the shape of the tails of the 55
distribution is measured by the kurtosis of the distribution; fatter tails lead to higher kurtosis. In investment terms, this represents the tendency of the price of this investment to jump (up or down from current levels) in either direction. In the special case where the distribution of returns is normal, investors do not have to worry about skewness and kurtosis, since there is no skewness (normal distributions are symmetric) and a normal distribution is defined to have a kurtosis of zero. Figure 4.3 illustrates the return distributions on two investments with symmetric returns. Figure 4.3 Return Distribution Com parisons
When return distributions are normal, the characteristics of any investment can be measured with two variables— the expected return, which represents the opportunity in the investment, and the standard deviation or variance, which represents the danger. In this scenario, a rational investor, faced with a choice between two investments with the same standard deviation but different expected returns, will always pick the one with the higher expected return. In the more general case, where distributions are neither symmetric nor normal, it is still conceivable that investors will choose between investments on the basis of only the expected return and the variance, if they possess utility functions that allow them to do so.1 It is far more likely, however, that they prefer positive skewed distributions to negatively skewed ones, and distributions with a lower likelihood of jumps (lower kurtosis) over those with a higher likelihood of jumps (higher kurtosis). In thi s world, investors will trade off the good (higher expected returns and more positive skewness) against the bad (higher variance and kurtosis) in making investments. In closing, it should be noted that the expected returns and variances that we run into in practice are almost always estimated using past returns rather than future returns. The assumption made when using historical variances is that past return distributions are good indicators of future return distributions. When this assumption is violated, as is the case when the asset’s characteristics have changed significantly over time, the historical estimates may not be good measures of risk.
optvar.xls: This is a dataset on the Web that summarizes standard deviations and variances of stocks in various sectors in the United States.
Diversifiable and Nondiversifiable Risk Although there are many reasons why actual returns may differ from expected returns, we can group the reasons into two categories: firm-specific and marketwide. The risks that arise from firm-specific actions affect one or a few investments, while the risks arising from marketwide reasons affect many or all investments. This distinction is critical to the way we assess risk in finance.
Components of Risk When an investor buys stock or takes an equity position in a firm, he or she is exposed to many risks. Some risk may affect only one or a few firms, and this risk is categorized as firm-specific risk. Within this category, we would consider a wide range of risks, starting with the risk that a firm may have misjudged the demand for a product from its customers; we call this project risk. For instance, consider Boeing’s investment in a Super Jumbo jet. This 56
investment is based on the assumption that airlines want a larger airplane and are willing to pay a high price for it. If Boeing has misjudged this demand, it will clearly have an impact on Boeing’s earnings and value, but it should not have a significant effect on other firms in the market. The risk could also arise from competitors proving to be stronger or weaker than anticipated, called competitive risk. For instance, assume that Boeing and Airbus are competing for an order from Qantas, the Australian airline. The possibility that Airbus may win the bid is a potential source of risk to Boeing and perhaps some of its suppliers, but again, few other firms will be affected by it. Similarly, Disney recently launched magazines aimed at teenage girls, hoping to capitalize on the success of its TV shows. Whether it succeeds is clearly important to Disney and its competitors, but it is unlikely to have an impact on the rest of the market. In fact, risk measures can be extended to include risks that may affect an entire sector but are restricted to that sector; we call this sector risk. For instance, a cut in the defense budget in the United States will adversely affect all firms in the defense business, including Boeing, but there should be no significant impact on other sectors. What is common across the three risks described—project, competitive, and sector risk—is that they affect only a small subset of firms. There is another group of risks that is much more pervasive and affects many if not all investments. For instance, when interest rates increase, all investments are negatively affected, albeit to different degrees. Similarly, when the economy weakens, all firms feel the effects, though cyclical firms (such as automobiles, steel, and housing) may feel it more. We term this risk market risk. Finally, there are risks that fall in a gray area, depending on how many assets they affect. For instance, when the dollar strengthens against other currencies, it has a significant impact on the earnings and values of firms with international operations. If most firms in the market have significant international operations, it could well be categorized as market risk. If only a few do, it would be closer to firm-specific risk. Figure 4.4 summarizes the spectrum of firm-specific and market risks. Figure 4.4 Breakdown of Risk
Why Diversification Reduces or Eliminates Firm-Specific Risk: An Intuitive Explanation As an investor, you could invest all your portfolio in one asset. If you do so, you are exposed to both firm-specific and market risk. If, however, you expand your portfolio to include other assets or stocks, you are diversifying, and by doing so you can reduce your exposure to firm-specific risk. There are two reasons why diversification reduces or, at the limit, eliminates firm-specific risk. The first is that each investment in a diversified portfolio is a much smaller percentage of that portfolio than would be the case if you were not diversified. Any action that increases or decreases the value of only that investment or a small group of investments will have only a small impact on your overall portfolio, whereas undiversified investors are much more exposed to changes in the values of the investments in their portfolios. The second reason is that the effects of firm-specific actions on the prices of individual assets in a portfolio can be either positive or negative for each asset for any period. Thus, in very large portfolios this risk will average out to zero and will not affect the overall value of the portfolio. In contrast, the effects of marketwide movements are likely to be in the same direction for most or all investments in a portfolio, though some assets may be affected more than others. For instance, other things being equal, an increase in interest rates will lower the values of most assets in a portfolio. Being more diversified does not eliminate this risk.
A Statistical Analysis of Diversification-Reducing Risk The effects of diversification on risk can be illustrated fairly dramatically by examining the effects of increasing the number of assets in a portfolio on portfolio variance. The variance in a portfolio is partially determined by the variances of the individual assets in the portfolio and partially by how they move together; the latter is measured 57
statistically with a correlation coefficient or the covariance across investments in the portfolio. It is the covariance term that provides an insight into why diversification will reduce risk and by how much. Consider a portfolio of two assets. Asset A has an expected return of μA and a variance in returns of σ2A, while asset B has an expected return of μΒ and a variance in returns of σ2B. The correlation in returns between the two assets, which measures how the assets move together, is ρAB. The expected returns and variances of a two-asset portfolio can be written as a function of these inputs and the proportion of the portfolio going to each asset.
where wA = Proportion of the portfolio in asset A The last term in the variance formulation is sometimes written in terms of the covariance in returns between the two assets, which is:
The savings that accrue from diversification are a function of the correlation coefficient. Other things remaining equal, the higher the correlation in returns between the two assets, the smaller are the potential benefits from diversification. It is worth adding, though, that the benefits of correlation exist even for positively correlated assets and are non-existent only when the correlation is equal to one.
Mean-Variance Models Measuring Market Risk While most risk and return models in use in finance agree on the first two steps of the risk analysis process (i.e., that risk comes from the distribution of actual returns around the expected return and that risk should be measured from the perspective of a marginal investor who is well diversified), they part ways when it comes to measuring nondiversifiable or market risk. This section will discuss the different models that exist in finance for measuring market risk and why they differ. It begins with what still is the most widely used model for measuring market risk in finance—the capital asset pricing model (CAPM)—and then discusses the alternatives to this model that have developed over the past two decades. While the discussion will emphasize the differences, it will also look at what the models have in common.
Capital Asset Pricing Model The risk and return model that has been in use the longest and is still the standard for most practitioners is the capital asset pricing model (CAPM). This section will examine the assumptions on which the model is based and the measures of market risk that emerge from these assumptions.
WHY IS THE MARGINAL INVESTOR ASSUMED TO BE DIVERSIFIED? The argument that diversification reduces an investor’s exposure to risk is clear both intuitively and statistically, but risk and return models in finance go further. These models look at risk through the eyes of the investor most likely to be trading on the investment at any point in time— the marginal investor. They argue that this investor, who sets prices for investments, is well diversified; thus, the only risk that he or she cares about is the risk added to a diversified portfolio or market risk. This argument can be justified simply. The risk in an investment will always be perceived to be higher for an undiversified investor than for a diversified one, since the latter does not shoulder any firm-specific risk and the former does. If both investors have the same expectations about future earnings and cash flows on an asset, the diversified investor will be willing to pay a higher price for that asset because of his or her perception of lower risk. Consequently, the asset, over time, will end up being held by diversified investors. This argument is powerful, especially in markets where assets can be traded easily and at low cost. Thus, it works well for a stock traded in developed markets, since investors can become diversified at fairly low cost. In addition, a significant proportion of the trading in developed market stocks is done by institutional investors, who tend to be well diversified. It becomes a more difficult argument to sustain when assets cannot be easily traded or the costs of trading are high. In these markets, the marginal investor may well be undiversified, and firm-specific risk may therefore continue to matter when looking at individual investments. For instance, real estate in most countries is still held by investors who are undiversified and have the bulk of their wealth tied up in these investments.
Assumptions While diversification reduces the exposure of investors to firm-specific risk, most investors limit their diversification to holding only a few assets. Even large mutual funds rarely hold more than a few hundred stocks, and many of them hold as few as 10 to 20. There are two reasons why investors stop diversifying. One is that an investor or mutual fund manager can obtain most of the benefits of diversification from a relatively small portfolio, because the marginal benefits of diversification become smaller as the portfolio gets more diversified. Consequently, these benefits may not cover the marginal costs of diversification, which include transactions and monitoring costs. Another reason for limiting diversification is that many investors (and funds) believe they can find undervalued assets and thus choose not to hold those assets that they believe to be fairly valued or overvalued. 58
The capital asset pricing model assumes that there are no transaction costs, all assets are traded, and investments are infinitely divisible (i.e., you can buy any fraction of a unit of the asset). It also assumes that everyone has access to the same information and that investors therefore cannot find under- or overvalued assets in the marketplace. By making these assumptions, it allows investors to keep diversifying without additional cost. At the limit, their portfolios will not only include every traded asset in the market but these assets will be held in proportion to their market value. The fact that this portfolio includes all traded assets in the market is the reason it is called the market portfolio, which should not be a surprising result, given the benefits of diversification and the absence of transaction costs in the capital asset pricing model. If diversification reduces exposure to firm-specific risk and there are no costs associated with adding more assets to the portfolio, the logical limit to diversification is to hold a small proportion of every traded asset in the economy. If this seems abstract, consider the market portfolio to be an extremely well diversified mutual fund that holds stocks and real assets. In the CAPM, all investors will hold combinations of the riskier asset and that supremely diversified mutual fund.2
Investor Portfolios in the CAPM If every investor in the market holds the identical marke t portfolio, how exactly do investors reflect their risk aversion in their investments? In the capital asset pricing model, investors adjust for their risk preferences in their allocation decision, where they decide how much to invest in a riskless asset and how much in the market portfolio. Investors who are risk averse might choose to put much or even all of their wealth in the riskless asset. Investors who want to take more risk will invest the bulk or even all of their wealth in the market portfolio. Investors who invest all their wealth in the market portfolio and are desirous of taking on still more risk would do so by borrowing at the riskless rate and investing in the same market portfolio as everyone else. These results are predicated on two additional assumptions. First, there exists a riskless asset, where the expected returns are known with certainty. Second, investors can lend and borrow at the riskless rate to arrive at their optimal allocations. While lending at the riskless rate can be accomplished fairly simply by buying Treasury bills or bonds, borrowing at the riskless rate might be more difficult for individuals to do. There are variations of the CAPM that allow these assumptions to be relaxed and still arrive at conclusions that are consistent with the model.
Measuring the Market Risk of an Individual Asset The risk of any asset to an investor is the risk added by that asset to the investor’s overall portfolio. In the CAPM world, where all investors hold the market portfolio, the risk to an investor of an individual asset will be the risk that this asset adds to the market portfolio. Intuitively, if an asset moves independently of the market portfolio, it will not add much risk to the market portfolio. In other words, most of the risk in this asset is firm-specific and can be diversified away. In contrast, if an asset tends to move up when the market portfolio moves up and down when it moves down, it will add risk to the market portfolio. This asset has more market risk and less firm-specific risk. Statistically, this added risk is measured by the covariance of the asset with the market portfolio.
Measuring the Nondiversifiable Risk In a world in which investors hold a combination of only two assets—the riskless asset and the market portfolio— the risk of any individual asset will be measured relative to the market portfolio. In particular, the risk of any asset will be the risk it adds to the market portfolio. To arrive at the appropriate measure of this added risk, assume that σ2m is the variance of the market portfolio prior to the addition of the new asset and that the variance of the individual asset being added to this portfolio is σ2i. The market value portfolio weight on this asset is wi, and the covariance in returns between the individual asset and the market portfolio is σim. The variance of the market portfolio prior to and after the addition of the individual asset can then be written as:
The market value weight on any individual asset in the market portfolio should be small, since the market portfolio includes all traded assets in the economy. Consequently, the first term in the equation should approach zero, and the second term should approach σ2m′ leaving the third term (σim′ the covariance) as the measure of the risk added by asset i.
Standardizing Covariances The covariance is a percentage value, and it is difficult to pass judgment on the relative risk of an investment by looking at this value. In other words, knowing that the covariance of Boeing with the market portfolio is 55 percent does not provide us a clue as to whether Boeing is riskier or safer than the average asset. We therefore standardize the risk measure by dividing the covariance of each asset with the market portfolio by the variance of the market portfolio. This yields a risk measure called the beta of the asset: 59
Since the covariance of the market portfolio with itself is its variance, the beta of the market portfolio (and, by extension, the average asset in it) is 1. Assets that are riskier than average (using this measure of risk) will have betas that exceed 1, and assets that are safer than average will have betas that are lower than 1. The riskless asset will have a beta of zero.
Getting Expected Returns The fact that every investor holds some combination of the riskless asset and the market portfolio leads to the next conclusion, which is that the expected return on an asset is linearly related to the beta of the asset. In particular, the expected return on an asset can be written as a function of the risk-free rate and the beta of that asset:
where E(Ri) = Expected return on asset i Rf = Risk-free rate E(Rm) = Expected return on market portfolio βi = Beta of asset i To use the capital asset pricing model, we need three inputs. While the next chapter looks at the estimation process in far more detail, each of these inputs is estimated as follows: The riskless asset is defined to be an asset for which the investor knows the expected return with certainty for the time horizon of the analysis. The risk premium is the premium demanded by investors for investing in the market portfolio, which includes all risky assets in the market, instead of investing in a riskless asset. The beta, defined as the covariance of the asset divided by the market portfolio, measures the risk added by an investment to the market portfolio. In summary, in the capital asset pricing model all the market risk is captured in one beta measured relative to a market portfolio, which at least in theory should include all traded assets in the marketplace held in proportion to their market value.
Arbitrage Pricing Model The restrictive assumptions on transaction costs and private information in the capital asset pricing model, and the model’s dependence on the market portfolio, have long been viewed with skepticism by both academics and practitioners. Ross (1976) suggested an alternative model for measuring risk called the arbitrage pricing model (APM).
Assumptions If investors can invest risklessly and earn more than the riskless rate, they have found an arbitrage opportunity. The premise of the arbitrage pricing model is that investors take advantage of such arbitrage opportunities, and in the process eliminate them. If two portfolios have the same exposure to risk but offer different expected returns, investors will buy the portfolio that has the higher expected returns and sell the portfolio with the lower expected returns, and earn the difference as a riskless profit. To prevent this arbitrage from occurring, the two portfolios have to earn the same expected return. Like the capital asset pricing model, the arbitrage pricing model begins by breaking risk down into firm-specific and market risk components. As in the capital asset pricing model, firm-specific risk covers information that affects primarily the firm. Market risk affects many or all firms and would include unanticipated changes in a number of economic variables, including gross national product, inflation, and interest rates. Incorporating both types of risk into a return model, we get:
where R is the actual return, E(R) is the expected return, m is the marketwide component of unanticipated risk, and ε is the firm-specific component. Thus, the actual return can be different from the expected return, because of either market risk or firm-specific actions.
Sources of Marketwide Risk While both the capital asset pricing model and the arbitrage pricing model make a distinction between firm-specific and marketwide risk, they measure market risk differently. The CAPM assumes that market risk is captured in the 60
market portfolio, whereas the arbitrage pricing model allows for multiple sources of marketwide risk and measures the sensitivity of investments to changes in each source. In general, the market component of unanticipated returns can be decomposed into economic factors:
where βj = Sensitivity of investment to unanticipated changes in market risk factor j Fj = Unanticipated changes in market risk factor j Note that the measure of an investment’s sensitivity to any macroeconomic (or market) factor takes the form of a beta, called a factor beta. In fact, this beta has many of the same properties as the market beta in the CAPM.
Effects of Diversification The benefits of diversification were discussed earlier, in the context of the breakdown of risk into market and firmspecific risk. The primary point of that discussion was that diversification eliminates firm-specific risk. The arbitrage pricing model uses the same argument and concludes that the return on a portfolio will not have a firm-specific component of unanticipated returns. The return on a portfolio can be written as the sum of two weighted averages —that of the anticipated returns in the portfolio and that of the market factors:
where wj = Portfolio weight on asset j (where there are n assets) Rj = Expected return on asset j βi,j = Beta on factor i for asset j
Expected Returns and Betas The final step in this process is estimating an expected return as a function of the betas just specified. To do this, we should first note that the beta of a portfolio is the weighted average of the betas of the assets in the portfolio. This property, in conjunction with the absence of arbitrage, leads to the conclusion that expected returns should be linearly related to betas. To see why, assume that there is only one factor and three portfolios. Portfolio A has a beta of 2.0 and an expected return of 20 percent; portfolio B has a beta of 1.0 and an expected return of 12 percent; and portfolio C has a beta of 1.5 and an expected return of 14 percent. Note that investors can put half of their wealth in portfolio A and half in portfolio B and end up with portfolios with a beta of 1.5 and an expected return of 16 percent. Consequently no investor will choose to hold portfolio C until the prices of assets in that portfolio drop and the expected return increases to 16 percent. By the same rationale, the expected returns of every portfolio should be a linear function of the beta. If they were not, we could combine two other portfolios, one with a higher beta and one with a lower beta, to earn a higher return than the portfolio in question, creating an opportunity for arbitrage. This argument can be extended to multiple factors with the same results. Therefore, the expected return on an asset can be written as:
where Rf = Expected return on a zero-beta portfolio E(Rj) = Expected return on a portfolio with a factor beta of 1 for factor j, and zero for all other factors (where j = 1, 2, ... , K factors) The terms in the brackets can be considered to be risk premiums for each of the factors in the model. The capital asset pricing model can be considered to be a special case of the arbitrage pricing model, where there is only one economic factor driving marketwide returns, and the market portfolio is the factor.
The APM in Practice The arbitrage pricing model requires estimates of each of the factor betas and factor risk premiums in addition to the riskless rate. In practice, these are usually estimated using historical data on asset returns and a factor analysis. Intuitively, in a factor analysis, we examine the historical data looking for common patterns that affect broad groups of assets (rather than just one sector or a few assets). A factor analysis provides two output measures: 1. It specifies the number of common factors that affected the historical return data. 2. It measures the beta of each investment relative to each of the common factors and provides an estimate of 61
the actual risk premium earned by each factor. The factor analysis does not, however, identify the factors in economic terms. In summary, in the arbitrage pricing model the market risk is measured relative to multiple unspecified macroeconomic variables, with the sensitivity of the investment relative to each factor being measured by a beta. The number of factors, the factor betas, and the factor risk premiums can all be estimated using the factor analysis.
Multifactor Models for Risk and Return The arbitrage pricing model’s failure to identify the factors specifically in the model may be a statistical strength, but it is an intuitive weakness. The solution seems simple: Replace the unidentified statistical factors with specific economic factors, and the resultant model should have an economic basis while still retaining much of the strength of the arbitrage pricing model. That is precisely what multifactor models try to do.
Deriving a Multifactor Model Multifactor models generally are determined by historical data rather than by economic modeling. Once the number of factors has been identified in the arbitrage pricing model, their behavior over time can be extracted from the data. The behavior of the unnamed factors over time can then be compared to the behavior of macroeconomic variables over that same period, to see whether any of the variables is correlated, over time, with the identified factors. For instance, Chen, Roll, and Ross (1986) suggest that the following macroeconomic variables are highly correlated with the factors that come out of factor analysis: industrial production, changes in default premium, shifts in the term structure, unanticipated inflation, and changes in the real rate of return. These variables can then be correlated with returns to come up with a model of expected returns, with firm-specific betas calculated relative to each variable.
where βGNP = Beta relative to changes in industrial production E(RGNP) = Expected return on a portfolio with a beta of one on the industrial production factor and zero on all other factors βI = Beta relative to changes in inflation E(RI) = Expected return on a portfolio with a beta of one on the inflation factor and zero on all other factors The costs of going from the arbitrage pricing model to a macroeconomic multifactor model can be traced directly to the errors that can be made in identifying the factors. The economic factors in the model can change over time, as will the risk premium associated with each one. For instance, oil price changes were a significant economic factor driving expected returns in the 1970s but are not as significant in other time periods. Using the wrong factor or missing a significant factor in a multifactor model can lead to inferior estimates of expected return.
ALTERNATIVE MODELS FOR EQUITY RISK The CAPM, arbitrage pricing model, and multifactor model represent attempts by financial economists to build risk and return models from the mean-variance base established by Harry Markowitz (1991). The re are many, though, who believe the basis for the model is flawed and that we should be looking at alternatives, and in this section, we will look at some of them.
Different Return Distributions From its very beginnings, the mean-variance framework has been controversial. While there have been many who have challenged its applicability, we will consider these challenges in three groups. The first group argues that stock prices, in particular, and investment returns, in general, exhibit too many large values to be drawn from a normal distribution. They argue that the fat tails on stock price distributions lend themselves better to a class of distributions, called power law distributions, which exhibit infinite variance and long periods of price dependence. The second group takes issue with the symmetry of the normal distribution and argues for measures that incorporate the asymmetry observed in actual return distributions into risk measures. The third group posits that distributions that allow for price jumps are more realistic and that risk measures should consider the likelihood and magnitude of price jumps.
Fat Tails and Power Law Distributions Benoit Mandelbrot (1961; Mandelbrot and Hudson, 2004), a mathematician who also did pioneering work on the behavior of stock prices, was one of those who took issue with the use of normal and lognormal distributions. He argued, based on his observation of stock and real asset prices, that a power law distribution characterized them better. In a power-law distribution, the relationship between two variables, Y and X, can be written as follows: 62
In this equation, α is a constant (constant of proportionality), and k is the power law exponent. Mandelbrot’s key point was that the normal and log normal distributions were best suited for series that exhibited mild and wellbehaved randomness, whereas power law distributions were more suited for series that exhibited large movements and what he termed wild randomness. Wild randomness occurs when a single observation can affect the population in a disproportionate way; stock and commodity prices exhibit wild randomness. Stock and commodity prices, with their long periods of relatively small movements, punctuated by wild swings in both directions, seem to fit better into the wild randomness group. What are the consequences for risk measures? If asset prices follow power law distributions, the standard deviation or volatility ceases to be a good risk measure and a good basis for computing probabilities. Assume, for instance, that the standard deviation in annual stock returns is 15 percent and that the average return is 10 percent. Using the normal distribution as the base for probability predictions, this will imply that the stock returns will exceed 40 percent (average plus two standard deviations) only once every 44 years and 55 percent only (average plus three standard deviations) only once every 740 years. In fact, stock returns will be greater than 85 percent (average plus five standard deviations) only once every 3.5 million years. In reality, stock returns exceed these values far more frequently, a finding consistent with power law distributions, where the probability of larger values declines linearly as a function of the power law exponent. As the value gets doubled, the probability of its occurrence drops by the square of the exponent. Thus, if the exponent in the distribution is 2, the likelihood of returns of 25 percent, 50 percent, and 100 percent can be computed as follows: Returns will exceed 25 percent: once every 6 years. Returns will exceed 50 percent: once every 24 years. Returns will exceed 100 percent: once every 96 years. Note that as the returns get doubled, the likelihood increases four-fold (the square of the exponent). As the exponent decreases, the likelihood of larger values increases; an exponent between 0 and 2 will yield extreme values more often than a normal distribution. An exponent between 1 and 2 yields power law distributions called stable Paretian distributions, which have infinite variance. In an early study, Fama (1965) estimated the exponent for stocks to be between 1.7 and 1.9, but subsequent studies have found that the exponent is higher in both equity and currency markets.3 In practical terms, the power law proponents argue that using measures such as volatility (and its derivatives such as beta) underestimate the risk of large movements. The power law exponents for assets, in their view, provide investors with more realistic risk measures for these assets. Assets with higher exponents are less risky (since extreme values become less common) than asset with lower exponents. Mandelbrot’s challenge to the normal distribution was more than a procedural one. Mandelbrot’s world, in contrast to the Gaussian mean-variance one, is a world where prices move jaggedly over time and look as though they have no pattern at a distance, but where patterns repeat themselves, when observed closely. In the 1970s, Mandelbrot created a branch of mathematics called fractal geometry where processes are not described by conventional statistical or mathematical measures but by fractals; a fractal is a geometric shape that when broken down into smaller parts replicates that shape. To illustrate the concept, he uses the example of the coastline that, from a distance, looks irregular and up close looks roughly the same—fractal patterns repeat themselves. In fractal geometry, higher fractal dimensions translate into more jagged shapes; the rugged Cornish coastline has a fractal dimension of 1.25 whereas the much smoother South African coastline has a fractal dimension of 1.02. Using the same reasoning, stock prices that look random, when observed at longer time intervals, start revealing selfrepeating patterns, when observed over shorter time periods. More volatile stocks score higher on measures of fractal dimension, thus making it a measure of risk. With fractal geometry, Mandelbrot was able to explain not only the higher frequency of price jumps (relative to the normal distribution) but also long periods where prices move in the same direction and the resulting price bubbles.
Asymmetric Distributions Intuitively, it should be downside risk that concerns us and not upside risk. In other words, it is not investments that go up significantly that create heartburn and unease but investments that go down significantly. The mean-variance framework, by weighting both upside volatility and downside movements equally, does not distinguish between the two. With a normal or any other symmetric distribution, the distinction between upside and downside risk is irrelevant because the risks are equivalent. With asymmetric distributions, though, there can be a difference between upside and downside risk. Studies of risk aversion in humans conclude that (1) they are loss averse; that is, they weigh the pain of a loss more than the joy of an equivalent gain and (2) they value very large positive payoffs—long shots—far more than they should given the likelihood of these payoffs. In practice, return distributions for stocks and most other assets are not symmetric. Instead, asset returns exhibit fat tails (i.e, more jumps) and are more likely to have extreme positive values than extreme negative values (simply 63
because returns are constrained to be no less than –100 percent). As a consequence, the distribution of stock returns has a higher incidence of extreme returns (fat tails or kurtosis) and a tilt toward very large positive returns (positive skewness). Critics of the mean-variance approach argue that it takes too narrow a view of both rewards and risk. In their view, a fuller return measure should consider not just the magnitude of expected returns but also the likelihood of very large positive returns or skewness, and a more complete risk measure should incorporate both variance and the possibility of big jumps (co-kurtosis). Note that even as these approaches deviate from the meanvariance approach in terms of how they define risk, they stay true to the portfolio measure of risk. In other words, it is not the possibility of large positive payoffs (skewness) or big jumps (kurtosis) that they argue should be considered, but only that portion of the skewness (co-skewness) and kurtosis (co-kurtosis) that is market-related and not diversifiable.
Jump Process Models The normal, power law, and asymmetric distributions that form the basis for the models we have discussed in this section are all continuous distributions. Observing the reality that stock prices do jump, there are some who have argued for the use of jump process distributions to derive risk measures. Press (1967), in one of the earliest papers that attempted to model stock price jumps, argued that stock prices follow a combination of a continuous price distribution and a Poisson distribution, where prices jump at irregular intervals. The key parameters of the Poisson distribution are the expected size of the price jump (μ), the variance in this value (δ2), and the likelihood of a price jump in any specified time period (λ), and Press estimated these values for 10 stocks. In subsequent papers, Beckers (1981) and Ball and Torous (1983) suggest ways of refining these estimates. In an attempt to bridge the gap between the CAPM and jump process models, Jarrow and Rosenfeld (1984) derive a version of the capital asset pricing model that includes a jump component that captures the likelihood of market jumps and an individual asset’s correlation with these jumps. While jump process models have gained some traction in option pricing, they have had limited success in equity markets, largely because the parameters of jump process models are difficult to estimate with any degree of precision. Thus, while everyone agrees that stock prices jump, there is little consensus on the best way to measure how often this happens, whether these jumps are diversifiable, and how best to incorporate their effect into risk measures.
Regression or Proxy Models The conventional models for risk and return in finance (CAPM, arbitrage pricing model, and even multifactor models) start by making assumptions about how investors behave and how markets work to derive models that measure risk and link those measures to expected returns. While these models have the advantage of a foundation in economic theory, they seem to fall short in explaining differences in returns across investments. The reasons for the failure of these models run the gamut: The assumptions made about markets are unrealistic (no transactions costs, perfect information) and investors don’t behave rationally (and behavioral finance research provides ample evidence of this). With proxy models, we essentially give up on building risk and return models from economic theory. Instead, we start with how investments are priced by markets and relate returns earned to observable variables. Rather than talk in abstractions, consider the work done by Fama and French in the early 1990s. Examining returns earned by individual stocks from 1962 to 1990, they concluded that CAPM betas did not explain much of the variation in these returns. They then took a different tack and looked for company-specific variables that did a better job of explaining return differences they pinpointed two variables—the market capitalization of a firm and its price-to-book ratio (the ratio of market cap to accounting book value for equity). Specifically, they concluded that small market cap stocks earned much higher annual returns than large market cap stocks and that low price to book ratio stocks earned much higher annual returns than stocks that traded at high price-to-book ratios. Rather than view this as evidence of market inefficiency (which is what prior studies that had found the same phenomena had done), they argued if these stocks earned higher returns over long time periods, they must be riskier than stocks that earned lower returns. In effect, market capitalization and price-to-book ratios were better proxies for risk, according to their reasoning, than betas. In fact, they regressed returns on stocks against the market capitalization of a company and its price-to-book ratio to arrive at the following regression for U.S. stocks;
In a pure proxy model, you could plug the market capitalization and book-to-market ratio for any company into this regression to get expected monthly returns. In the two decades since the Fama-French paper brought proxy models to the fore, researchers have probed the data (which has become more detailed and voluminous over time) to find better and additional proxies for risk. Some of the proxies are highlighted here:
Earnings momentum. Equity research analysts will find vindication in research that seems to indicate that 64
companies that have reported stronger than expected earnings growth in the past earn higher returns than the rest of the market. Price momentum. Chartists will smile when they read this, but researchers have concluded that price momentum carries over into future periods. Thus, the expected returns will be higher for stocks that have outperformed markets in recent time periods and lower for stocks that have lagged. Liquidity. In a nod to real-world costs, there seems to be clear evidence that stocks that are less liquid (lower trading volume, higher bid-ask spreads) earn higher returns than more liquid stocks. While the use of pure proxy models by practitioners is rare, they have adapted the findings for these models into their day-to-day use. Many analysts have melded the CAPM with proxy models to create composite or melded models. For instance, many analysts who value small companies derive expected returns for these companies by adding a small cap premium to the CAPM expected return:
The threshold for small capitalization varies across time but is generally set at the bottom decile of publicly traded companies, and the small cap premium itself is estimated by looking at the historical premium earned by small cap stocks over the market. Using the Fama-French findings, the CAPM has been expanded to include market capitalization and price-to-book ratios as additional variables, with the expected return stated as:
The size and the book-to-market betas are estimated by regressing a stock’s returns against the size premium and book-to-market premiums over time; this is analogous to the way we get the market beta, by regressing stock returns against overall market returns. While the use of proxy and melded models offers a way of adjusting expected returns to reflect market reality, there are three dangers in using these models. 1. Data mining. As the amount of data that we have on companies increases and becomes more accessible, it is inevitable that we will find more variables that are related to returns. It is also likely that most of these variables are not proxies for risk and that the correlation is a function of the time period that we look at. In effect, proxy models are statistical models and not economic models. Thus, there is no easy way to separate the variables that matter from those that do not. 2. Standard error. Since proxy models come from looking at historical data, they carry all of the burden of the noise in the data. Stock returns are extremely volatile over time, and any historical premia that we compute (for market capitalization or any other variable) are going to have significant standard errors. The standard errors on the size and book-to-market betas in the three-factor Fama-French model may be so large that using them in practice creates almost as much noise as it adds in precision. 3. Pricing error or risk proxy. For decades, value investors have argued that you should invest in stocks with low PE ratios that trade at low multiples of book value and have high dividend yields, pointing to the fact that you will earn higher returns by doing so. (In fact, a scan of Benjamin Graham’s screens from security analysis4 for cheap companies unearths most of the proxies that you see in use today.) Proxy models incorporate all of these variables into the expected return and thus render these assets to be fairly priced. Using the circular logic of these models, markets are always efficient because any inefficiency that exists is just another r isk proxy that needs to get built into the model.
A COMPARATIVE ANALYSIS OF EQUITY RISK MODELS When faced with the choice of estimating expected returns on equity or cost of equity, we are therefore faced with several choices, ranging from the CAPM to proxy models. Table 4.1 summarizes the different models and presents their pluses and minuses. Table 4.1 Alternative Models for Cost of Equity
65
The decision has to be based as much on theoretical considerations as it will be on pragmatic considerations. The CAPM is the simplest of the models, insofar as it requires only one firm-specific input (the beta), and that input can be estimated readily from public information. To replace the CAPM with an alternative model, whether it be from the mean variance family (arbitrage pricing model or multifactor models), alternative return process families (power, asymmetric, and jump distribution models), or proxy models, we need evidence of substantial improvement in accuracy in future forecasts (and not just in explaining past returns). Ultimately, the survival of the capital asset pricing model as the default model for risk in real-world applications is a testament to both its intuitive appeal and the failure of more complex models to deliver significant improvement in terms of estimating expected returns. We would argue that a judicious use of the capital asset pricing model, without an over reliance on historical data, is still the most effective way of dealing with risk in valuation in most cases. In some sectors (commodities) and segments (closely held companies, illiquid stocks), using other, more complete models will be justified. We will return to the question of how improvements in estimating the inputs to the CAPM can generate far more payoff than switching to more complicated models for cost of equity.
MODELS OF DEFAULT RISK The risk discussed so far in this chapter relates to cash flows on investments being different from expected cash flows. There are some investments, however, in which the cash flows are promised when the investment is made. This is the case, for instance, when you lend to a business or buy a corporate bond; the borrower may default on interest and principal payments on the borrowing. Generally speaking, borrowers with higher default risk should pay higher interest rates on their borrowing than those with lower default risk. This section examines the measurement of default risk and the relationship of default risk to interest rates on borrowing. In contrast to the general risk and return models for equity, which evaluate the effects of market risk on expected returns, models of default risk measure the consequences of firm-specific default risk on promised returns. While diversification can be used to explain why firm-specific risk will not be priced into expected returns for equities, the same rationale cannot be applied to securities that have limited upside potential and much greater downside potential from firm-specific events. To see what is meant by limited upside potential, consider investing in the bond issued by a company. The coupons are fixed at the time of the issue, and these coupons represent the promised cash flow on the bond. The best-case scenario for you as an investor is that you receive the promised cash flows; you are not entitled to more than these cash flows even if the company is wildly successful. All other scenarios contain only bad news, though in varying degrees, with the delivered cash flows being less than the promised cash flows. Consequently, the expected return on a corporate bond is likely to reflect the firm-specific default risk of the firm issuing the bond.
Determinants of Default Risk The default risk of a firm is a function of two variables. The first is the firm’s capacity to generate cash flows from operations, and the second is its financial obligations—including interest and principal payments.5 Firms that generate high cash flows relative to their financial obligations should have lower default risk than do firms that generate low cash flows relative to obligations. Thus, firms with significant existing investments that generate high cash flows will have lower default risk than will firms that do not have such investm ents. In addition to the magnitude of a firm’s cash flows, the default risk is also affected by the volatility in these cash flows. The more stability there is in cash flows, the lower is the default risk in the firm. Firms that operate in predictable and stable businesses will have lower default risk than will otherwise similar firms that operate in cyclical or volatile businesses. 66
Most models of default risk use financial ratios to measure the cash flow coverage (i.e., the magnitude of cash flows relative to obligations) and control for industry effects in order to evaluate the variability in cash flows.
Bond Ratings and Interest Rates The most widely used measure of a firm’s default risk is its bond rating, which is generally assigned by an independent ratings agency. The two best known are Standard & Poor’s (S&P) and Moody’s. Thousands of companies are rated by these two agencies, and their views carry significant weight with financial markets.
The Ratings Process The process of rating a bond starts when the issuing company requests a rating from a bond ratings agency. The ratings agency then collects information from both publicly available sources, such as financial statements, and the company itself and makes a decision on the rating. If the company disagrees with the rating, it is given the opportunity to present additional information. This process is presented schematically for one ratings agency, Standard & Poor’s, in Figure 4.5. Figure 4.5 The Ratings Process
The ratings assigned by these agencies are letter ratings. A rating of AAA from Standard & Poor’s and Aaa from Moody’s represents the highest rating, granted to firms that are viewed as having the lowest default risk. As the default risk increases, the ratings decline toward D for firms in default (Standard & Poor’s). A rating at or above BBB by Standard & Poor’s (or Baa by Moody’s) is categorized as above investment grade, reflecting the view of the ratings agency that there is relatively little default risk in investing in bonds issued by these firms.
Determinants of Bond Ratings The bond ratings assigned by ratings agencies are primarily based on publicly available information, though private information conveyed by the firm to the ratings agency does play a role. The rating assigned to a company’s bonds will depend in large part on financial ratios that measure the capacity of the company to meet debt payments and generate stable and predictable cash flows. While a multitude of financial ratios exist, Table 4.2 summarizes some of the key ratios used to measure default risk. Table 4.2 Definition of Financial Ratios: S&P Financial Ratio
Definition
EBITDA/Revenues
EBITDA/Revenues
ROIC
ROIC = EBIT/(BV of debt + BV of equity – Cash)
EBIT/Interest expenses Interest coverage ratio
67
EBITDA/Interest
EBITDA/Interest expenses
FFO/debt
(Net Income + Depreciation)/Debt
Free operating CF/Debt Funds from operations/Debt Discounted cash flows/Debt Discounted CF/Debt Debt/EBITDA
BV of Debt/EBITDA
D/(D+E)
BV of Debt/(BV of Debt + BV of equity)
There is a strong relationship between the bond rating a company receives and its performance on these financial ratios. Table 4.3 provides a summary of the median ratios6 from 2007 to 2009 for different S&P ratings classes for manufacturing firms. Table 4.3 Financial Ratios and S&P Ratings
Not surprisingly, firms that generate income and cash flows significantly higher than debt payments, that are profitable, and that have low debt ratios are more likely to be highly rated than are firms that do not have these characteristics. There will be individual firms whose ratings are not consistent with their financial ratios, however, because the ratings agency does add subjective judgments into the final mix. Thus a firm that performs poorly on financial ratios but is expected to improve its performance dramatically over the next period may receive a higher rating than is justified by its current financials. For most firms, however, the financial ratios should provide a reasonable basis for estimating the bond rating.
Bond Ratings and Interest Rates The interest rate on a corporate bond should be a function of its default risk, which is measured by its rating. If the rating is a good measure of the default risk, higher-rated bonds should be priced to yield lower interest rates than those of lower-rated bonds. In fact, the difference between the interest rate on a bond with default risk and a default-free government bond is the default spread. This default spread will vary by maturity of the bond and can also change from period to period, depending on economic conditions. Chapter 7 considers how best to estimate these default spreads and how they might vary over time.
CONCLUSION Risk, as defined in finance, is measured based on deviations of actual returns on an investment from its expected returns. There are two types of risk. The first, called equity risk, arises in investments where there are no promised cash flows, but there are expected cash flows. Th e second, default risk, arises on investments with promised cash flows. On investments with equity risk, the risk is best measured by looking at the variance of actual returns around the expected returns, with greater variance indicating greater risk. This risk can be broken down into risk that affects one or a few investments, called firm-specific risk, and risk that affects many investments, refered to as market risk. When investors diversify, they can reduce their exposure to firm-specific risk. If we assume that the investors who trade at the margin are well diversified, the risk we should be looking at with equity investments is the nondiversifiable or market risk. The different models of equity risk introduced in this chapter share this objective of measuring market risk, but they differ in the way they do it. In the capital asset pricing model, exposure to market risk is measured by a market beta, which estimates how much risk an individual investment will add to a portfolio that includes all traded assets. The arbitrage pricing model and the multifactor model allow for multiple sources of market risk and estimate betas for an investment relative to each source. Regression or proxy models for risk look for firm characteristics, such as size, that have been correlated with high returns in the past and use these to measure market risk. In all these models, the risk measures are used to estimate the expected return on an equity investment. This expected return can be considered the cost of equity for a company. 68
On investments with default risk, risk is measured by the likelihood that the promised cash flows might not be delivered. Investments with higher default risk should have higher interest rates, and the premium that we demand over a riskless rate is the default spread. For many U.S. companies, default risk is measured by rating agencies in the form of a bond rating; these ratings determine, in large part, the interest rates at which these firms can borrow. Even in the absence of ratings, interest rates will include a default spread that reflects the lenders’ assessments of default risk. These default-risk-adjusted interest rates represent the cost of borrowing or debt for a business.
QUESTIONS AND SHORT PROBLEMS In the problems following, use an equity risk premium of 5.5 percent if none is specified. 1. The following table list s the stock prices for Microsoft from 1989 to 1998. The company did not pay any dividends during the period. Year Price 1989 $ 1.20 1990 $ 2.09 1991 $ 4.64 1992 $ 5.34 1993 $ 5.05 1994 $ 7.64 1995 $10.97 1996 $20.66 1997 $32.31 1998 $69.34
a. Estimate the average annual return you would have made on your investment. b. Estimate the standard deviation and variance in annual returns. c. If you were investing in Microsoft today, would you expect the historical standard deviations and variances to continue to hold? Why or why not? 2. Unicom is a regulated utility serving northern Illinois. The following table lists the stock prices and dividends on Unicom from 1989 to 1998. Year Price
Dividends
1989 $36.10 $3.00 1990 $33.60 $3.00 1991 $37.80 $3.00 1992 $30.90 $2.30 1993 $26.80 $1.60 1994 $24.80 $1.60 1995 $31.60 $1.60 1996 $28.50 $1.60 1997 $24.25 $1.60 1998 $35.60 $1.60
a. Estimate the average annual return you would have made on your investment. b. Estimate the standard deviation and variance in annual returns. c. If you were investing in Unicom today, would you expect the historical standard deviations and variances to continue to hold? Why or why not? 3. The following table summarizes the annual returns you would have made on two companies—Scientific Atlanta, a satellite and data equipment manufacturer, and AT&T, the telecommunications giant—from 1989 to 1998. Year Scientific Atlanta AT&T 1989 80.95%
58.26%
1990 –47.37%
–33.79%
69
1991 31.00%
29.88%
1992 132.44%
30.35%
1993 32.02%
2.94%
1994 25.37%
–4.29%
1995 –28.57%
28.86%
1996 0.00%
–6.36%
1997 11.67%
48.64%
1998 36.19%
23.55%
a. Estimate the average annual return and standard deviation in annual returns in each company. b. Estimate the covariance and correlation in returns between the two companies. c. Estimate the variance of a portfolio composed, in equal parts, of the two investments. 4. You are in a world where there are only two assets, gold and stocks. You are interested in investing your money in one, the other, or both assets. Consequently you collect the following data on the returns on the two assets over the past six years. Gold Stock Market Average return
8%
20%
Standard deviation 25% 22% Correlation
–0.4
a. If you were constrained to pick just one, which one would you choose? b. A friend argues that this is wrong. He says that you are ignoring the big payoffs that you can get on the other asset. How would you go about alleviating his concern? c. How would a portfolio composed of equal proportions in gold and stocks do in terms of mean and variance? d. You now learn that GPEC (a cartel of gold-producing countries) is going to vary the amount of gold it produces in relation to stock prices in the United States. (GPEC will produce less gold when stock markets are up and more when they are down.) What effect will this have on your portfolio? Explain. 5. You are interested in creating a portfolio of two stocks—Coca-Cola and Texas Utilities. Over the past decade, an investment in Coca-Cola stock would have earned an average annual return of 25%, with a standard deviation in returns of 36%. An investment in Texas Utilities stock would have earned an average annual return of 12%, with a standard deviation of 22%. The correlation in returns across the two stocks is 0.28. a. Assuming that the average return and standard deviation, estimated using past returns, will continue to hold in the future, estimate the future average returns and standard deviation of a portfolio composed 60% of Coca-Cola and 40% of Texas Utilities stock. b. Now assume that Coca-Cola’s international diversification will reduce the correlation to 0.20, while increasing Coca-Cola’s standard deviation in returns to 45%. Assuming all of the other numbers remain unchanged, estimate one standard deviation of the portfolio in (a). 6. Assume that you have half your money invested in Times Mirror, the media company, and the other half invested in Unilever, the consumer product company. The expected returns and standard deviations on the two investments are: Times Mirror Unilever Expected return
14%
18%
Standard deviation 25%
40%
Estimate the variance of the portfolio as a function of the correlation coefficient (start with –1 and increase the correlation to +1 in 0.2 increments). 7. You have been asked to analyze the standard deviation of a portfolio composed of the following three assets: Expected Return Standard Deviation Sony Corporation
11%
23%
Tesoro Petroleum
9%
27%
Storage Technology 16%
50%
70
You have also been provided with the correlations across these three investments:
Estimate the variance of a portfolio, equally weighted across all three assets. 8. Assume that the average variance of return for an individual security is 50 and that the average covariance is 10. What is the expected variance of a portfolio of 5, 10, 20, 50, and 100 securities? How many securities need to be held before the risk of a portfolio is only 10% more than the minimum? 9. Assume you have all your wealth (a million dollars) invested in the Vanguard 500 index fund, and that you expect to earn an annual return of 12%, with a standard deviation in returns of 25%. Since you have become more risk averse, you decide to shift $200,000 from the Vanguard 500 index fund to Treasury bills. The T-bill rate is 5%. Estimate the expected return and standard deviation of your new portfolio. 10. Every investor in the capital asset pricing model owns a combination of the market portfolio and a riskless asset. Assume that the standard deviation of the market portfolio is 30% and that the expected return on the portfolio is 15%. What proportion of the following investors’ wealth would you suggest investing in the market portfolio and what proportion in the riskless asset? (The riskless asset has an expected return of 5%.) a. An investor who desires a portfolio with no standard deviation. b. An investor who desires a portfolio with a standard deviation of 15%. c. An investor who desires a portfolio with a standard deviation of 30%. d. An investor who desires a portfolio with a standard deviation of 45%. e. An investor who desires a portfolio with an expected return of 12%. 11. The following table lists returns on the market portfolio and on Scientific Atlanta, each year from 1989 to 1998. Year Scientific Atlanta Market Portfolio 1989 80.95%
31.49%
1990 –47.37%
–3.17%
1991 31.00%
30.57%
1992 132.44%
7.58%
1993 32.02%
10.36%
1994 25.37%
2.55%
1995 –28.57%
37.57%
1996 0.00%
22.68%
1997 11.67%
33.10%
1998 36.19%
28.32%
a. Estimate the covariance in returns between Scientific Atlanta and the market portfolio. b. Estimate the variances in returns on both investments. c. Estimate the beta for Scientific Atlanta. 12. United Airlines has a beta of 1.5. The standard deviation in the market portfolio is 22%, and United Airlines has a standard deviation of 66%. a. Estimate the correlation between United Airlines and the market portfolio. b. What proportion of United Airlines’ risk is market risk? 13. You are using the arbitrage pricing model to estimate the expected return on Bethlehem Steel, and have derived the following estimates for the factor betas and risk premium: Factor Beta Risk Premium
71
1
1.2
2.5%
2
0.6
1.5%
3
1.5
1.0%
4
2.2
0.8%
5
0.5
1.2%
a. Which risk factor is Bethlehem Steel most exposed to? Is there any way, within the arbitrage pricing model, to identify the risk factor? b. If the risk-free rate is 5%, estimate the expected return on Bethlehem Steel. c. Now assume that the beta in the capital asset pricing model for Bethlehem Steel is 1.1, and that the risk premium for the market portfolio is 5%. Estimate the expected return using the CAPM. d. Why are the expected returns different using the two models? 14. You are using the multifactor model to estimate the expected return on Emerson Electric, and have derived the following estimates for the factor betas and risk premiums:
With a riskless rate of 6%, estimate the expected return on Emerson Electric. 15. The following equation is reproduced from the study by Fama and French of returns between 1963 and 1990.
where MV is the market value of equity in hundreds of millions of dollars and BV is the book value of equity in hundreds of millions of dollars. The return is a monthly return. a. Estimate the expected annual return on Lucent Technologies if the market value of its equity is $180 billion and the book value of its equity is $73.5 billion. b. Lucent Technologies has a beta of 1.55. If the riskless rate is 6% and the risk premium for the market portfolio is 5.5%, estimate the expected return. c. Why are the expected returns different under the two approaches? A utility function is a way of summarizing investor preferences into a generic term called “utility” on the basis of some choice variables. In this case, for instance, the investors’ utility or satisfaction is stated as a function of wealth. By doing so, we effectively can answer questions such as, Will investors be twice as happy if they have twice as much wealth? Do es each marginal increase in wealth lead to less additional utility than the prior marginal increase? In one specific form of this function, the quadratic utility function, the entire utility of an investor can be compressed into the expected wealth measure and the standard deviation in that wealth. 1
The significance of introducing the riskless asset into the choice mix and the implications for portfolio choice were first noted in Sharpe (1964) and Lintner (1965). Hence, the model is sometimes called the Sharpe-Lintner model. 2
3 In a paper in Nature (Gabaix, X., Gopikrishnan, P., Plerou, V., and Stanley, H.E., 2003, A theory of power law distributions in financial market fluctuations, Nature 423, 267–70), researchers looked at stock prices on 500 stocks between 1929 and 1987 and con cluded that the exponent for stock returns is roughly 3. 4
Graham, B., 1949, The Intelligent Investor (New York: HarperBusiness, reprinted in 2005).
Financial obligation refers t o any payment that the firm has legally obligated itself to make, such as interest and principal payments. It does not include discretionary cash flows, such as dividend payments or new capital expenditures, which can be deferred or delayed without legal consequences, though there may be economic consequences. 5
72
6
See the Standard & Poor’s online site (www.standardandpoors.com/ratings/criteria/index.htm).
73
CHAPTER 5 Option Pricing Theory and Models In general, the value of any asset is the present value of the expected cash flows on that asset. This chapter considers an exception to that rule when it looks at assets with two s pecific characteristics: 1. The assets derive their value from the values of other assets. 2. The cash flows on the assets are contingent on the occurrence of specific events. These assets are called options, and the present value of the expected cash flows on these assets will understate their true value. This chapter describes the cash flow characteristics of options, considers the factors that determine their value, and examines how best to value them.
BASICS OF OPTION PRICING An option provides the holder with the right to buy or sell a specified quantity of an underlying asset at a fixed price (called a strike price or an exercise price) at or before the expiration date of the option. Since it is a right and not an obligation, the holder can choose not to exercise the right and can a llow the option to expire. There are two types of options—call options and put options.
Call and Put Options: Description and Payoff Diagrams A call option gives the buyer of the option the right to buy the underlying asset at the strike price or the exercise price at any time prior to the expiration date of the option. The buyer pays a price for this right. If at expiration the value of the asset is less than the strike price, the option is not exercised and expires worthless. If, however, the value of the asset is greater than the strike price, the option is exercised—the buyer of the option buys the stock at the exercise price, and the difference between the asset value and the exercise price comprises the gross profit on the investment. The net profit on the investment is the difference between the gross profit and the price paid for the call initially. A payoff diagram illustrates the cash payoff on an option at expiration. For a call, the net payoff is negative (and equal to the price paid for the call) if the value of the underlying asset is less than the strike price. If the price of the underlying asset exceeds the strike price, the gross payoff is the difference between the value of the underlying asset and the strike price, and the net payoff is the difference between the gross payoff and the price of the call. This is illustrated in Figure 5.1. Figure 5.1 Payoff on Call Option
A put option gives the buyer of the option the right to sell the underlying asset at a fixed price, again called the strike or exercise price, at any time prior to the expiration date of the option. The buyer pays a price for this right. If the price of the underlying asset is greater than the strike price, the option will not be exercised and will expire worthless. But if the price of the underlying asset is less than the strike price, the owner of the put option will exercise the option and sell the stock at the strike price, claiming the difference between the strike price and the market value of the asset as the gross profit. Again, netting out the initial cost paid for the put yields the net profit from the transaction. A put has a negative net payoff if the value of the underlying asset exceeds the strike price, and has a gross payoff equal to the difference between the strike price and the value of the underlying asset if the asset value is less than the strike price. This is summarized in Figure 5.2. 74
Figure 5.2 Payoff on Put Option
DETERMINANTS OF OPTION VALU E The value of an optio n is determined by six variables relating to the underlying asset and financial markets. 1. Current value of the underlying asset. Options are assets that derive value from an underlying asset. Consequently, changes in the value of the underlying asset affect the value of the options on that asset. Since calls provide the right to buy the underlying asset at a fixed price, an increase in the value of the asset will increase the value of the calls. Puts, on the other hand, become less valuable as the value of the asset increases. 2. Variance in value of the underlying asset. The buyer of an option acquires the right to buy or sell the underlying asset at a fixed price. The higher the variance in the value of the underlying asset, the greater the value of the option.1 This is true for both calls and puts. While it may seem counterintuitive that an increase in a risk measure (variance) should increase value, options are different from other securities since buyers of options can never lose more than the price they pay for them; in fact, they have the potential to earn significant returns from large price movements. 3. Dividends paid on the underlying asset. The value of the underlying asset can be expected to decrease if dividend payments are made on the asset during the life of the option. Consequently, the value of a call on the asset is a decreasing function of the size of expected dividend payments, and the value of a put is an increasing function of expected dividend payments. A more intuitive way of thinking about dividend payments, for call options, is as a cost of delaying exercise on in-the-money options. To see why, consider an option on a traded stock. Once a call option is in-the-money (i.e., the holder of the option will make a gross payoff by exercising the option), exercising the call option will provide the holder with the stock and entitle him or her to the dividends on the stock in subsequent periods. Failing to exercise the option will mean that these dividends are forgone. 4. Strike price of the option. A key characteristic used to describe an option is the strike price. In the case of calls, where the holder acquires the right to buy at a fixed price, the value of the call will decline as the strike price increases. In the case of puts, where the holder has the right to sell at a fixed price, the value will increase as the strike price increases. 5. Time to expiration on the option. Both calls and puts are more valuable the greater the time to expiration. This is because the longer time to expiration provides more time for the value of the underlying asset to move, increasing the value of both types of options. Additionally, in the case of a call, where the buyer has the right to pay a fixed price at expiration, the present value of this fixed price decreases as the life of the option increases, increasing the value of the call. 6. Riskless interest rate corresponding to life of the option. Since the buyer of an option pays the price of the option up front, an opportunity cost is involved. This cost will depend on the level of interest rates and the time to expiration of the option. The riskless interest rate also enters into the valuation of options when the present value of the exercise price is calculated, since the exercise price does not have to be paid (received) until expiration on calls (puts). Increases in the interest rate will increase the value of calls and reduce the value of puts. Table 5.1 summarizes the variables and their predicted effects on call and put prices. Table 5.1 Summary of Variables Affecting Call and Put Prices Effect On Factor
Call Value Put Value
Increase in underlying asset’s value
Increases
Decreases
Increase in variance of underlying asset Increases
Increases
Increase in strike price
Decreases Increases
Increase in dividends paid
Decreases Increases
75
Increase in time to expiration
Increases
Increases
Increase in interest rates
Increases
Decreases
American versus European Options: Variables Relating to Early Exercise A primary distinction between American and European options is that an American option can be exercised at any time prior to its expiration, while European options can be exercised only at expiration. The possibility of early exercise makes American options more valuable than otherwise similar European options; it also makes them more difficult to value. There is one compensating factor that enables the former to be valued using models designed for the latter. In most cases, the time premium associated with the remaining life of an option and transaction costs make early exercise suboptimal. In other words, the holders of in-the-money options generally get much more by selling the options to someone else than by exercising the options.
OPTION PRICING MODELS Option pricing theory has made vast strides since 1972, when Fischer Black and Myron Scholes published their pathbreaking paper that provided a model for valuing dividend-protected European options. Black and Scholes used a “replicating portfolio”—a portfolio composed of the underlying asset and the risk-free asset that h ad the same cash flows as the option being valued—and the notion of arbitrage to come up with their final formulation. Although their derivation is mathematically complicated, there is a simpler binomial model for valuing options that draws on the same logic.
Binomial Model The binomial option pricing model is based on a simple formulation for the asset price process in which the asset, in any time period, can move to one of two possible prices. The general formulation of a stock price process that follows the binomial path is shown in Figure 5.3. In this figure, S is the current stock price; the price moves up to Su with probability p and down to Sd with probability 1 – p in any time period. Figure 5.3 General Formulation for Binomial Price Path
Creating a Replicating Portfolio T he objective in creating a replicating portfolio is to use a combination of risk-free borrowing/lending and the underlying asset to create the same cash flows as the option being valued. The principles of arbitrage apply then, and the value of the option must be equal to the value of the replicating portfolio. In the case of the general formulation shown in Figure 5.3, where stock prices can move either up to Su or down to Sd in any time period, the replicating portfolio for a call with strike price K will involve borrowing $B and acquiring Δ of the underlying asset, where:
76
where Cu = Value of the call if the stock price is Su Cd = Value of the call if the stock price is Sd In a multiperiod binomial process, the valuation has to proceed iteratively (i.e., starting with the final time period and moving backward in time until the current point in time). The portfolios replicating the option are created at each step and valued, providing the values for the option in that time period. The final output from the binomial option pricing model is a statement of the value of the option in terms of the replicating portfolio, composed of Δ shares (option delta) of the underlying asset and risk-free borrowing/lending.
ILLUSTRATION 5.1: Binomial Option Valuation Assume that the objective is to value a call with a strike price of $50, which is expected to expire in two time periods, on an underlying asset whose price currently is $50 and is expected to follow a binomial process:
Now assume that the interest rate is 11%. In addition, define:
The objective is to combined Δ shares of stock and B dollars of borrowing to replicate the cash flows from the call with a strike price of $50. This can be done iteratively, starting with the last period and working back through the binomial tree. Step 1: Start with the end nodes and work backward:
Thus, if the stock price is $70 at t = 1, borrowing $45 and buying one share of the stock will give the same cash flows as buying the call. The value of the call at t = 1, if the stock price is $70, is therefore:
Considering the other leg of the binomial tree at t = 1,
77
If the stock price is $35 at t = 1, then the call is worth nothing. Step 2: Move backward to the earlier time period and create a replicating portfolio that will provide the cash flows the option will provide.
In other words, borrowing $22.50 and buying five-sevenths of a share will provide the same cash flows as a call with a strike price of $50 over the call’s lifetime. The value of the call therefore has to be the same as the cost of creating this position.
The Determinants of Value The binomial model provides insight into the determinants of option value. The value of an option is not determined by the expected price of the asset but by its current price, which, of course, reflects expectations about the future. This is a direct consequence of arbitrage. If the option value deviates from the value of the replicating portfolio, investors can create an arbitrage position (i.e., one that requires no investment, involves no risk, and delivers positive returns). To illustrate, if the portfolio that replicates the call costs more than the call does in the market, an investor could buy the call, sell the replicating portfolio, and be guaranteed the difference as a profit. The cash flows on the two positions will offset each other, leading to no cash flows in subsequent periods. The call option value also increases as the time to expiration is extended, as the price movements (u and d) increase, and with increases in the interest rate. While the binomial model provides an intuitive feel for the determinants of option value, it requires a large number of inputs, in terms of expected future prices at each node. As time periods are made shorter in the binomial model, you can make one of two assumptions about asset prices.You can assume that price changes become smaller as periods get shorter; this leads to price changes becoming infinitesimally small as time periods approach zero, leading to a continuous price process. Alternatively, you can assume that price changes stay large even as the period gets shorter; this leads to a jump price process, where prices can jump in any period. This section considers the option pricing models that emerge with each of these assumptions.
Black-Scholes Model When the price process is continuous (i.e., price changes become smaller as time periods get shorter), the binomial model for pricing options converges on the Black-Scholes model. The model, named after its cocreators, Fischer Black and Myron Scholes, allows us to estimate the value of any option using a small number of inputs, and has been shown to be robust in valuing many listed options.
The Model 78
While the derivation of the Black-Scholes model is far too complicated to present here, it is based on the idea of creating a portfolio of the underlying asset and the riskless asset with the same cash flows, and hence the same cost, as the option being valued. The value of a call option in the Black-Scholes model can be written as a function of the five variables: S = Current value of the underlying asset K = Strike price of the option t = Life to expiration of the option r = Riskless interest rate corresponding to the life of the option σ2 = Variance in the ln(value) of the underlying asset The value of a call is then:
Note that e–rt is the present value factor, and reflects the fact that the exercise price on the call option does not have to be paid until expiration, since the model values European options. N(d1) and N(d2) are probabilities, estimated by using a cumulative standardized normal distribution, and the values of d1 and d2 obtained for an option. The cumulative distribution is shown in Figure 5.4. Figure 5.4 Cumulative Normal Distribution
In approximate terms, N(d 2) yields the likelihood that an option will generate positive cash flows for its owner at exercise (i.e., that S > K in the case of a call option and that K > S in the case of a put option). The portfolio that replicates the call option is created by buying N(d1) units of the underlying asset, and borrowing Ke–rt N(d2). The portfolio will have the same cash flows as the call option, and thus the same value as the option. N(d1), which is the number of units of the underlying asset that are needed to create the replicating portfolio, is called the option delta.
A NOTE ON ESTIMATING THE INPUTS TO THE BLACK-SCHOLES MODEL The Black-Scholes model requires inputs that are consistent on time measurement. There are two places where this affects estimates. The first relates to the fact that the model works in continuous time, rather than discrete time. That is why we use the continuous time version of present value (exp–rt) rather than the discrete version, (1 + r)–t. It also means that the inputs such as the riskless rate have to be modified to make them continuous time inputs. For instance, if the one-year Treasury bond rate is 6.2 percent, the risk-free rate that is used in the BlackScholes model should be:
The second relates to the period over which the inputs are estimated. For instance, the preceding rate is an annual rate. The variance that is entered into the model also has to be an annualized variance. The variance, estimated from ln(asset prices), can be annualized easily because variances are linear in time if the serial correlation is zero. Thus, if monthly or weekly prices are used to estimate variance, the variance is annualized by multiplying by 12 or 52, respectively.
79
ILLUSTRATION 5.2: Valuing an Option Using the Black-Scholes Model On March 6, 2001, Cisco Systems was trading at $13.62. We will attempt to value a July 2001 call option with a strike price of $15, trading on the CBOE on the same day for $2. The following are the other parameters of the options: The annualized standard deviation in Cisco Systems stock price over the previous year was 81%. This standard deviation is estimated using weekly stock prices over the year, and the resulting number was annualized as follows:
The option expiration date is Friday, July 20, 2001. There are 103 days to expiration, and the annualized Treasury bill rate corresponding to this option life is 4.63%. The inputs for the Black-Scholes model are as follows: Current stock price (S) = $13.62 Strike price on the option = $15 Option life = 103/365 = 0.2822 Standard deviation in ln(stock prices) = 81% Riskless rate = 4.63% Inputting these numbers into the model, we get:
Using the normal distribution, we can estimate the N(d1) and N(d2):
The value of the call can now be estimated:
Since the call is trading at $2, it is slightly overvalued, assuming that the estimate of standard deviation used is correct.
IMPLIED VOLATILITY The only input in the Black Scholes on which there can be significant disagreement among investors is the variance. While the variance is often estimated by looking at historical data, the values for options that emerge from using the historical variance can be different from the market prices. For any option, there is some variance at which the estimated value will be equal to the market price. This variance is called an implied variance. Consider the Cisco option valued in Illustration 5.2. With a standard deviation of 81 percent, the value of the call option with a strike price of $15 was estimated to be $1.87. Since the market price is higher than the calculated value, we tried higher standard deviations, and at a standard deviation 85.40 percent the value of the option is $2 (which is the market price). This is the implied standard deviation or implied volatility.
Model Limitations and Fixes The Black-Scholes model was designed to value European options that can be exercised only at maturity and whose underlying assets do not pay dividends. In addition, options are valued based on the assumption that option exercise does not affect the value of the underlying asset. In practice, assets do pay dividends, options sometimes get exercised early, and exercising an option can affect the value of the underlying asset. Adjustments exist that, while not perfect, provide partial corrections to the Black-Scholes model.
Dividends The payment of a dividend reduces the stock price; note that on the ex-dividend day, the stock price generally 80
declines. Consequently, call options become less valuable and put options more valuable as expected dividend payments increase. There are two ways of dealing with dividends in the Black-Scholes model: 1. Short-term options. One approach to dealing with dividends is to estimate the present value of expected dividends that will be paid by the underlying asset during the option life and subtract it from the current value of the asset to use as S in the model.
2. Long-term options. Since it becomes less practical to estimate the present value of dividends the longer the option life, an alternate approach can be used. If the dividend yield (y = Dividends/Current value of the asset) on the underlying asset is expected to remain unchanged during the life of the option, the Black-Scholes model can be modified to take dividends into account.
From an intuitive standpoint, the adjustments have two effects. First, the value of the asset is discounted back to the present at the dividend yield to take into account the expected drop in asset value resulting from dividend payments. Second, the interest rate is offset by the dividend yield to reflect the lower carrying cost from holding the asset (in the replicating portfolio). The net effect will be a reduction in the value of calls estimated using this model.
ILLUSTRATION 5.3: Valuing a Short-Term Option with Dividend Adjustments—The Black-Scholes Correction Assume that it is March 6, 2001, and that AT&T is trading at $20.50 a share. Consider a call option on the stock with a strike price of $20, expiring on July 20, 2001. Using past stock prices, the annualized standard deviation in the log of stock prices for AT&T is estimated at 60%. There is one dividend, amounting to $0.15, and it will be paid in 23 days. The riskless rate is 4.63%. Present value of expected dividend = $0.15/1.046323/365 = $0.15 Dividend-adjusted stock price = $20.50 – $0.15 = $20.35 Time to expiration = 103/365 = 0.2822 Variance in ln(stock prices) = 0.62 = 0.36 Riskless rate = 4.63% The value from the Black-Scholes model is:
The call option was trading at $2.60 on that day.
ILLUSTRATION 5.4: Valuing a Long-Term Option with Dividend Adjustments— Primes and Scores The CBOE offers longer-term call and put options on some stocks. On March 6, 2001, for instance, you could have purchased an AT&T call expiring on January 17, 2003. The stock price for AT&T is $20.50 (as in the previous example). The following is the valuation of a call option with a strike price of $20. Instead of estimating the present value of dividends over the next two years, assume that AT&T’s dividend yield will remain 2.51% over this period and that the risk-free rate for a two-year Treasury bond is 4.85%. The inputs to the Black-Scholes model are:
81
The value from the Black-Scholes model is:
The call was trading at $5.80 on March 8, 2001.
optst.xls: This spreadsheet allows you to estimate the value of a short-term option when the expected dividends during the option life can be estimated.
optlt.xls: This spreadsheet allows you to estimate the value of an option when the underlying asset has a constant dividend yield.
Early Exercise The Black-Scholes model was designed to value European options that can be exercised only at expiration. In contrast, most options that we encounter in practice are American options and can be exercised at any time until expiration. As mentioned earlier, the possibility of early exercise makes American options more valuable than otherwise similar European options; it also makes them more difficult to value. In general, though, with traded options, it is almost always better to sell the option to someone else rather than exercise early, since options have a time premium (i.e., they sell for more than their exercise value). There are two exceptions. One occurs when the underlying asset pays large dividends, thus reducing the expected value of the asset. In this case, call options may be exercised just before an ex-dividend date, if the time premium on the options is less than the expected decline in asset value as a consequence of the dividend payment. The other exception arises when an investor holds both the underlying asset and deep in-the-money puts (i.e., puts with strike prices well above the current price of the underlying asset) on that asset at a time when interest rates are high. In this case, the time premium on the put may be less than the potential gain from exercising the put early and earning interest on the exercise price. There are two basic ways of dealing with the possibility of early exercise. One is to continue to use the unadjusted Black-Scholes model and to regard the resulting value as a floor or conservative estimate of the true value. The other is to try to adjust the value of the option for the possibility of early exercise. There are two approaches for doing so. One uses the Black-Scholes model to value the option to each potential exercise date. With options on stocks, this basically requires that the investor values options to each ex-dividend day and chooses the maximum of the estimated call values. The second approach is to use a modified version of the binomial model to consider the possibility of early exercise. In this version, the up and the down movements for asset prices in each period can be estimated from the variance. 2
Approach 1: Pseudo-American Valuation Step 1: Define when dividends will be paid and how much the dividends will be. Step 2: V alue the call option to each ex-dividend date using the dividend-adjusted approach described earlier, where the stock price is reduced by the present value of expected dividends. Step 3: Choose the maximum of the call values estimated for each ex-dividend day.
82
ILLUSTRATION 5.5: Using Pseudo-American Option Valuation to Adjust for Early Exercise Consider an option with a strike price of $35 on a stock trading at $40. The variance in the ln(stock prices) is 0.05, and the riskless rate is 4%. The option has a remaining life of eight months, and there are three dividends expected during this period:
Expected Dividend Ex-Dividend Day $0.80
In 1 month
$0.80
In 4 months
$0.80
In 7 months
The call option is first valued to just before the first ex-dividend date:
The value from the Black-Scholes model is:
The call option is then valued to before the second ex-dividend date:
The value of the call based on these parameters is:
The call option is then valued to before the third ex-dividend date:
The value of the call based on these parameters is:
The call option is then valued to expiration:
The value of the call based on these parameters is:
Approach 2: Using the Binomial Model The binomial model is much more capable of handling early exercise because it considers the cash flows at each time period, rather than just at expiration. The biggest limitation of the binomial model is determining what stock prices will be at the end of each period, but this can be overcome by using a variant that allows us to estimate the up and the down movements in stock prices from the estimated variance. There are four steps involved:
Step 1: If the variance in ln(stock prices) has been estimated for the Black-Scholes valuation, convert these into inputs for the binomial model:
where u and d are the up and the down movements per unit time for the binomial, and dt is the number of periods within each year (or unit time). 83
Step 2: Specify the period in which the dividends will be paid and make the assumption that the price will drop by the amount of the dividend in that period. Step 3: Value the call at each node of the tree, allowing for the possibility of early exercise just before ex-dividend dates. There will be early exercise if the remaining time premium on the option is less than the expected drop in option value as a consequence of the dividend payment. Step 4: Value the call at time 0, using the standard binomial approach.
bstobin.xls: This spreadsheet allows you to estimate the parameters for a binomial model from the inputs to a Black-Scholes model.
Impact of Exercise on Underlying Asset Value The Black-Scholes model is based on the assumption that exercising an option does not affect the value of the underlying asset. This may be true for listed options on stocks, but it is not true for some types of options. For instance, the exercise of warrants increases the number of shares outstanding and brings fresh cash into the firm, both of which will affect the stock price.3 The expected negative impact (dilution) of exercise will decrease the value of warrants, compared to otherwise similar call options. The adjustment for dilution to the stock price is fairly simple in the Black-Scholes valuation. The stock price is adjusted for the expected dilution from the exercise of the options. In the case of warrants, for instance:
where S = Current value of the stock nw = Number of warrants outstanding W = Value of warrants outstanding ns = Number of shares outstanding When the warrants are exercised, the number of shares outstanding will increase, reducing the stock price. The numerator reflects the market value of equity, including both stocks and warrants outstanding. The reduction in S will reduce the value of the call option. There is an element of circularity in this analysis, since the value of the warrant is needed to estimate the dilutionadjusted S and the dilution-adjusted S is needed to estimate the value of the warrant. This problem can be resolved by starting the process off with an assumed value for the warrant (e.g., the exercise value or the current market price of the warrant). This will yield a value for the warrant, and this estimated value can then be used as an input to reestimate the warrant’s value until there is convergence.
FROM BLACK-SCHOLES TO BINOMIAL The process of converting the continuous variance in a Black-Scholes model to a binomial tree is a fairly simple one. Assume, for instance, that you have an asset that is trading at $30 currently and that you estimate the annualized standard deviation in the asset value to be 40 percent; the annualized riskless rate is 5 percent. For simplicity, let us assume that the option that you are valuing has a four-year life and that each period is a year. To estimate the prices at the end of each of the four years, we begin by first estimating the up and down movements in the binomial:
Based on these estimates, we can obtain the prices at the end of the first node of the tree (the end of the first year):
84
Progressing through the rest of the tree, we obtain the following numbers:
ILLUSTRATION 5.6: Valuing a Warrant on Avatek Corporation Avatek Corporation is a real estate firm with 19.637 million shares outstanding, trading at $0.38 a share. In March 2001 the company had 1.8 million options outstanding, with four years to expiration and with an exercise price of $2.25. The stock paid no dividends, and the standard deviation in ln(stock prices) was 93%. The four-year Treasury bond rate was 4.9%. (The options were trading at $0.12 apiece at the time of this analysis.) The inputs to the warrant valuation model are as follows:
The results of the Black-Scholes valuation of this option are:
The options were trading at $0.12 in March 2001. Since the value was equal to the price, there was no need for further iterations. If there had been a difference, we would have reestimated the adjusted stock price and option value. If the options had been non-traded (as is the case with management options), this calculation would have required an iterative process, where the option value is used to get the adjusted value per share and the value per share to get the option value.
warrant.xls: This spreadsheet allows you to estimate the value of an option when there is a potential dilution from exercise.
The Black-Scholes Model for Valuing Puts The value of a put can be derived from the value of a call with the same strike price and the same expiration date:
85
where C is the value of the call and P is the value of the put. This relationship between the call and put values is called put-call parity, and any deviations from parity can be used by investors to make riskless profits. To see why put-call parity holds, consider selling a call and buying a put with exercise price K and expiration date t, and simultaneously buying the underlying asset at the current price S. The payoff from this position is riskless and always yields K at expiration (t). To see this, assume that the stock price at expiration is S*. The payoff on each of the positions in the portfolio can be written as follows: Position
Payoffs at t if S* > K Payoffs at t if S* < K
Sell call
–(S* – K)
Buy put
K – S*
Buy stock S*
S*
Total
K
K
Since this position yields K with certainty, the cost of creating this position must be equal to the present value of K at the riskless rate (K e–rt).
Substituting the Black-Scholes equation for the value of an equivalent call into this equation, we get:
Thus, the replicating portfolio for a put is created by selling short [1 – N(d1)] shares of stock and investing K e–rt[1 – N(d2)] in the riskless asset.
ILLUSTRATION 5.7: Valuing a Put Using Put-Call Parity: Cisco Systems and AT&T Consider the call on Cisco Systems that we valued in Illustration 5.2. The call had a strike price of $15 on the stock, had 103 days left to expiration, and was valued at $1.87. The stock was trading at $13.62, and the riskless rate was 4.63%. The put can be valued as follows:
The put was trading at $3.38. Also, a long-term call on AT&T was valued in Illustration 5.4. The call had a strike price of $20, 1.8333 years left to expiration, and a value of $6.63. The stock was trading at $20.50 and was expected to maintain a dividend yield of 2.51% over the period. The riskless rate was 4.85%. The put value can be estimated as follows:
The put was trading at $3.80. Both the call and the put were trading at different prices from our estimates, which may indicate that we have not correctly estimated the stock’s volatility.
Jump Process Option Pricing Models If price changes remain larger as the time periods in the binomial model are shortened, it can no longer be assumed that prices change continuously. When price changes remain large, a price process that allows for price jumps is much more realistic. Cox and Ross (1976) valued options when prices follow a pure jump process, where the jumps can only be positive. Thus, in the next interval, the stock price will either have a large positive jump with a specified probability or drift downward at a given rate. Merton (1976) considered a distribution where there are price jumps superimposed on a continuous price process. He specified the rate at which jumps occur (λ) and the average jump size (k), measured as a percentage of the stock price. The model derived to value options with this process is called a jump diffusion model. In this model, the value of an option is determined by the five variables specified in the Black-Scholes model, and the parameters of the jump process (λ, k). Unfortunately, the estimates of the jump process parameters are so difficult to make for most firms that they overwhelm any advantages that accrue from using a more realistic model. These models, therefore, have seen limited use in practice.
86
EXTENSIONS OF OPTION PRICING All the option pricing models described so far—the binomial, the Black-Scholes, and the jump process models—are designed to value options with clearly defined exercise prices and maturities on underlying assets that are traded. However, the options we encounter in investment analysis or valuation are often on real assets rather than financial assets. Categorized as real options, they can take much more complicated forms. This section considers some of these variations.
Capped and Barrier Options With a simple call option, there is no specified upper limit on the profits that can be made by the buyer of the call. Asset prices, at least in theory, can keep going up, and the payoffs increase proportionately. In some call options, though, the buyer is entitled to profits up to a specified price but not above it. For instance, consider a call option with a strike price of K1 on an asset. In an unrestricted call option, the payoff on this option will increase as the underlying asset’s price increases above K1. Assume, however, that if the price reaches K2, the payoff is capped at (K2 – K1). The payoff diagram on this option is shown in Figure 5.5. Figure 5.5 Payoff on Capped Call
This option is called a capped call. Notice, also, that once the price reaches K 2, there is no longer any time premium associated with the option, and the option will therefore be exercised. Capped calls are part of a family of options called barrier options, where the payoff on and the life of the option are functions of whether the underlying asset price reaches a certain level during a specified period. The value of a capped call is always lower than the value of the same call without the payoff limit. A simple approximation of this value can be obtained by valuing the call twice, once with the given exercise price and once with the cap, and taking the difference in the two values. In the preceding example, then, the value of the call with an exercise price of K1 and a cap at K2 can be written as:
Barrier options can take many forms. In a knockout option, an option ceases to exist if the underlying asset reaches a certain price. In the case of a call option, this knockout price is usually set below the strike price, and this option is called a down-and-out option. In the case of a put option, the knockout price will be set above the exercise price, and this option is called an up-and-out option. Like the capped call, these options are worth less than their unrestricted counterparts. Many real options have limits on potential upside, or knockout provisions, and ignoring these limits can result in the overstatement of the value of these options.
Compound Options Some options derive their value not from an underlying asset, but from other options. These options are called compound options. Compound options can take any of four forms—a call on a call, a put on a put, a call on a put, or a put on a call. Geske (1979) developed the analytical formulation for valuing compound options by replacing the standard normal distribution used in a simple option model with a bivariate normal distribution in the calculation. Consider, for instance, the option to expand a project that is discussed in Chapter 30. While we will value this option using a simple option pricing model, in reality there could be multiple stages in expansion, with each stage representing an option for the following stage. In this case, we will undervalue the option by considering it as a simple rather than a compound option. Notwithstanding this discussion, the valuation of compound options becomes progressively more difficult as more options are added to the chain. In this case, rather than wreck the valuation on the shoals of estimation error, it may be better to accept the conservative estimate that is provided with a simple valuation model as a floor on the value.
Rainbow Options 87
In a simple option, the uncertainty is about the price of the underlying asset. Some options are exposed to two or more sources of uncertainty, and these options are rainbow options. Using the simple option pricing model to value such options can lead to biased estimates of value. As an example, consider an undeveloped oil reserve as an option, where the firm that owns the reserve has the right to develop the reserve. Here there are two sources of uncertainty. The first is obviously the price of oil, and the second is the quantity of oil that is in the reserve. To value this undeveloped reserve, we can make the simplifying assumption that we know the quantity of oil in the reserve with certainty. In reality, however, uncertainty about the quantity will affect the value of this option and make the decision to exercise more difficult.4
CONCLUSION An option is an asset with payoffs that are contingent on the value of an underlying asset. A call option provides its holder with the right to buy the underlying asset at a fixed price, whereas a put option provides its holder with the right to sell at a fixed price, at any time b efore the expiration of the option. The value of an option is determined by six variables—the current value of the underlying asset, the variance in this value, the expected dividends on the asset, the strike price and life of the option, and the riskless interest rate. This is illustrated in both the binomial and the Black-Scholes models, which value options by creating replicating portfolios composed of the underlying asset and riskless lending or borrowing. These models can be used to value assets that have option like characteristics.
QUESTIONS AND SHORT PROBLEMS In the problems following, use an equity risk premium of 5.5 percent if none is specified. 1. The following are prices of options traded on Microsoft Corporation, which pays no dividends.
The stock is trading at $83, and the annualized riskless rate is 3.8%. The standard deviation in ln(stock prices) (based on historical data) is 30%. a. Estimate the value of a three-month call with a strike price of $85. b. Using the inputs from the Black-Scholes model, specify how you would replicate this call. c. What is the implied standard deviation in this call? d. Assume now that you buy a call with a strike price of $85 and sell a call with a strike price of $90. Draw the payoff diagram on this position. e. Using put-call parity, estimate the value of a three-month put with a strike price of $85. 2. You are trying to value three-month call and put options on Merck with a strike price of $30. The stock is trading at $28.75, and the company expects to pay a quarterly dividend per share of $0.28 in two months. The annualized riskless interest rate is 3.6%, and the standard deviation in log stock prices is 20%. a. Estimate the value of the call and put options, using the Black-Scholes model. b. What effect does the expected dividend payment have on call values? On put values? Why? 3. There is the possibility that the options on Merck described in the preceding problem could be exercised early. a. Use the pseudo-American call option technique to determine whether this will affect the value of the call. b. Why does the possibility of early exercise exist? What types of options are most likely to be exercised early? 4. You have been provided the following information on a three-month call:
88
a. If you wanted to replicate buying this call, how much money would you need to borrow? b. If you wanted to replicate buying this call, how many shares of stock would you need to buy? 5. Go Video, a manufacturer of video recorders, was trading at $4 per share in May 1994. There were 11 million shares outstanding. At the same time, it had 550,000 one-year warrants outstanding, with a strike price of $4.25. The stock has had a standard deviation of 60%. The stock does not pay a dividend. The riskless rate is 5%. a. Estimate the value of the warrants, ignoring dilution. b. Estimate the value of the warrants, allowing for dilution. c. Why does dilution reduce the value of the warrants? 6. You are trying to value a long-term call option on the NYSE Composite index, expiring in five years, with a strike price of 275. The index is currently at 250, and the annualized standard deviation in stock prices is 15%. The average dividend yield on the index is 3% and is expected to remain unchanged over the next five years. The fiveyear Treasury bond rate is 5%. a. Estimate the value of the long-term call option. b. Estimate the value of a put option with the same parameters. c. What are the implicit assumptions you are making when you use the Black-Scholes model to value this option? Which of these assumptions are likely to be violated? What are the consequences for your valuation? 7. A new security on AT&T will entitle the investor to all dividends on AT&T over the next three years, limiting upside potential to 20% but also providing downside protection below 10%. AT&T stock is trading at $50, and three-year call and put options are traded on the exchange at the following prices:
How much would you be willing to pay for this security? Note, though, that higher variance can reduce the value of the underlying asset. As a call option becomes more in-the-money, the more it resembles the underlying asset. For very deep in-the-money call options, higher variance can reduce the value of the option. 1
To illustrate, if σ2 is the variance in ln(stock prices), the up and the down movements in the binomial can be estimated as follows: 2
where u and d are the up and down movements per unit time for the binomial, T is the life of the option , and m is the number of periods within that lifetime. Warrants are call options issued by firms, either as part of management compensation contracts or to raise equity. 3
The analogy to a listed option on a stock is the case where you do not know with certainty what the stock price is when you exercise the option. The more uncertain you are about the stock price, the more margin for error you have to give yourself when you exercise the option, to ensure that you are in fact earning a profit. 4
89
CHAPTER 6 Market Efficiency—Definition, Tests, and Evidence What is an efficient market? What does it imply for investment and valuation models? Clearly, market efficiency is a concept that is controversial and attracts strong views, pro and con, partly because of differences between individual s about what it really means, and partly because whether markets are efficient or not is a core belief that in large part determines how an investor approaches investing. This chapter provides a definition of market efficiency, considers the implications of an efficient market for investors, and summarizes some of the basic approaches that are used to test investment schemes, thereby proving or disproving market efficiency. It also provides a summary of the voluminous research on whether markets are efficient.
MARKET EFFICIENCY AND INVESTMENT VALUATION The question of whether markets are efficient, and, if not, where the inefficiencies lie, is central to investment valuation. If markets are in fact efficient, the market price provides the best estimate of value, and the process of valuation becomes one of justifying the market price. If markets are not efficient, the market price may deviate from the true value, an d the process of valuation is directed toward obtaining a reasonable estimate of this value. Those who do valuation well, then, will be able to make higher returns than other investors because of their capacity to spot under- and overvalued firms. To make these higher returns, though, markets have to correct their mistakes (i.e., become efficient) over time. Whether these corrections occur over six months or over five years can have a profound impact on which valuation approach an investor chooses to use and the time horizon that is needed for it to succeed. There is also much that can be learned from studies of market efficiency, which highlight segments where the market seems to be inefficient. These inefficiencies can provide the basis for screening all stocks to come up with a subsample that is more likely to contain undervalued stocks. Given the size of the universe of stocks, this not only saves time for the analyst, but it increases the odds significantly of finding under- and overvalued stocks. For instance, some efficiency studies suggest that stocks that are neglected by institutional investors are more likely to be undervalued and earn excess returns. A strategy that screens firms for low institutional investment (as a percentage of the outstanding stock) may yield a subsample of neglected firms, which can then be valued to arrive at a portfolio of undervalued firms. If the research is correct, the odds of finding undervalued firms should increase in this subsample.
WHAT IS AN EFFICIENT MARKET? An efficient market is one where the market price is an unbiased estimate of the true value of the investment. Implicit in this derivation are several key concepts: Contrary to popular view, market efficiency does not require that the market price be equal to true value at every point in time. All it requires is that errors in the market price be unbiased; prices can be greater than or less than true value, as long as these deviations are random. The fact that the deviations from true value are random implies, in a rough sense, that there is an equal chance that any stock is under- or overvalued at any point in time, and that these deviations are uncorrelated with any observable variable. For instance, in an efficient market, stocks with lower PE ratios should be no more or no less likely to be undervalued than stocks with high PE ratios. If the deviations of market price from true value are random, it follows that no group of investors should be able to consistently find under- or overvalued stocks using any investment strategy. Definitions of market efficiency have to be specific not only about the market that is being considered but also the investor group that is covered. It is extremely unlikely that all markets are efficient to all investors at all times, but it is entirely possible that a particular market (for instance, the New York Stock Exchange) is efficient with respect to the average investor. It is also possible that some markets are efficient while others are not, and that a market is efficient with respect to some investors and not to others. This is a direct consequence of differential tax rates and transaction costs, which confer advantages on some investors relative to others. Definitions of market efficiency are also linked up with assumptions about what information is available to investors and reflected in the price. For instance, a strict definition of market efficiency that assumes that all information, public as well as private, is reflected in market prices would imply that even investors with precise inside information will be unable to beat the market. One of the earliest classifications of levels of market efficiency was provided by Fama (1971), who argued that markets could be efficient at three levels, based on what information was reflected in prices. Under weak form efficiency, the current price reflects the information contained in all past prices, suggesting that charts and technical analyses that use past prices alone would not be useful in finding undervalued stocks. Under semi-strong form efficiency, the current price reflects the information contained not only in past prices but all public information (including financial statements and news reports) and no approach that is 90
predicated on using and massaging this information would be useful in finding undervalued stocks. Under strong form efficiency, the current price reflects all information, public as well as private, and no investors will be able to find undervalued stocks consistently.
IMPLICATIONS OF MARKET EFFICIENCY An immediate and direct implication of an efficient market is that no group of investors should be able to beat the market consistently using a common investment strategy. An efficient market would also carry negative implications for many investment strategies: In an efficient market, equity research and valuation would be a costly task that would provide no benefits. The odds of finding an undervalued stock would always be 50–50, reflecting the randomness of pricing errors. At best, the benefits from information collection and equity research would cover the costs of doing the research. In an efficient market, a strategy of randomly diversifying across stocks or indexing to the market, carrying little or no information cost and minimal execution costs, would be superior to any other strategy that created larger information and execution costs. There would be no value added by active portfolio managers and investment strategists. In an efficient market, a strategy of minimizing trading (i.e., creating a portfolio and not trading unless cash was needed) would be superior to a strategy that required frequent trading. It is therefore no wonder that the concept of market efficiency evokes such strong reactions on the part of portfolio managers and analysts, who view it, quite rightly, as a challenge to their existence. It is also important that there be clarity about what market efficiency does not imply. An efficient market does not imply that: Stock prices cannot deviate from true value; in fact, there can be large deviations from true value. The only requirement is that the deviations be random. No investor will beat the market in any time period. To the contrary, approximately half of all investors, prior to transaction costs, should beat the market in any period.1 No group of investors will beat the market in the long term. Given the number of investors in financial markets, the laws of probability would suggest that a fairly large number are going to beat the market consistently over long periods, not because of their investment strategies but because they are lucky. It would not, however, be consistent if a disproportionately large number 2 of these investors used the same investment strategy. In an efficient market, the expected returns from any investment will be consistent with the risk of that investment over the long term, though there may be deviations from these expected returns in the short term.
NECESSARY CONDITIONS FOR MARKET EFFI CIENCY Markets do not become efficient automatically. It is the actions of investors, sensing bargains and putting into effect schemes to beat the market, that make markets efficient. The necessary conditions for a market inefficiency to be eliminated are: The market inefficiency should provide the basis for a scheme to beat the market and earn excess returns. For this to hold true: The asset or assets that are the source of the inefficiency have to be traded. The transaction costs of executing the scheme have to be smaller than the expected profits from the scheme. There should be profit-maximizing investors who: Recognize the potential for excess return. Can replicate the beat-the-market scheme that earns the excess return. Have the resources to trade on the stock(s) until the inefficiency disappears. The internal contradiction of claiming that there is no possibility of beating the market in an efficient market and requiring profit-maximizing investors to constantly seek out ways of beating the market and thus making it efficient has been explored by many. If markets were in fact efficient, investors would stop looking for inefficiencies, which would lead to markets becoming inefficient again. It makes sense to think about an efficient market as a selfcorrecting mechanism, where inefficiencies appear at regular intervals but disappear almost instantaneously as investors find them and trade on them.
PROPOSITIONS ABOUT MARKET EFFICIENCY A reading of the conditions under which markets become efficient leads to general propositions about where 91
investors are most likely to find inefficiencies in financial markets. Proposition 1: The probability of finding inefficiencies in an asset market decreases as the ease of trading on the asset increases. To the extent that investors have difficulty trading on an asset, either because open markets do not exist or because there are significant barriers to trading, inefficiencies in pricing can continue for long periods. This proposition can be used to shed light on the differences between different asset markets. For instance, it is far easier to trade on stocks than it is on real estate, since markets are much more open, prices are in smaller units (reducing the barriers to entry for new traders), and the asset itself does not vary from transaction to transaction (one share of IBM is identical to another share, whereas one piece of real estate can be very different from another piece that is a stone’s throw away). Based on these differences, there should be a greater likelihood of finding inefficiencies (both under- and overvaluation) in the real estate market. Proposition 2: The probability of finding an inefficiency in an asset market increases as the transactions and information cost of exploiting the inefficiency increases. The cost of collecting information and trading varies widely across markets and even across investments in the same markets. As these costs increase, it pays less and less to try to exploit these inefficiencies. Consider, for instance, the perceived wisdom that investing in “loser” stocks (i.e., stocks that have done very badly in some prior time period) should yield excess returns. This may be true in terms of raw returns, but transaction costs are likely to be much higher for these stocks since: They tend to be low-priced stocks, leading to higher brokerage commissions and expenses. The bid-ask spread, a transaction cost paid at the time of purchase, becomes a much higher fraction of the total price paid. Trading is often thin on these stocks, and small trades can cause prices to change, resulting in a higher buy price or a lower sell price. Corollary 1: Investors who can establish a cost advantage (either in information collection or in transaction costs) will be more able to exploit small inefficiencies than other investors who do not possess this advantage. There are a number of studies that look at the effect of block trades on prices and conclude that while block trades do affect prices, investors will not exploit these inefficiencies because of the number of times they will have to trade and their associated transaction costs. These concerns are unlikely to hold for a specialist on the floor of the exchange, who can trade quickly, often and at no or very low costs. It should be pointed out, however, that if the market for specialists is efficient, the value of a seat on the exchange should reflect the present value of potential benefits from being a specialist. This corollary also suggests that investors who work at establishing a cost advantage, especially in relation to information, may be able to generate excess returns on the basis of these advantages. Thus John Templeton, who started investing in Japanese and other Asian markets well before other portfolio managers, might have been able to exploit the informational advantages he had over his peers to make excess returns on his portfolios, at least for a few years. Proposition 3: The speed with which an inefficiency is resolved will be directly related to how easily the scheme to exploit the inefficiency can be replicated by other investors. The ease with which a scheme can be replicated is related to the time, resources, and information needed to execute it. Since very few investors single-handedly possess the resources to eliminate an inefficiency through trading, it is much more likely that an inefficiency will disappear quickly if the scheme used to exploit the inefficiency is transparent and can be copied by other investors. To illustrate this point, assume that stocks are consistently found to earn excess returns in the month following a stock split. Since firms announce stock splits publicly and any investor can buy stocks right after these splits, it would be surprising if this inefficiency persisted over time. This can be contrasted with the excess returns made by some arbitrage funds in index arbitrage, where index futures are bought (or sold), and stocks in the index are sold short or (bought). This strategy requires that investors be able to obtain information on the index and spot prices instantaneously, have the capacity (in terms of margin requirements and resources) to trade index futures and to sell short on stocks, and to have the resources to take and hold very large positions until the arbitrage unwinds. Consequently, inefficiencies in index futures pricing are likely to persist at least for the most efficient arbitrageurs, with the lowest execution costs and the speediest execution times.
TESTING MARKET EFFICIENCY Tests of market efficiency look at the whether specific investment strategies earn excess returns. Some tests also account for transactions costs and execution feasibility. Since an excess return on an investment is the difference between the actual and expected return on that investment, there is implicit in every test of market efficiency a model for this expected return. In some cases, th is expected return adjusts for risk using the capital asset pricing model or the arbitrage pricing model, and in others the expected return is based on returns on similar or equivalent 92
investments. In every case, a test of market efficiency is a joint test of market efficiency and the efficacy of the model used for expected returns. When there is evidence of excess returns in a test of market efficiency, it can indicate that markets are inefficient or that the model used to compute expected returns is wrong (or both). Although this may seem to present an insoluble dilemma, if the conclusions of the study are insensitive to different model specifications, it is much more likely that the results are being driven by true market inefficiencies and not just by model misspecifications. There are a number of different ways of testing for market efficiency, and the approach used will depend in great part on the investment scheme being tested. A scheme based on trading on information events (stock splits, earnings announcements, or acquisition announcements) is likely to be tested using an “event study” in which returns around the event are scrutinized for evidence of excess returns. A scheme based on trading on an observable characteristic of a firm (price-earnings ratios, price–book value ratios, or dividend yields) is likely to be tested using a portfolio approach, where portfolios of stocks with these characteristics are created and tracked over time to see whether in fact they make excess returns. The following pages summarize the key steps involved in each of these approaches, and some potential pitfalls to watch out for when conducting or using these tests.
Event Study An event study is designed to examine market reactions to and excess returns around specific information events. The information events can be marketwide, such as macroeconomic announcements, or firm-specifc, such as earnings or dividend announcements. The five steps in an event study are: 1. The event to be studied is clearly identified, and the date on which the event was announced is pinpointed. The presumption in event studies is that the timing of the event is known with a fair degree of certainty. Since financial markets react to the information about an event rather than the event itself, most event studies are centered around the announcement date for the event.3
2. Once the event dates are known, returns are collected around these dates for each of the firms in the sample. In doing so, two decisions have to be made. First, the researcher has to decide whether t o collect weekly, daily, or shorter-interval returns around the event. This will be determined in part by how precisely the event date is known (the more precise, the more likely it is that shorter return intervals can be used) and by how quickly information is reflected in prices (the faster the adjustment, the shorter the return interval to use). Second, the analyst has to determine how many periods of returns before and after the announcement date will be considered as part of the event window. That decision also will be determined by the precision of the event date, since more imprecise dates will require longer windows.
where Rjt = Returns on firm j for period t(t = –n, ..., 0, ..., +n) 3. The returns, by period, around the announcement date, are adjusted for market performance and risk to arrive at excess returns for each firm in the sample. For instance, if the capital asset pricing model is used to control for risk:
where ERjt = Excess returns on firm j for period t(t = n, . . . , 0, . . . , +n) = Rjt E(Rjt) 4. The excess returns, by period, are averaged across all events in the sample, and a standard error is computed.
93
where N = Number of events (firms) in the event study 5. The question of whether the excess returns around the announcement are different from zero is answered by estimating the t statistic for each period, by dividing the average excess return by the standard error:
If the t statistics are statistically significant,4 the event affects returns; the sign of the excess return determines whether the effect is positive or negative.
ILLUSTRATION 6.1: Example of an Event Study—Effects of Option Listing on Stock Prices Academics and practitioners have long argued about the consequenc es of option listing for stock price volatility. On the one hand, there are those who argue that options attract speculators and hence increase stock price volatility. On the other hand, there are others who argue that options increase the available choices for investors and increase the flow of information to financial markets, and thus lead to lower stock price volatility and higher stock prices. One way to test these alternative hypotheses is to do an event study, examining the effects of listing options on the underlying stocks’ prices. Conrad (1989) did such a study, following these steps:
Step 1: The date of the announcement that options on a particular stock would be listed on the Chicago Board Options Exchange was collected. Step 2: The prices of the underlying stock (j) were collected for each of the 10 days prior to the option listing announcement date, for the day of the announcement, and for each of the 10 days after. Step 3: The returns on the stock (Rjt) were computed for each of these trading days. Step 4: The beta for the stock (βj) was estimated using the returns from a time period outside the event window (using 100 trading days from before the event and 100 trading days after the event). Step 5: The returns on the market index (Rmt) were computed for each of the 21 trading days. Step 6: The excess returns were computed for each of the 21 trading days:
The excess returns are cumulated for each trading day.
Step 7: The average and standard error of excess returns across all stocks with option listings were computed for each of the 21 trading days. The t statistics are computed using the averages and standard errors for each trading day. The following table summarizes the average excess returns and t statistics around option listing announcement dates:
94
Based on these excess returns, there is no evidence of an announcement effect on the announcement day alone, but there is mild evidence of a positive effect over the entire announcement period.5
Portfolio Study In some investment strategies, firms with specific characteristics are viewed as more likely to be undervalued, and therefore to have excess returns, than firms without these characteristics. In these cases, the strategies can be tested by creating portfolios of firms possessing these c haracteristics at the beginning of a time period and then examining returns over the time period. To ensure that these results are not colored by the idiosyncracies of one time period, this analysis is repeated for a number of periods. The seven steps in doing a portfolio study are: 1. The variable on which firms will be classified is defined, using the investment strategy as a guide. This variable has to be observable, though it does not have to be numerical. Examples would include market value of equity, bond ratings, stock prices, price-earnings ratios, and price–book value ratios. 2. The data on the variable is collected for every firm in the defined universe6 at the start of the testing period, and firms are classified into portfolios based on the magnitude of the variable. Thus, if the price-earnings ratio is the screening variable, firms are classified on the basis of PE ratios into portfolios from lowest PE to highest PE classes. The number of classes will depend on the size of the universe, since there have to be sufficient firms in each portfolio to get some measure of diversificati on. 3. The returns are collected for each firm in each portfolio for the testing period, and the returns for each portfolio are computed, making the decision to weight them either equally or based on value. 4. The beta (if using a single-factor model) or betas (if using a multifactor model) of each portfolio are estimated, either by taking the average of the betas of the individual stocks in the portfolio or by regressing the portfolio’s returns against market returns over a prior time period (for instance, the year before the testing period). If you want to control for any other variables that have been shown to be correlated with returns such as market capitalization or price to book ratio, they can be incorporated into the expected return as well. 5. The excess returns earned by each portfolio are computed, in conjunction with the standard error of the excess returns. 6. There are a number of statistical tests available to check whether the average excess returns are, in fact, different across the portfolios. Some of these tests are parametric7 (they make certain distributional assumptions about excess returns), and some are nonparametric.8 7. As a final test, the extreme portfolios can be matched against each other to see whether there are statistically significant differences across these portfolios.
ILLUSTRATION 6.2: Example of a Portfolio Study—Price-Earnings Ratios Practitioners have claimed that low price-earnings ratio stocks are generally bargains and do much better than the market or stocks with high price-earnings ratios. This hypothesis can be tested using a portfolio approach:
Step 1: Using data on price-earnings ratios from the end of 1987, firms on the New York Stock Exchange were classified into five groups, the first group consisting of stocks with the lowest PE ratios and the fifth group consisting of stocks with the highest PE ratios. Firms with negative price-earnings ratios were ignored (which may bias the results). Step 2: The returns on each portfolio were computed using data from 1988 to 1992. Stocks that went bankrupt or were delisted were assigned a return of –100%. Step 3: The betas for each stock in each portfolio were computed using monthly returns from 1983 to 1987, and the average beta for each portfolio was estimated. The portfolios were assumed to be equally weighted. Step 4: The returns on the market index were computed from 1988 to 1992. Step 5: The excess returns on each portfolio were computed from 1988 to 1992. The following table summarizes the excess returns each year from 1988 to 1992 for each portfolio.
Step 6: While the ranking of the returns across the portfolio classes seems to confirm our hypothesis that low-PE stocks earn a higher return, we have to consider whether the differences across portfolios are statistically significant. There are several tests available, but these are a few: An F test can be used to accept or reject the hypothesis that the average returns are the same across all portfolios. A high F score would lead us to conclude that the differences are too large to be random. A chi-squared test is a nonparametric test that can be used to test the hypothesis that the means are the same across the five portfolio
95
classes. We could isolate just the lowest-PE and highest-PE stocks and estimate a t statistic that the averages are different across these two portfolios.
CARDINAL SINS IN TESTING MARKET EFFICIENCY In the process of testing investment strategies, there are a number of pitfalls that have to be avoided. Six of them are: 1. Using anecdotal evidence to support/reject an investment strategy. Anecdotal evidence is a double-e dged sword. It can be used to support or reject the same hypothesis. Since stock prices are noisy and all investment schemes (no matter how absurd) will succeed sometimes and fail at other times, there will always be cases where the scheme works or does not work. 2. Testing an investment strategy on the same data and time period from which it was extracted. This is the tool of choice for the unscrupulous investment strategist. An investment scheme is extracted from hundreds through an examination of the data for a particular time period. This investment scheme is then tested on the same time period, with predictable results. (The scheme does miraculously well and makes immense returns.) An investment scheme should always be tested out on a time period different from the one it is extracted from or on a universe different from the one used to derive the scheme. 3. Choosing a biased sample. There may be bias in the sample on which the test is run. Since there are thousands of stocks that could be considered part of this universe, researchers often choose to use a smaller sample. When this choice is random, this does limited damage to the results of the study. If the choice is biased, it can provide results that are not true in the larger universe. 4. Failure to control for market performance. A failure to control for overall market performance can lead you to conclude that your investment scheme works just because it makes good returns (most schemes will make good returns if the overall market does well; the question is whether they made better returns than expected) or does not work just because it makes bad returns (most schemes will do badly if the overall market performs poorly). It is crucial therefore that investment schemes control for market performance during the period of the test. 5. Failure to control for risk. A failure to control for risk leads to a bias toward accepting high-risk investment schemes and rejecting low-risk investment schemes, since the former should make higher returns than the market and the latter lower, without implying any excess returns. 6. Mistaking correlation for causation. Consider the study on PE stocks cited in the earlier section. We concluded that low-PE stocks have higher excess returns than high-PE stocks. It would be a mistake to conclude that a low price-earnings ratio causes excess returns, since the high returns and the low PE ratio themselves might have been caused by the high risk associated with investing in the stock. In other words, high risk is the causative factor that leads to both the observed phenomena—low PE ratios on the one hand and high returns on the other. This insight would make us more cautious about adopting a strategy of buying low-PE stocks in the first place.
SOME LESSER SINS THAT CAN BE A PROBLEM 1. Survival bias. Most researchers start with an existing universe of publicly traded companies and work back through time to test investment strategies. This can create a subtle bias since it automatically eliminates firms that failed during the period, with obvious negative consequences for returns. If the investment scheme is particularly susceptible to picking firms that have high bankruptcy risk, this may lead to an overstatement of returns on the scheme. For example, assume that the investment scheme recommends investing in stocks that have very negative earnings, using the argument that these stocks are the most likely to benefit from a turnaround. Some of the firms in this portfolio will go bankrupt, and a failure to consider these firms will overstate the returns from this strategy. 2. Not allowing for transaction costs. Some investment schemes are more expensive than others because of transaction costs—execution fees, bid-ask spreads, and price impact. A complete test will take these into account before it passes judgment on the strategy. This is easier said than done, because different investors have different transaction costs, and it is unclear which investor’s trading cost schedule should be used in the test. Most researchers who ignore transaction costs argue that individual investors can decide for themselves, given their transaction costs, whether the excess returns justify the investment strategy. 3. Not allowing for difficulties in execution. Some strategies look good on paper but are difficult to execute in practice, either because of impediments to trading or because trading creates a price impact. Thus a strategy of investing in very small companies may seem to create excess returns on paper, but these excess returns may not exist in practice because the price impact is significant.
EVIDENCE ON MARKET EFFICIENCY 96
This section of the chapter attempts to summarize the evidence from studies of market efficiency. Without claiming to be comprehensive, the evidence is classified into four sections—the study of price changes and their time series properties, the research on the efficiency of market reaction to information announcements, the existence of return anomalies across firms and over time, and the analysis of the performa nce of insiders, analysts, and money managers.
TIME SERIES PROPERTIES OF PRICE CHANGES Investors have used price charts and price patterns as tools for predicting future price movements for as long as there have been financial markets. It is not surprising, therefore, that the first studies of market efficiency focused on the relationship between price changes over time, to see if in fact such predictions were feasible. Some of this testing was spurred by the random walk theory of price move ments, which contended that price changes over time followed a random walk. As the studies of the time series properties of prices have proliferated, the evidence can be classified into two categories—studies that focus on short-term price behavior (intraday, daily, and weekly price movements) price behavior and research that examines the longer term (monthly, annual, and multi-year).
Short-Term Price Movements The notion that today’s price change conveys information about tomorrow’s price change is deeply rooted in most investors’ psyches. There are several ways in which this hypothesis can be tested in financial markets.
Serial Correlation The serial correlation measures the correlation between price changes in consecutive time periods, whether hourly, daily, or weekly, and is a measure of how much the price change in any period depends on the price change over the previous time period. A serial correlation of zero would therefore imply that price changes in consecutive time periods are uncorrelated with each other, and can thus be viewed as a rejection of the hypothesis that investors can learn about future price changes from past ones. A serial correlation that is positive and statistically significant could be viewed as evidence of price momentum in markets, and would suggest that returns in a period are more likely to be positive if the prior period’s returns were positive, and negative if previous returns were negative. A serial correlation that is negative and statistically significant could be evidence of price reversals, and would be consistent with a market where positive returns are more likely to follow negative returns and vice versa. From the viewpoint of investment strategy, serial correlations can be exploited to earn excess returns. A positive serial correlation would be exploited by a strategy of buying after periods with positive returns and selling after periods with negative returns. A negative serial correlation would suggest a strategy of buying after periods with negative returns and selling after periods with positive returns. Since these strategies generate transactions costs, the correlations have to be large enough to allow investors to generate profits to cover these costs. It is therefore entirely possible that there is serial correlation in returns, without any opportunity to earn excess returns for most investors. The earliest studies of serial correlation—Alexander (1964), Cootner (1962), and Fama (1965)—all looked at large U.S. stocks and concluded that the serial correlation in stock prices was small. Fama, for instance, found that 8 of the 30 stocks listed in the Dow had negative serial correlations and that most of the serial correlations were less than 0.05. Other studies confirm these findings not only for smaller stocks in the United States, but also for other markets. For instance, Jennergren and Korsvold (1974) report low serial correlations for the Swedish equity market, and Cootner (1961) concludes that serial correlations are low in commodity markets as well. Although there may be statistical significance associated with some of these correlations, it is unlikely that there is enough correlation to generate excess returns. The serial correlation in short period returns is affected by market liquidity and the presence of a bid-ask spread. Not all stocks in an index are liquid, and in some cases stocks may not trade during a period. When the stock trades in a subsequent period, the resulting price changes can create positive serial correlation. To see why, assume that the market is up strongly on day 1, but that three stocks in the index do not trade on that day. On day 2, if these stocks are traded, they are likely to go up in price to reflect the increase in the market the previous day. The net result is that you should expect to see positive serial correlation in daily or hourly returns in illiquid market indexes. The bid-ask spread creates a bias in the opposite direction, if transaction prices are used to compute returns, since prices have an equal chance of ending up at the bid or the ask price. The bounce that this induces in prices—from bid to ask to bid again—will result in negative serial correlations in returns. Roll (1984) provides a simple measure of this relationship:
where the serial covariance in returns measures the covariance between return changes in consecutive time periods. For very short return intervals, this bias induced in serial correlations might dominate and create the mistaken view that price changes in consecutive time periods are negatively correlated. 97
Filter Rules In a filter rule, an investor buys an investment if the price rises X percent from a previous low and holds the investment until the price drops X percent from a previous high. The magnitude of the change (X percent) that triggers the trades can vary from filter rule to filter rule, with smaller changes resulting in more transactions per period and higher transaction costs. Figure 6.1 graphs out a typical filter rule. Figure 6.1 Filter Rule
This strategy is based on the assumption that price changes are serially correlated and that there is price momentum (i.e., stocks that have gon e up strongly in the past are more likely to keep going up than to go down). Table 6.1 summarizes results—Fama and Blume (1966) and Jensen and Bennington (1970)—from a study on returns, before and after transactions costs, on a trading strategy based on filter rules ranging from 0.5 percent to 20 percent. (A 0.5 percent rule implies that a stock is bought when it rises 0.5 percent from a previous low and is sold when it falls 0.5 percent from a prior high.) Table 6.1 Returns on Filt er Rule Strategies
The only filter rule that beats the returns from the buy-and-hold strategy is the 0.5 percent rule, but it does so before transaction costs. This strategy creates 12,514 trades during the period, which generate enough transaction costs to wipe out the principal invested by the investor. Wh ile this test is dated, it also illustrates basic problems with strategies that require frequent short-term trading. Even though these strategies may earn excess returns prior to transaction costs, adjusting for these costs can wipe out the excess returns. One popular indicator among investors that is a variant on the filter rule is the relative strength measure, which relates recent prices on stocks or other investments either to average prices over a specified period, say over six months, or to the price at the beginning of the period. Stocks that score high on the relative strength measure are considered good investments. This investment strategy is also based upon the assumption of price momentum.
Runs Tests A runs test is a nonparametric variation on the serial correlation, and it is based on a count of the number of runs 98
(i.e., sequences of price increases or decreases) in the price changes. Thus, the following time series of price changes, where U is an increase and D is a decrease, would result in the following runs:
There were 18 runs in this price series of 33 periods. The actual number of runs in the price series is compared against the number that can be expected in a series of this length, assuming that price changes are random.9 If the actual number of runs is greater than the expected number, there is evidence of negative correlation in price changes. If it is lower, there is evidence of positive correlation. A 1966 study by Niederhoffer and Osborne of price changes in the Dow 30 stocks assuming daily, four-day, nine-day, and 16-day return intervals provided the following results:
Based on these results, there is evidence of positive correlation in daily returns but no evidence of deviations from normality for longer return intervals. Again, while the evidence is dated, it serves to illustrate the point that long strings of positive and negative changes are, by themselves, insufficient evidence that markets are not random, since such behavior is consistent with price changes following a random walk. It is the recurrence of these strings that can be viewed as evidence against randomness in price behavior.
Longer-term Price Movements While most of the earlier studies of price behavior focused on shorter return intervals, more attention has been paid to price movements over longer periods (one-year to five-year periods) in recent years. Here, there is an interesting dichotomy in the results. When long term is defined as months rather than years, there seems to be a tendency toward positive serial correlation or price momentum. However, when long term is defined in terms of years, there is substantial negative correlation in the returns, suggesting that markets reverse themselves over long periods.
Weekly and Monthly Price Momentum In the preceding section, we noted that the evidence of short-term price patterns is weak and that any price dependence over very short time periods (minutes or hours) can be attributed more to market structure (liquidity, bid-ask spreads) than to inefficiency. We also argued that while chartists who track these short-term price movements abound, few seem to emerge as consistent winners. As we extend our time periods from minutes to days and from days to weeks, there is some evidence of price momentum. Put differently, stocks that have gone up in the last few weeks or months seem to have a tendency to continue to outperform the market in the next few weeks or months, and stocks that have gone down in the recent weeks or months continue to languish in the next few weeks or months. Jegadeesh and Titman (1993, 2001) present evidence of what they call price momentum in stock prices over time periods of up to eight months — stocks that have gone up in the last six months tend to continue to go up, whereas stocks that have gone down in the last six months tend to continue to go down. The momentum effect is just as strong in the European markets, though it seems to be weaker in emerging markets. What may cause this momentum? One potential explanation is that mutual funds are more likely to buy past winners and dump past losers, thus generating price continuity.
Annual or Multi-year Price Reversal When the long term is defined in terms of years, there is negative correlation in returns, suggesting that markets reverse themselves over very long periods. Fama and French (1988) examined five-year returns on stocks from 1941 to 1985 and present evidence of this phenomenon. They found that serial correlation is more negative in fiveyear returns than in one-year returns, and is much more negative for smaller stocks rather than larger stocks. Figure 6.2 summarizes one-year and five-years serial correlation by size class for stocks on the New York Stock Exchange. Figure 6.2 One-Year and Five-Year Correlations: Market Value Class, 1941 to 1985 Source: Fama and French (1988).
99
Since there is evidence that prices reverse th emselves in the long term, it might be worth examining whether such price reversals can be used by investors to profit. To isolate the effect of such price reversals on the extreme portfolios, DeBondt and Thaler constructed a winner portfolio of 35 stocks that had gone up the most over the prior year, and a loser portfolio of 35 stocks that had gone down the most over the prior year, each year from 1933 to 1978, and examined returns on these portfolios for the sixty months following the creation of the portfolio. Figure 6.3 summarizes the excess returns for winner and loser portfolios. This analysis suggests that loser portfolios clearly outperform winner portfolios in the 60 months following creation. This evidence is consistent with market overreaction and correction in long return intervals. Figure 6.3 Excess Returns for Winner and Loser Portfolios Source: DeBondt and Thaler (1985).
There are many academics as well as practitioner, who suggest that these findings may be interesting but that they overstate potential returns on loser portfolios. For instance, there is evidence that loser portfolios are more likely to contain low-priced stocks (selling for less than $5), which generate higher transactions costs and are also more likely to offer heavily skewed returns; that is, the excess returns come from a few stocks making phenomenal returns rather than from consistent performance. One study of the winner and loser portfolios attributes the bulk of the excess returns of loser portfolios to low-priced stocks and also finds that the results are sensitive to when the portfolios are created. Loser portfolios created every December earn significantly higher returns than portfolios created every June.
Speculative Bubbles, Crashes, and Panics Historians who have examined the behavior of financial markets over time have challenged the assumption of rationality that underlies much of efficient market theory. They point to the frequency with which speculative bubbles have formed in financial markets as investors buy into fads or get-rich-quick schemes, and the crashes when these bubbles have ended, and suggest that there is nothing to prevent the recurrence of this phenomenon in today’s financial markets. There is some evidence in the literature of irrationality on the part of market players.
Experimental Studies of Rationality Some of the most interesting evidence on market efficiency and rationality in recent years has come from experimental studies. While most experimental studies suggest that traders are rational, there are some examples 100
of irrational behavior in some of these studies. One such study was done at the University of Arizona. In an experimental study, traders were told that a payout would be declared after each trading day, determined randomly from four possibilities—0, 8, 28, or 60 cents. The average payout was 24 cents. Thus the share’s expected value on the first trading day of a 15-day experiment was $3.60 (24 cents times 15), the second day was $3.36, and so on. The traders were allowed to trade each day. The results of 60 such experiments are summarized in Figure 6.4. Figure 6.4 Trading Price by Trading Day
There is clear evidence here of a speculative bubble forming during periods 3 to 5, where prices exceed expected values by a significant amount. The bubble ultimately burst s, and prices approach expected value by the end of the period. If this mispricing is feasible in a simple market, where every investor obtains the same information, it is clearly feasible in real financial markets, where there is much more differential information and much greater uncertainty about expected value. It should be pointed out that some of the experiments were run with students, and some with Tucson businessmen with real-world experience. The results were similar for both groups. Furthermore, when price curbs of 15 cents were introduced, the booms lasted even longer because traders knew that prices would not fall by more than 15 cents in a period. Thus, the notion that price limits can control speculative bubbles seems misguided.
Behavioral Finance The irrationality sometimes exhibited by investors has given rise to a whole new area of finance called behavioral finance. Using evidence gathered from experimental psychology, researchers have tried to both model how investors react to information and predict how prices will change as a consequence. They have been far more successful at the first endeavor than the second. For instance, the evidence seems to suggest that: Investors do not like to admit their mistakes. Consequently, they tend to hold on to losing stocks far too long, or in some cases double up their bets (investments) as stocks drop in value. More information does not always lead to better investment decisions. Investors seem to suffer both from information overload and from a tendency to react to the latest piece of information. Both result in investment decisions that lower returns in the long term. If the evidence on how investors behave is so clear-cut, you might ask, why are the predictions that emerge from these models so noisy? The answer, perhaps, is that any model that tries to forecast human foibles and irrationalities is, by its very nature, unlikely to be a stable one. Behavioral finance may emerge ultimately as a trump card in explaining why and how stock prices deviate from true value, but its role in devising investment strategy still remains questionable.
BEHAVORIAL FINANCE AND VALUATION In 1999, Robert Shiller made waves in both academia and investment houses with his book titled Irrational Exuberance . His thesis is that investors are often not just irrational but irrational in predictable ways—overreacting to some information and buying and selling in herds. His work forms part of a growing body of theory and evidence of behavioral finance, which can be viewed as a congruence of psychology, statistics, and finance. While the evidence presented for investor irrationality is strong, the implications for valuation are less so. You can consider discounted cash flow valuation to be the antithesis of behavioral finance, because it takes the point of view that the value of an asset is the present value of the expected cash flows generated by that asset. With this context, there are two ways in which you can look at the findings in behavioral finance: 1. Irrational behavior may explain why prices can deviate from value (as estimated in a discounted cash flow model). Consequently, it provides the foundation for the excess returns earned by rational investors who base decisions on estimated value. Implicit here is the assumption that markets ultimately recognize their irrationality and correct themselves.
101
2. It may also explain why discounted cash flow values can deviate from relative values (estimated using multiples). Since the relative value is estimated by looking at how the market prices similar assets, irrationalities that exist will be priced into the asset.
MARKET REACTION TO INFORMATION EVENTS Some of the most powerful tests of market efficiency are event studies where market reaction to informational events (such as earnings and takeover announcements) has been scrutinized for evidence of inefficiency. While it is consistent with market efficiency for markets to react to new information, the reaction has to be instantaneous and unbiased. This point is made in Figure 6.5 by contrasting three different market reactions to information announcements containing good news. Figure 6.5 Information and Price Adjustment
Of the three market reactions pictured here, only the first one is consi stent with an efficient market. In the second market, the information announcement is followed by a gradual increase in prices, allowing investors to make excess returns after the announcement. This is a s low learning market in which some investors will make excess returns on the price drift. In the third market, the price reacts instantaneously to the announcement, but corrects itself in the days that follow, suggesting that the initial price change was an overreaction to the information. Here again, an enterprising investor could have sold short after the announcement and expected to make excess returns as a consequence of the price correction.
Earnings Announcements When firms make earnings announcements, they convey information to financial markets about their current and future prospects. The magnitude of the information, and the size of the market reaction, should depend on how much the earnings report exceeds or falls short of investor expectations. In an efficient market, there should be an instantaneous reaction to the earnings report, if it contains surprising information, and prices should increase following positive surprises and decline following negative surprises. Since actual earnings are compared to investor expectations, one of the key parts of an earnings event study is the measurement of these expectations. Some of the earlier studies used earnings from the same quarter in the prior year as a measure of expected earnings (i.e., firms that report increases in quarter-to-quarter earnings provide positive surprises, and those that report decreases in quarter-to-quarter earnings provide negative surprises). In more recent studies, analyst estimates of earnings have been used as a proxy for expected earnings and compared to the actual earnings. Figure 6.6 provides a graph of price reactions to earnings surprises, classified on the basis of magnitude into different classes from “most negative” earnings reports (group 1) to most positive earnings reports (group 10). The evidence contained in this graph is consistent with the evidence in most earnings announcement studies: Figure 6.6 Post-Announcement Drift after Unexpected Qua rterly Earnings Surprises: U.S. Companies from 1988 to 2002 Source: Updated version of Rendleman, Jones, and Latrané (1982).
102
The earnings announcement clearly conveys valuable information to financial markets; there are positive excess returns (cumulative abnormal returns) around positive announcements and negative excess returns around negative announcements. There is some evidence of a market reaction in the days immediately prior to the earnings announcement that is consistent with the nature of the announcement (i.e., prices tend to go up on the day before positive announcements and down on the day before negative announcements). This can be viewed as evidence of either insider trading, information leakage, or getting the announcement date wrong.10 There is some evidence, albeit weak, of a price drift in the days following an earnings announcement. Thus a positive report evokes a positive market reaction on the announcement date, and there are mildly positive excess returns in the days following the earnings announcement. Similar conclusions emerge for negative earnings reports. The management of a firm has some discretion on the timing of earnings reports, and there is some evidence that the timing affects expected returns. A 1989 study of earnings reports, classified by the day of the week that the earnings are reported, reveals that earnings and dividend reports on Fridays are much more likely to contain negative information than announcements on any other day of the week. This is shown in Figure 6.7. Figure 6.7 Earnings and Dividend Reports by Day of the Week Source: Damodaran (1989).
There is also some evidence discussed by Chambers and Penman (1984) that earnings anno uncements that are delayed, relative to the expected announcement date, are much more likely to contain bad news than earnings announcements that are early or on time. This is graphed in Figure 6.8. Earnings announcements that are more than six days late relative to the expected announcement date are much more likely to contain bad news and evoke negative market reactions than earnings announcements that are on time or early. Figure 6.8 Cumulated Abnormal Returns and Earnings Delay Source: Chambers and Penman (1984).
Investment and Project Announcements Firms frequently make announcements of their intentions of investing resources in projects and research and development. There is evidence that financial markets react to these announcements. The question of whether markets have a long-term or short-term perspective can be partially answered by looking at these market reactions. 103
If financial markets are as short-term as some of their critics claim, they should react negatively to announcements by the firm that it plans to invest in research and development. As Table 6.2, which looks at market reactions to various investment announcements makes clear, the evidence suggests that the market reaction to investment announcements is generally positive, albeit discriminating. Table 6.2 Market Reactions to Investment Announcements Source: Chan, Ma rtin, and Kensinger (1990); McConnell and Muscarella (1985).
Abnormal Returns On
In
Type of Announcement Announcement Day Announcement Month Joint venture formations
0.399%
1.412%
R&D expenditures
0.251%
1.456%
Product strategies
0.440%
–0.35%
Capital expenditures
0.290%
1.499%
All announcements
0.355%
0.984%
This table excludes the largest investments that most firms make, which are acquisitions of other firms. Here the evidence is not so favorable. In about 55 percent of all acquisitions, the stock price of the acquiring firm drops on the announcement of the acquisition, reflecting the market’s beliefs that firms tend to overpay on acquisitions.
MARKET ANOMALIES Merriam-Webster’s Collegiate Dictionary defines an anomaly as a “deviation from the common rule.” Studies of market efficiency have uncovered numerous examples of market behavior that are inconsistent with existing models of risk and return and often defy rational explanation. The persistence of some of these patterns of behavior suggests that the problem, in at least some of these anomalies, lies in the models being used for risk and r eturn rather than in the behavior of financial markets. The following section summarizes some of the more widely noticed anomalies in financial markets in the United States and elsewhere.
Anomalies Based on Firm Characteristics There are a number of anomalies that have been related to observable firm characteristics, including the market value of equity, price-earnings ratios, and price–book value ratios.
The Small Firm Effect Studies such as Banz (1981) and Keim (1983) have consistently found that smaller firms (in terms of market value of equity) earn higher returns than larger firms of equivalent risk, where risk is defined in terms of the market beta. Figure 6.9 summarizes returns for stocks in 10 market value classes for the period from 1927 to 2010. Figure 6.9 Annual Returns by Size Class, 1927 to 1983
The size of the small firm premium, while it has varied across time, has been generally positive. It was highest during the 1970s and early 1980s and lowest during the 1990s before returning in the first half of the last decade. The persistence of this premium has led to several poss ible explanations. 104
Figure 6.10 Returns on CRSP Small Stocks versus DFA Small Stock Fund
1. The transaction costs of investing in small stocks are significantly higher than the transaction costs of investing in larger stocks, and the premiums are estimated prior to these costs. While this is generally true, the dif ferential transaction costs are unlikely to explain the magnitude of the premium across time, and are likely to become even less critical for longer investment horizons. The difficulties of replicating the small firm premiums that are observed in the studies in real time are illustrated in Figure 6.10, which compares the returns on a hypothetical small firm portfolio (CRSP Small Stocks) with the actual returns on a small firm mutual fund (DFA Small Stock Fund), which passively invests in small stocks. 2. The capital asset pricing model may not be the right model for risk, and betas underestimate the true risk of small stocks. Thus, the small firm premium is really a measure of the failure of beta to capture risk. The additional risk a ssociated with small stocks may come from several sources. First, the estimation risk associated with estimates of beta for small firms is much greater than the estimation risk associated with beta estimates for larger firms. The small firm premium may be a reward for this additional estimation risk. Second, there may be additional risk in investing in small stocks because far less information is available on these stocks. In fact, studies indicate that stocks that are neglected by analysts and institutional investors earn an excess return that parallels the small firm premium. There is evidence of a small firm premium in markets outside the United States as well. Dimson and Marsh (1986) examined stocks in the United Kingdom from 1955 to 1984 and found that the annual returns on small stocks exceeded those on large stocks by 6 percent annually over the period. Chan, Hamao, and Lakonishok (1991) report a small firm premium of about 5 percent for Japanese stocks between 1971 and 1988.
Price-Earnings Ratios Investors have long argued that stocks with low price-earnings ratios are more likely to be undervalued and earn excess returns. For instance, Benjamin Graham, in his investment classic The Intelligent Investor,11 used low priceearnings ratios as a screen for finding undervalued stocks. Studies [Basu (1977); Basu (1983)] that have looked at the relationship between PE ratios and excess returns confirm these priors. Figure 6.11 summarizes annual returns by PE ratio classes for stocks from 1952 to 2010. Firms in the lowest PE ratio class earned an average return of 18.9 percent during the period, while firms in the highest PE ratio class earned an average return of only 1 1.4 percent. Figure 6.11 PE Ratios and Stock Returns, 1952 to 2010
The excess returns earned by low PE ratio stocks also persist in other international markets. Table 6.3 summarizes 105
the results of studies looking at this phenomenon in markets outside the Unit ed States. Table 6.3 Excess Returns on Low PE Ratio Stocks by Country, 1989–1994 Annual premium: Premium earned over an index of equally weighted stocks in that market between January 1, 1989, and De cember 31, 1994. These numbers were obtained from a Merrill Lynch Survey of Proprietary Indices.
Annual Premium Earned by Country
Lowest-PE Stocks (Bottom Quintile)
Australia
3.03%
France
6.40%
Germany
1.06%
Hong Kong
6.60%
Italy
14.16%
Japan
7.30%
Switzerland
9.02%
United Kingdom 2.40%
The excess returns earned by low price-earnings ratio stocks are difficult to justify using a variation of the argument used for small stocks (i.e., that the risk of low PE ratios stocks is understated in the CAPM). Low PE ratio stocks generally are characterized by low growth, large size, and stable businesses, all of which should work toward reducing their risk rather than increasing it. The only explanation that can be given for this phenomenon, which is consistent with an efficient market, is that low PE ratio stocks generate large dividend yields, which would have created a larger tax burden because dividends are taxed at higher rates.
Price–Book Value Ratios Another statistic that is widely used by investors in investment strategy is price–book value ratios. A low price–book value ratio has been considered a reliable indicator of undervaluation in firms. In studies that parallel those done on price-earnings ratios, the relationship between returns and price–book value ratios has been examined. The consistent finding from these studies is that there is a negative relationship between returns and price–book value ratios—low price–book value ratio stocks earn higher returns than high price–book value ratio stocks. Rosenberg, Reid, and Lanstein (1985) find that the average returns on U.S. stocks are positively related to the ratio of a firm’s book value to market value. Between 1973 and 1984, the strategy of picking stocks with high book-price ratios (low price-book values) yielded an excess return of 36 basis points a month. Fama and French (1992), in examining the cross section of expected stock returns between 1963 and 1990, established that the positive relationship between book-to-price ratios and average returns persists in both the univariate and multivariate tests, and is even stronger than the size effect in explaining returns. When they classified firms on the basis of book-toprice ratios into 12 portfolios, firms in the lowest book-to-price (highest price-book) class earned an average monthly return of 0.30 percent, while firms in the highest book-to-price (lowest price-book) class earned an average monthly return of 1.83 percent for the 1963–1990 period. Chan, Hamao, and Lakonishok (1991) find that the book-to-market ratio has a strong role in explaining the cross section of average returns on Japanese stocks. Capaul, Rowley, and Sharpe (1993) extend the analysis of price– book value ratios across other international markets, and conclude that value stocks (i.e., stocks with low price– book value ratios) earned excess returns in every market that they analyzed between 1981 and 1992. Their annualized estimates of the return differential earned by stocks with low price–book value ratios, over the market index, were: Country
Added Return to Low Price–Book Value Portfolio
France
3.26%
Germany
1.39%
Switzerland
1.17%
United Kingdom 1.09% Japan
3.43%
United States
1.06%
Europe
1.30%
Global
1.88%
106
A caveat is in order. Fama and French pointed out that low price–book value ratios may operate as a measure of risk, since firms with prices well below book value are more likely to be in trouble and go out of business. Investors therefore have to evaluate for themselves whether the additional returns made by such firms justify the additional risk taken on by investing in them.
Temporal Anomalies There are a number of peculiarities in return differences across calendar time that not only are difficult to rationalize but are also suggestive of inefficiencies. Furthermore, some of these temporal anomalies are related to the small firm effect described in the previous section.
January Effect Studies of returns in the United States and other major financial markets [Roll (1983); Haugen and Lakonishok (1988)] consistently reveal strong differences in return behavior across the months of the year. Figure 6.12 reports average returns by months of the year from 1926 to 1983. Returns in January are significantly higher than returns in any other month of the year. This phenomenon is called the year-end or January effect, and it can be traced to the first two weeks in January. Figure 6.12 Returns by Month of the Year: U.S. stocks from 1927 to 2010
The relationship between the January effect and the small firm effect [(Keum (1983) and Reinganum (1983)] adds to the complexity of this phenomenon. The January effect is much more accentuated for small firms than for larger firms, and roughly half of the small firm premium described in the prior section is earned in the first two weeks of January. Figure 6.13 graphs returns in January by size and risk class for data from 1935 to 1986. Figure 6.13 Returns in January by Size and Risk Class, 1935–1986 Source: Ritter and Chopra (1989).
107
A number of explanations have been advanced for the January effect, but few hold up to serious scrutiny. Reinganum su ggested that there is tax loss selling by investors at the end of the year on stocks that have lost money to capture the capital gain, driving prices down, presumably below true value, in December, and a buying back of the same stocks in January,12 resulting in the high returns. The fact that the January effect is accentuated for stocks that have done worse over the prior year is offered as evidence for this explanation. There are several pieces of evidence that contradict it, though. First, there are countries, like Australia, that have a different tax year but continue to have a January effect. Second, the January effect is no greater, on average, in years following bad years for the stock market than in other years. A second rationale is that the January effect is related to institutional trading behavior around the turn of the years. It has been noted, for instance, that the ratio of buys to sells for institutions drops significantly below average in the days before the turn of the year and picks up to above average in the months that follow. This is illustrated in Figure 6.14. It is argued that the absence of institutional buying pushes down prices in the days before the turn of the year and pushes up prices in the days after. Figure 6.14 Institutional Buying/Selling around Year-End
The universality of the January effect is illustrated in Figure 6.15, which examines returns in January versus the other months of the year in several major financial markets, and finds strong evidence of a January effect i n every 108
market [Haugen and Lakonishok (1988); Gultekin and Gultekin (1983)]. Figure 6.15 Returns in January versus Other Months—Major Financial Markets Source: Gultek in and Gultekin (1983).
Weekend Effect The weekend effect is another return phenomenon that has persisted over extraordinarily long periods and over a number of international markets. It refers to the differences in returns between Mondays and other days of the week. The significance of the return difference is brought out in Figure 6.16, which graphs returns by days of the week from 1962 to 1978 [Gibbons and Hess (1981)]. Figure 6.16 Average Daily Returns by Day of the Week, 1962–1978 Source: Gibbons and Hess (1981).
The returns on Mondays are significantly negative, whereas the returns on every other day of the trading week are not. Ther e are a number of other findings on the Monday effect that have fleshed this out. First, the Monday effect is really a weekend effect since the bulk of the negative returns is manifested in the Friday close to Monday open returns. The intraday returns on Monday are not the culprits in creating the negative returns. Second, the Monday effect is worse for small stocks than for larger stocks. Third, the Monday effect is no worse following three-day weekends than following two-day weekends. There are some who have argued that the weekend effect is the result of bad news being revealed after the close of trading on Friday and during the weekend. They point to Figure 6.7, which reveals that more negative earnings 109
reports are revealed after close of trading on Friday. Even if this were a widespread phenomenon, the return behavior would be inconsistent with a rational market, since rational investors would build the expectation of the bad news over the weekend into the price before the weekend, leading to an elimination of the weekend effect. The weekend effect is fairly strong in most major international markets, as shown in Figure 6.17. The presence of a strong weekend effect in Japan, which allowed Saturday trading for a portion of the period studied here, indicates that there might be a more direct reason for negative returns on Mondays than bad information over the weekend. Figure 6.17 Weekend Effect in International Markets
As a final note, the negative returns on Mondays cannot be attributed to just the absence of trading over the weekend. The returns on days following trading holidays in general are characterized by positive, not negative, returns. Figure 6.18 summarizes returns on trading days following major holidays and confirms this pattern. Figure 6.18 A Holiday Effect? Average Market Returns on Trading Days Following Holidays
EVIDENCE ON INSIDERS AND INVESTMENT PROFESSIONALS There is a sense that insiders, analysts, and portfolio managers must possess an advantage over the average investors in the market and b e able to convert this advantage into excess returns. The evidence on the performance of these investors is actually surprisingly mixed.
Insider Trading The Securities and Ex change Commission (SEC) defines an insider to be an officer or director of the firm or a major stockholder (holding more than 5 percent of the outstanding stock in the firm). Insiders are barred from trading in advance of specific information on the company and are required to file with the SEC when they buy or sell stock in the company. If it is assumed, as seems reasonable, that insiders have better information about the company and consequently better estimates of value than other investors, the decisions by insiders to buy and sell stock should affect stock prices. Figure 6.19, derived from an early study of insider trading by Jaffe (1974), examines excess returns on two groups of stock, classified on the basis of insider trades. The “buy group” includes stocks where buys exceeded sells by the biggest margin, and the “sell group” includes stocks where sells exceeded buys by the biggest margin. Figure 6.19 Cumulative Returns Following Insider Trading: Buy versus Sell Group Source: Jaffe (1974).
110
While it seems like the buy group does significantly better than the sell group in this study, advances in information technolog y have made this information on insider trading available to more and more investors. A more recent study [Seyhun (1998)] of insider trading examines excess returns around both the date the insiders report to the SEC and the date that information becomes available to investors in the official summary. Figure 6.20 presents the contrast between the two event dates. Figure 6.20 Abnormal Returns around Reporting Day versus Official Summary Availability Day
Given the opportunity to buy on the date the insider reports to the SEC, investors could have marginal excess returns, but these return s diminish and become statistically insignificant if investors are forced to wait until the official summary date. None of these studies examine the question of whether insiders themselves make excess returns. The reporting process, as set up now by the SEC, is biased toward legal and less profitable trades and away from illegal and more profitable trades. Though direct evidence cannot be offered for this proposition, insiders trading illegally on private information must make excess returns.
Analyst Recommendations Analysts clearly hold a privileged position in the market for information, operating at the nexus of private and public information. Using both types of information, analysts issue buy and sell recommendations to their clients, who trade on this basis. While both buy and sell recommendations affect stock prices, sell recommendations affect prices much more adversely than buy recommendations affect them positively. Interestingly, Womack (1996) documents that the price effect of buy recommendations tends to be immediate and there is no evidence of price drifts after the announcement, whereas prices continue to trend down after sell recommendations. Figure 6.21 graphs his findings. Stock prices increase by about 3 percent on buy recommendations, whereas they drop by about 4 percent on sell recommendations at the time of the recommendations (three days around reports). In the six months following, prices decline an additional 5 percent for sell recommendations, while leveling off for buy recommendations. Figure 6.21 Market Reaction to Recommendations, 1989 to 1990 Source: Womack (1996).
111
Though analysts provide a valuable service in collecting private information, or maybe because they do, there is a negative relationship in the cross section between returns earned by stocks and the number of analysts following the stock. The same kind of relationship exists between another proxy for interest—institutional ownership—and returns. This evidence [Arbel and Strebel (1983)] suggests that neglected stocks—those followed by few analysts and not held widely by institutions—earn higher returns than widely followed and held stocks.
Money Managers Professional money managers operate as the experts in the field of investments. They are supposed to be better informed, have lower transaction costs, and be better investors overall than smaller investors. The earliest study of mutual funds by Jensen (1968) suggested that this supposition might not hold in practice. His findings, summarized in Figure 6.22 as excess returns on mutual funds, were that the average portfolio manager actually underperformed the market between 1955 and 1964. Figure 6.22 Mutual Fund Performance, 1955 to 1964—the Jensen Study Source: Jensen (1968).
These results have been replicated with mild variations in their conclusions. In the studies that are most favorable for professional money managers , they break even against the market after adjusting for transaction costs, and in those that are least favorable they underpeform the market even before adjusting for transaction costs. The results, when categorized on a number of different bases, do not offer much solace. For instance, Figure 6.23 shows excess returns from 1983 to 1990 and the percentage of money managers beating the market, categorized by investment style. Money managers in every investment style underperform the market index and updated studies since have yielded similar findings. Figure 6.23 Performance of Equity Funds, 1983 to 1990
112
Figure 6.24 looks at the payoff to active portfolio management by measuring the added value from trading actively during the course of the year, and finds that returns drop between 0.5 percent and 1.5 percent a year as a consequence. Figure 6.24 The Payoff to Active Money Management: Equity Funds Note: This chart me asures the difference between actual return on equity funds and return on hypothetical portfolio frozen at beginning of period.
Finally, we find no evidence of continuity in performance. It classified money managers into quartiles and examined the probabilities of movement from one quartile to another each year from 1983 to 1990. The results are summarized in Table 6.4. Table 6.4 Probabilities of Transition from One Quartile to Another
113
Table 6.4 indicates that a money manager who was ranked in the first quartile in a period had a 26 percent chance of being ranked in the first quartile in the next period and a 27 perce nt chance of being ranked in the bottom quartile. There is some evidence of reversal in the portfolio mana gers in the lowest quartile, though some of that may be a reflection of the higher-risk portfolios that they put together. Thus, the overall evidence is that money managers collectively add little value for investors with their active investing strategies. So, is there any good news for money managers in these studies? There are some glimmers:
Hot hands phenomenon: While there is little evidence of overall continuing in mutual fund performance, there is some evidence that money managers who have performed well in the recent past are likely to outperform the market in the near future. It is unclear, however, how long this momentum lasts, before there is reversal. This evidence mirrors our earlier analysis of stock prices showing momentum over weeks and months and reversing themselves in the long term. Even this good news has to be taken with an ounce of caution, since it is possible that sheer chance and selection bias can still explain these positive runs of superior performance. Skill versus luck: There is some debate as to whether the differences in returns that money managers earn over long periods can be entirely attributed to luck. Fama and French (2010) argue that when returns are measured net of costs, there is little evidence that mutual funds beat the market. However, when they look at gross returns, there is some evidence of differences in skill; they estimate that superior managers generate about 1.25% more than the average. Tax, liquidity, and time horizon arbitrage: A money manager with a tax rate much lower than that of other money managers, lower need for liquidity or lower transactions costs than other investors, and/or a longer time horizon may be able to exploit these differences to get investments at bargain prices and earn excess returns. Although there is anecdotal evidence that such investors exist, many of them are undone by their own success; as they attract more money, they lose their focus and their competitive edge. The bottom line: It is possible to beat the market, but it is hard work. There are no magic bullets or simple formulae for investment success. The money managers who beat the market consistently over time tend to be few and far between. Although they may adopt different strategies and have different views on markets, they share some common characteristics. They have well thought out investment philosophies, play to their strengths, and stay disciplined.
CONCLUSION The question of whether markets are efficient will always be a provocative one, given the implications that efficient markets have for investment management and research. If an efficient market is defined as one where the market price is an unbiased estimate of the true value, it is quite clear that some markets will always be more efficient than others and that markets will always be more efficient to some investors than to others. The capacity of a market to correct inefficiencies quickly will depend, in part, on the ease of trading, t he transaction costs, and the vigilance of profit-seeking investors in that market. While market efficiency can be tested in a number of different ways, the two most widely used tests to test efficiency are event studies, which examine market reactions to information events, and portfolio studies, which evaluate the returns of portfolios created on the basis of observable characteristics. It does make sense to be vigilant, because bias can enter these studies, intentionally or otherwise, in a number of different ways and can lead to unwarranted conclusions and, worse still, wasteful investment strategies. There is substantial evidence of irregularities in market behavior related to systematic factors such as size, priceearnings ratios, and price–book value ratios, as well as to time—the January and the weekend effects. While these irregularities may be inefficiencies, there is also the sobering evidence that professional money managers, who are in a position to exploit these inefficiencies, have a very difficult time consistently beating financial markets. Read together, the persistence of the irregularities and the inability of money managers to beat the market are testimony to the gap between empirical tests on paper and real-world money management in some cases, and the failure of the models of risk and return in others.
MARKET INEFFICIENCIES AND MONEY MANAGER PERFORMANCE 114
The evidence on markets is contradictory. On the one hand, there seem to be numerous patterns in stock prices—stock prices reverse course in the long term and returns are higher in January—and evidence of market anomalies—small-market-cap firms with low price-to-book and priceto-earnings ratios seem to handily beat the market. On the other hand, there seems to be little evidence of money managers being able to exploit these findings to beat the market. There are a number of possible explanations. The most benign one is that the inefficiences show up mostly in hypothetical studies and that the transaction cost and execution problems associated with converting these inefficiencies into portfolios overwhelm the excess returns. A second possible explanation is that the studies generally look at the long term; many are over 20 to 50 years. Over shorter periods, there is substantially more uncertainty about whether small stocks will outperform large stocks and whether buying losers will generate excess returns. There are no investment strategies that are sure bets for short periods. Pradhuman (2000) illustrates this phenomenon by noting that small-cap stocks have underperformed large-cap stocks in roughly one out of every four years in the past 50 years. Bernstein (1998) notes that while value investing (buying low PE and low price-to-book value stocks) may earn excess returns over long periods, growth investing has outperformed value investing over many five-year periods during the past three decades. A third explanation is that portfolio managers do not consistently follow any one strategy but jump from one strategy to another, both increasing their expenses and reducing the likelihood that the strategy can generate excess returns in the long term.
QUESTIONS AND SHORT PROBLEMS In the problems following, use an equity risk premium of 5.5 percent if none is specified. 1. Which of the following is an implication of market efficiency? (There may be more than one right answer.) a. Resources are allocated among firms efficiently (i.e., put to best use). b. No investor will do better than the market in any time period. c. No investor will do better than the market consistently. d. No investor will do better than the market consistently after adjusting for risk. e. No investor will do better than the market consistently after adjusting for risk and transaction costs. f. No group of investors will do better than the market consistently after adjusting for risk and transaction costs. 2. Suppose you are following a retailing stock that has a strong seasonal pattern to sales. Would you expect to see a seasonal pattern in the stock price as well? 3. Tests of market efficiency are often referred to as joint tests of two hypotheses—the hypothesis that the market is efficient and an expected returns model. Explain. Is it ever possible to test market efficiency alone (i.e., without jointly testing an asset pricing model)? 4. You are in a violent argument with a chartist. He claims that you are violating the fundamental laws of economics by trying to find intrinsic value. “Price is determined by demand and supply, not by some intrinsic value.” Is finding an intrinsic value inconsistent with demand and supply? 5. You are testing the effect of merger announcements on stock prices. (This is an event study.) Your procedure goes through the following steps: Step 1: You choose the 20 biggest mergers of the year. Step 2: You isolate the date the merger became effective as the key day around which you will examine the data. Step 3: You look at the returns for the five days after the effective merger date. By looking at these returns (0.13%) you conclude that you could not have made money on merger announcements. Are there any flaws that you can detect in this test? How would you correct for them? Can you devise a stronger test? 6. In an efficient market, the market price is defined to be an “unbiased estimate” of the true value. This implies that (choose one): a. The market price is always equal to true value. b. The market price has nothing to do with true value. c. Markets make mistakes about true value, and investors can exploit these mistakes to make money. d. Market prices contain errors, but the errors are random and therefore cannot be exploited by investors. e. No one can beat the market. 7. Evaluate whether the following actions are likely to increase stock market efficiency, decrease it, or leave it unchanged, and explain why. 115
a. The government imposes a transaction tax of 1% on all stock transactions. Increase efficiency ____ Decrease efficiency ____ Leave unchanged ____ b. The securities exchange regulators impose a restriction on all short sales to prevent rampant speculation. Increase efficiency ____ Decrease efficiency ____ Leave unchanged ____ c. An options market, trading call and put options, is opened up, with options traded on many of the stocks listed on the exchange. Increase efficiency ____ Decrease efficiency ____ Leave unchanged ____ d. The stock market removes all restrictions on foreign investors acquiring and holding stock in companies. Increase efficiency ____ Decrease efficiency ____ Leave unchanged ____ 8. The following is a graph of cumulative abnormal returns around the announcement of asset divestitures by major corporations.
How best would you explain the: a. Market behavior before the announcement? b. Market reaction to the announcement? c. Market reaction after the announcement? 9. What is the phenomenon of the size effect in stock performance? How does it relate to the turn-of-the-year effect? Can you suggest any good reasons why small stocks, after adjusting for beta, still do better than large stocks? What strategy would you follow to exploit this anomaly? What factors do you have to keep in mind? 10. A study examining market reactions to earnings surprises found that prices tend to drift after earnings surprises. What does this tell you about the market’s capacity to learn from events and new information? What cross-sectional differences would you expect to find in this learning behavior? (Would you expect to see a greater price drift in some types of firms than in others? Why?) How would you try to exploit this anomaly? What possible costs would you have to keep in mind? 11. One explanation of the turn-of-the-year or January effect has to do with sales and purchases related to the tax year. a. Present the tax effect hypothesis. b. Studies have shown that the January effect occurs internationally, even in countries where the tax year does not start in January. Speculate on a good reason for this. 12. The following are the expected price appreciation and dividend yield components of returns on two portfolios —a high dividend yield portfolio and a low dividend yield portfolio. Portfolio
Expected Price Appreciation Expected Dividend Yield
High yield 9%
5%
116
Low yield 12%
1%
You are a taxable investor who faces a tax rate of 40% on dividends. What would your tax rate on capital gains need to be for you to be indifferent between these two portfolios? 13. Answer true or false to the following questions: a. Low price-earnings stocks, on average, earn returns in excess of expectations, while high price-earnings stocks earn less than expected. This is primarily because lower PE ratio stocks have lower risk. True ____ False ____ b. The small firm effect, which refers the positive excess returns earned, on average, by small firms, is primarily caused by a few small firms that make very high positive returns. True ____ False ____ c. Investors generally cannot make money on analyst recommendations, because stock prices are not affected by these recommendations. True ____ False ____ 14. You are examining the performance of two mutual funds. AD Value Fund has been in existence since January 1, 1988, and invests primarily in stocks with low price-earnings ratios and high dividend yields. AD Growth Fund has also been in existence since January 1, 1988, but it invests primarily in high-growth stocks, with high PE ratios and low or no dividends. The performance of these funds over the past five years is summarized as follows:
The average risk-free rate during the period was 6%. The current risk-free rate is 3%. a. How well or badly did these funds perform after adjusting for risk? b. Assume that the front-end load on each of these funds is 5% (i.e., if you put $1,000 in each of these funds today, you would only be investing $950 after the initial commission). Assume also that the excess returns you have calculated in part (a) will continue into the future and that you choose to invest in the fund that outperformed the market. How many years would you have to hold this fund to break even? Since returns are positively skewed—that is, large positive returns are more likely than large negative returns (you cannot lose more than 100% on a stock)—less than half of all investors will probably beat the market. 1
One of the enduring pieces of evidence against market efficiency lies in the performance records posted by many of the investors who learned their lessons from Benjamin Graham in the 1950s. No probabi lity statistics could ever explain the consistency and superiority of their records. 2
In most financial transactions, the announcement date tends to precede the event date by several days and, sometimes, weeks. 3
4
The standard levels of significance for t statistics are:
Level One-Tailed Two-Tailed 1%
2.33
2.55
5%
1.66
1.96
5
The t statistics are marginally significant at the 5% level.
Though there are practicial limits on how big the universe can be, care should be taken to make sure that no biases enter at this stage of the process. An obvious bias would be to pick only stocks that have done well over the time period for the universe. 6
One parametric test is an F test, which tests for equality of means across groups. This test can be conducted assuming either that the groups have the same variance or that they have different variances. 7
An example of a nonparametric test is a rank sum test, which ranks returns across the entire sample and then sums the ranks within each group to check whether the rankings are random or systematic. 8
9
There are statistical tables that summarize the expected number of runs, assuming randomness, in a series of 117
any length. 10 The Wall Street Journal is often used as an information source to extract announcement dates for earnings. For some firms, news of the announcement may actually cross the news wire the day before the Wall Street J ournal announcement, leading to a misidentification of the report date and the drift in returns the day before the announcement. 11
Graham, B., 1949, The Intelligent Investor (New York: HarperBusiness, reprinted in 2005).
Since wash sales rules would prevent an investor from selling and buying back the same stock within 30 days, there has to be some substitution among the stocks. Thus investor 1 sells stock A and investor 2 sells stock B, but when it comes time to buy back the stock, investor 1 buys stock B and investor 2 buys stock A. 12
118
CHAPTER 7 Riskless Rates and Risk Premiums All models of risk and return in finance are built around a rate that investors can make on riskless investments and the risk premium or premiums that investors should charge for investing in the average-risk investment. In the capital asset pricing model (CAPM), where there is only one source of market risk captured in the market portfolio, this risk premium becomes the premium that investors would demand when investing in that portfolio. In multifactor models, there are multiple risk premiums, each one measuring the premium demanded by investors for exposure to a specific market risk factor. This chapter examines how best to measure a riskless rate and to estimate a risk premium or premiums for use in these models. As noted in Chapter 4, risk is measured in terms of default risk for bonds, and this default risk is captured in a default spread that firms have to pay over and above the riskless rate. This chapter closes by considering how best to estimate these default spreads and the factors that may cause these spreads to change over time.
THE RISK-FREE RATE Most risk and return models in finance start off with an asset that is defined as risk free, and use the expected return on that asset as the risk-free rate. The expected returns on risky investments are then measured relative to the risk-free rate, with the risk creating an expected risk premium that is added to the risk-free rate. But what makes an asset risk free? And what do we do when we cannot find such an asset? These are the questions that will be dealt with in this section.
Requirements for an Asset to Be Risk Free Chapter 4 considers some of the requirements for an asset to be risk free. In particular, an asset is risk free if we know the expected returns on it with certainty (i.e., the actual return is always equal to the expected return). Under what conditions will the actual returns on an investment be equal to the expected returns? There are two basic conditions that have to be met. The first is that there can be no default risk. Essentially, this rules out any security issued by a private entity, since even the largest and safest ones have some measure of default risk. The only securities that have a chance of being risk free are government securities, not because governments are better run than corporations, but because they usually control the printing of currency. At least in nominal terms, they should be able to fulfill their promises. Even this assumption, straightforward though it might seem, does not always hold up, especially when governments refuse to honor claims made by previous regimes and when they borrow in currencies other than their own. There is a second condition that riskless securities need to fulfill that is often forgotten. For an investment to have an actual return equal to its expected return, there can be no reinvestment risk. To illustrate this point, assume that you are trying to estimate the expected return over a five-year period and that you want a risk-free rate. A sixmonth Treasury bill rate, while default free, will not be risk free, because there is the reinvestment risk of not knowing what the Treasury bill rate will be in six months. Even a five-year Treasury bond is not risk free, since the coupons on the bond will be reinvested at rates that cannot be predicted today. The risk-free rate for a five-year time horizon has to be the expected return on a default-free (government) five-year zero coupon bond. This clearly has painful implications for anyone doing corporate finance or valuation, where expected returns often have to be estimated for periods ranging from 1 to 10 years. A purist’s view of risk-free rates would then require different riskfree rates for each period, and different expected returns. As a practical compromise, however, it is worth noting that the present value effect of using year-specific risk-free rates tends to be small for most well-behaved term structures.1 In these cases, we could use a duration matching strategy, where the duration of the default-free security used as the risk-free asset is matched up to the duration2 of the cash flows in the analysis. If, however, there are very large differences, in either direction, between short-term and long-term rates, it does pay to stick with year-specific risk-free rates in computing expected returns.
Practical Implications When a Default-Free Entity Exists In most developed markets, where the government can be viewed as a default-free entity, at least when it comes to borrowing in the local currency, the implications are simple. When d oing investment analysis on longer-term projects or valuations, the risk-free rate should be the long-term government bond rate. If the analysis is shorterterm, the short-term government security rate can be used as the risk-free rate. The choice of a risk-free rate also has implications for how risk premiums are estimated. If, as is often the case, historical risk premiums are used, where the excess return earned by stocks over and above a government security rate over a past period is used as the risk premium, the government security chosen has to be the same one as that used for the risk-free rate. Thus, the historical risk premium used in the United States should be the excess return earned by stocks over Treasury bonds, and not Treasury bills, for purposes of long-term analysis. 119
Cash Flows and Risk-Free Rates: The Consistency Principle The risk-free rate used to come up with expected returns should be measured consistently with how the cash flows are measured. Thus, if cash flows are estimated in nominal U.S. dollar terms, the risk-free rate will be the U.S. Treasury bond rate. This also implies that it is not where a firm is domiciled that determines the choice of a risk-free rate, but the currency in which the cash flows on the firm are estimated. Thus, Nestlé can be valued using cash flows estimated in Swiss francs, discounted back at an expected return estimated using a Swiss long-term government bond rate as the riskfree rate, or it can be valued in British pounds, with both the cash flows and the risk-free rate being in British pounds. Given that the same firm can be valued in different currencies, will the final results always be consistent? If we assume purchasing power parity, then differences in interest rates reflect differences in expected inflation. Both the cash flows and the discount rate are affected by expected inflation; thus, a low discount rate arising from a low risk-free rate will be exactly offset by a decline in expected nominal growth rates for cash flows, and the value will remain unchanged. If the difference in interest rates across two currencies does not adequately reflect the difference in expected inflation in these currencies, the values obtained using the different currencies can be different. In particular, firms will be valued more highly when the currency used is the one with low interest rates relative to inflation. The risk, however, is that the interest rates will have to rise at some point to correct for this divergence, at which point the values will also converge.
Real versus Nominal Risk-Free Rates Under conditions of high and unstable inflation, valuation is often done in real terms. Effectively, this means that cash flows are estimated using real growth rates and without allowing for the growth that comes from price inflation. To be consistent, the discount rates used in these cases have to be real discount rates. To get a real expected rate of return, we need to start with a real risk-free rate. While government bills and bonds offer returns that are risk free in nominal terms, they are not risk free in real terms, since expected inflation can be volatile. The standard approach of subtracting an expected inflation rate from the nominal interest rate to arrive at a real riskfree rate provides at best an estimate of the real risk-free rate. Until recently, there were few traded default-free securities that could be used to estimate real risk-free rates, but the introduction of inflation-indexed Treasuries (TIPs) has filled this void. An inflation-indexed Treasury security does not offer a guaranteed nominal return to buyers, but instead provides a guaranteed real return. Thus, an inflation-indexed Treasury that offers a 3 percent real return will yield approximately 7 percent in nominal terms if inflation is 4 percent and only 5 percent in nominal terms if inflation is only 2 percent. The only problem is that real valuations are seldom called for or done in the United States, which has historically had stable and low expected inflation. The markets where we would most need to do real valuations, unfortunately, are markets without inflation-indexed default-free securities. The real risk-free rates in these markets can be estimated by using one of two arguments: 1. The first argument is that as long as capital can flow freely to those economies with the highest real returns, there can be no differences in real risk-free rates across markets. Using this argument, the real risk-free rate for the United States, estimated from the inflation-indexed Treasury, can be used as the real risk-free rate in any market. 2. The second argument applies if there are frictions and constraints in capital flowing across markets. In that case, the expected real return on an economy, in the long term, should be equal to the expected real growth rate, again in the long term, of that economy, for equilibrium. Thus, the real risk-free rate for a mature economy like Germany should be much lower than the real risk-free rate for a economy with greater growth potential, such as Hungary’s.
Risk-Free Rates When There Is No Default-Free Entity Our discussion, hitherto, has been predicated on the assumption that governments do not default, at least on local borrowing. There are many emerging market economies and quite a few devloped markets where this assumption might not be viewed as reasonable. Governments in these markets are perceived as capable of defaulting even on local borrowing. When this is coupled with the fact that some governments do not borrow long-term in the local currency, there are scenarios where obtaining a local risk-free rate, especially for the long term, becomes difficult. We consider four alternatives in the section following.
Local Currency Government Bond If the government issues long-term bonds denominated in the local currency and these bonds are traded, you can use the interest rates on these bonds as a starting point for estimating the risk-free rate in that currency. In early 2011, for instance, the Indian government issued 10-year rupee-denominated bonds that were trading at a yield of 8 percent. This rate, though, is not a risk-free rate, because investors perceive default risk in the Indian 120
government. To back out how much of the yield can be attributed to the default risk, we used the local currency sovereign rating3 of Ba2 assigned to India by Moody’s and estimated a default spread of 2.40 percent for that rating. 4
The resulting risk-free rate in rupees is:
It is true that this number assumes that the ratings agency is correct in its assessment of sovereign risk and that the default spread based on the rating is correct. An alternative approach to estimating default spreads that has become available in recent years is the credit default swap (CDS) market, where investors can b uy insurance against default. While there were no traded CDSs on India in early 2011, there were CDS contracts traded on approximately 60 countries in March 2011. The Brazil CDS in March 2011 was trading at 75 basis points (0.75 percent), which can be used in conjunction with the 10-year Brazilian real-denominated government bond rate of 8.25 percent to compute the risk-free rate in Brazilian real (BR):
The CDS market does provide a more dynamic and updated measure of the default spread but, being a markettraded number, is much more volatile. It also provides a dollar or euro based spread which may not apply to the local currency bonds.
Build-Up Approach There are countries where either the government does not issue bonds denominated in the local currency or these bonds do not trade. In this case, one alternative is to build up to a risk-free rate from fundamentals:
Since the risk-free rate in any currency can be written as the sum of expected inflation in that currency and the expected real rate, we can try to estimate the two components separately. To estimate expected inflation, we can start with the current inflation rate and extrapolate from that to expected inflation in the future. For the real rate, we can use the rate on the inflation-indexed U.S. Treasury bond rate, with the rationale that real rates should be the same globally. In 2011, for instance, adding the expected inflation rate of 6 percent in India to the interest rate of 1 percent on the inflation-indexed U.S. Treasury would have yielded a risk-free rate of 7 percent in Indian rupees.
Derivatives Markets Forward and futures contracts on exchange rates provide information about interest rates in the currencies involved, since interest rate parity governs the relationship between spot and forward rates. For instance, the forward rate between the Thai baht and the U.S. dollar can be written as follows:
For example, if the current spot rate is 38.10 Thai baht per U.S. dollar, the 10-year forward rate is 61.36 baht per dollar, and the current 10-year U.S. Treasury bond rate is 5 percent, the 10-year Thai risk-free rate (in nominal baht) can be estimated as follows:
Solving for the Thai interest rate yields a 10-year risk-free rate of 10.12 percent. The biggest limitation of this approach, however, is that forward rates are difficult to come by for periods beyond a year5 for many of the emerging markets, where we would be most interested in using them.
Risk-Free Rate Conversion If the only reason for differences in risk-free rates in different currencies is expected inflation, you can convert the risk-free rate in a mature market currency (U.S. dollars, euros) into a risk-free rate in an emerging market currency, using differences in inflation across currencies.
121
For example, assume that the risk-free rate in U.S. dollars is 4 percent and that the expected inflation rate in Indonesian rupiah is 11 percent (compared to the 2 percent inflation rate in U.S. dollars). The Indonesian risk-free rate can be written as follows:
To make this conversion, we still have to estimate the expected inflation in the local currency and the mature market currency. What if none of these choices listed work? In other words, what if the government has no local currency bonds outstanding, there are no forward or futures contract on the currency, and/or expected inflation in the local currency is difficult to estimate? Faced with these problems, it is best to switch and do your valuation in a different currency. Thus, rather than value a Nigerian or Vietnamese company in the local currency, you would value it in euros or dollars. You will still have to estimate expected exchange rates in the future in order to convert local currency cash flows to foreign currency cash flows, but that may be a more manageable exercise.
EQUITY RISK PREMIUM The notion that risk matters, and that riskier investments should have a higher expected return than safer investments to be considered good investments, is intuitive. Thus, the expected return on any investment can be written as the sum of the risk-free rate and an extra return to compensate for the risk. The disagreement, in both theoretical and practical terms, remains on how to measure this risk, and how to convert the risk measure into an expected return that compensates for risk. This section looks at the estimation of an appropriate equity risk premium (ER P) to use in risk and return models, in general, and in the capital asset pricing model, in particular.
Competing Views on Risk Premiums In Chapter 4, we considered several competing models of risk ranging from the capital asset pricing model to multifactor models. Notwithstanding their different conclusions, they all share some common views about risk. First, they all define risk in terms of variance in actual returns around an expected return; thus, an investment is riskless when actual returns are always equal to the expected return. Second, they all argue that risk has to be measured from the perspective of the marginal investor in an asset, and that this marginal investor is well diversified. Therefore, the argument goes, it is only the risk that an investment adds on to a diversified portfolio that should be measured and compensated. In fact, it is this view of risk that leads models of risk to break the risk in any investment into two components. There is a firm-specific component that measures risk that relates only to that investment or to a few investments like it, and a market component that contains risk that affects a large subset or all investments. It is the latter risk that is not diversifiable and should be rewarded. While all risk and return models agree on this fairly crucial distinction, they part ways when it comes to how to measure this market risk. Table 7.1 summarizes four models and the way each model attempts to measure risk. Table 7.1 Comparing Risk and Return Models Model
Assumptions Capital asset pricing model (CAPM) There are no transaction costs or private information. Therefore, the diversified portfolio includes all traded investments, held in proportion to their market value. Investments with the same exposure to market risk Arbitrage pricing model (APM) have to trade at the same price (no arbitrage). There is the same no-arbitrage assumption as with Multifactor model the APM. Over very long periods, higher returns on Proxy model investments must be compensation for higher market risk.
Measure of Market Risk Beta measured against this market portfolio
Betas measured against multiple (unspecified) market risk factors Betas measured against multiple specified macroeconomic factors Proxies for market risk, for example, include market capitalization and price book value ratios.
In the first three models, the expected return on any investment can be written as:
where βj = Beta of investment relative to factor j 122
Risk premiumj = Risk premium for factor j Note that in the special case of a single-factor model, like the CAPM, each investment’s expected return will be determined by its beta relative to the risk premium for that factor. Assuming that the risk-free rate is known, these models all require two inputs. The first is the beta or betas of the investment being analyzed, and the second is the appropriate risk premium(s) for the factor or factors in the model. The issue of beta estimation is examined in the next chapter, this section concentrates on the measurement of the risk premium.
What We Would Like to Measure As far as the risk premium is concerned, we would like to know for each factor what investors, on average, require as a premium over the risk-free rate for an investment with average risk. Without any loss of generality, let us consider the estimation of the beta and the equity risk premium in the capital asset pricing model. Here, the risk premium should measure what investors, on average, demand as extra return for investing in the market portfolio relative to the risk-free asset.
Historical Risk Premiums In practice, we usually estimate the risk premium by looking at the historical premium earned by stocks over default-free securities over long time periods. The historical premium approach is simple. The actual returns earned on stocks over a long time period are estimated, and then compared to the actual returns earned on a default-free (usually government) security. The difference, on an annual basis, between the two returns is computed and represents the historical risk premium. This approach might yield reasonable estimates in markets like the United States, with a large and diversified stock market and a long history of returns on both stocks and government securities. However, they yield meaningless estimates for the risk premiums in other countries, where the equity markets represent a small proportion of the overall economy and the historical returns are available only for short periods. While users of risk and return models may have developed a consensus that historical premium is, in fact, the best estimate of the risk premium looking forward, there are surprisingly large differences in the actual premiums we observe being used in practice. For instance, the risk premium estimated in the U.S. markets by different investment banks, consultants, and corporations range from 3 percent at the lower end to 12 percent at the upper end. Given that they almost all use the same database of historical returns, provided by Ibbotson Associates,6 summarizing data from 1926, these differences may seem surprising. There are, however, three reasons for the divergence in risk premiums: Table 7.2 Standard Errors in Risk Premium Estimates Estimation Period Standard Error of Risk Premium Estimate 5 years 10 years 25 years 50 years
1. Time period used. While there are many who use all the data going back to 1926 (or earlier), there are almost as many using data over shorter time periods, such as 50, 20, or even 10 years, to come up with historical risk premiums. The rationale presented by those who use shorter periods is that the risk aversion of the average investor is likely to change over time, and that using a shorter time period provides a more updated estimate. This has to be offset against a cost associated with using shorter time periods, which is the greater noise in the risk premium estimate. In fact, given the annual standard deviation in stock prices7 between 1926 and 2010 of 20 percent, the standard error8 associated with the risk premium estimate can be estimated for different estimation periods in Table 7.2. Note that to get reasonable standard errors, we need very long time periods of historical returns. Conversely, the standard errors from 10-year and 20-year estimates are likely to be almost as large as or larger than the actual risk premium estimated. Thi s cost of using shorter time periods seems, in our view, to overwhelm any advantages associated with getting a more updated premium. 2. Choice of risk-free security. The Ibbotson database reports returns on both Treasury bills (T-bills) and Treasury bonds (T-bonds), and the risk premium for stocks can be estimated relative to each. Given that the yield curve in the United States has been upward-sloping for most of the past seven decades, the risk premium is larger when estimated relative to shorter-term government securities (such as Treasury bills). The risk-free rate chosen in computing the premium has to be consistent with the risk-free rate used to compute expected returns. Thus, if the Treasury bill rate is used as the risk-free rate, the premium has to be the premium earned by stocks over that rate. 123
If the Treasury bond rate is used as the risk-free rate, the premium has to be estimated relative to that rate. For the most part, in corporate finance and valuation, the risk-free rate will be a long-term default-free Treasury (government) bond rate and not a Treasury bill rate. Thus, the risk premium used should be the premium earned by stocks over Treasury bonds. 3. Arithmetic and geometric averages. The final sticking point when it comes to estimating historical premiums relates to how the average returns on stocks, Treasury bonds, and Treasury bills are computed. The arithmetic average return measures the simple mean of the series of annual returns, whereas the geometric average looks at the compounded return.9 Conventional wisdom argues for the use of the arithmetic average. In fact, if annual returns are uncorrelated over time, and our objective were to estimate the risk premium for the next year, the arithmetic average is the best unbiased estimate of the premium. In reality, however, there are strong arguments that can be made for the use of geometric averages. First, empirical studies seem to indicate that returns on stocks are negatively correlated over time.10 Consequently, the arithmetic average return is likely to overstate the premium. Second, while asset pricing models may be single-period models, the use of these models to get expected returns over long periods (such as 5 or 10 years) suggests that we are interested in returns over longer periods. In this context, the argument for geometric average premiums becomes even stronger. In summary, the risk premium estimates vary across users because of differences in time periods used, the choice of Treasury bills or bonds as the risk-free rate and the use of arithmetic averages as opposed to geometric averages. The effect of these choices is summarized in Table 7.3, which uses returns from 1928 to 2010. Note that the premiums can range from –4.11 percent to 7.62 percent, depending on the choices made. In fact, these differences are exacerbated by the fact that many risk premiums that are in use today were estimated using historical data three, four, or even 10 years ago. If forced to choose an equity risk premium on this table, we would be inclined to go with 4.31 percent, the geometric average risk premium for stocks over Treasury bonds from 1928 to 2010. Table 7.3 Histo rical Risk Premiums for the United States
histretSP.xls: There is a dataset on the Web that summar izes historical returns on stocks, T-bonds, and T-bills in the United States going back to 1928.
Historical Risk Premiums: Other Markets If it is difficult to estimate a reliable historical premium for the U.S. market, it becomes doubly so when looking at markets with short and volatile histories. This is clearly true for emerging markets, but it is also true for the European equity markets. While the economies of Germany, Italy, and France may be mature, their equity markets do not share the same characteristic. Until two decades ago, they tend to be dominated by a few large companies; many businesses remain private; and trading was, except on a few stocks. There are some practitioners who still use historical premiums for these markets. To capture some of the danger in this practice, Table 7.4 summarizes historical risk premiums11 for major non-U.S. markets for 1970 to 2010. Table 7.4 Historical Equity Risk Premiums: Markets outside the United States
124
Note that some of the countries have very low historical risk premiums, and a few oth ers have high risk premiums. Before an attempt is made to come up with a rationale for why this might be so, it is worth noting that th e standard error on each and every one of these estimates is high, notwithstanding the fact that the premiums are estimated over a very long time period.
HISTORICAL RISK PREMIUM APPROACH: SOME CAVEATS Given how widely the historical risk premium approach is used, it is surprising how flawed it is and how little attention these flaws have attracted. Consider first the underlying assumption that investors’ risk premiums have not changed over time and that the average risk investment (in the market portfolio) has remained stable over the period examined. We would be hard-pressed to find anyone who would be willing to sustain this argument with fervor. The obvious fix for this problem, which is to use a more recent time period, runs directly into a second problem, which is the large standard error associated with risk premium estimates. Though these standard errors may be tolerable for very long time periods, they clearly are unacceptably high when shorter periods are used. Finally, even if there is a sufficiently long time period of history available and investors’ risk aversion has not changed in a systematic way over that period, there is a final problem. Markets that exhibit this characteristic, and let us assume that the U.S. market is one such example, represent so-called survivor markets. In other words, assume that one had invested in the 10 largest equity markets in the world in 1928, of which the United States was one. In the period extending from 1928 to 2010, investments in few of the other equity markets would have earned as large a premium as the U.S. equity market, and some of them (like Austria) would have resulted in investors earning little or even negative returns over the period. Thus, the survivor bias will result in historical premiums that are larger than expected premiums for markets like the United States, even assuming that investors are rational and factor risk into prices.
If the standard errors on these estimates are high, consider how much more noise there is in estimates of historical risk premiums for emerging market equity markets, which often have a reliable history of 10 years or less and very large standard deviations in annual stock returns. Historical risk premiums for emerging markets may provide for interesting anecdotes, but they clearly should not be used in risk and return models.
Modified Historical Risk Premium While historical risk premiums for markets outside the United States cannot be used in risk models, we still need to estimate a risk premium for use in these markets. To approach this estimation question, let us start with the basic proposition that the risk premium in any equity market can be written as:
The country premium could reflect the extra risk in a specific market. This boils down our estimation to answering two questions: 1. What should the base premium for a mature equity market be? 2. Should there be a country premium, and if so, how do we estimate the premium? To answer the first question, one can argue that the U.S. equity market is a mature market and that there is sufficient historical data in the United States to make a reasonable estimate of the risk premium. In fact, reverting back to our discussion of historical premiums in the U.S. market, we will use the geometric average premium earned by stocks over Treasury bonds of 4.31 percent between 1928 and 2010. We have chosen the long time 125
period to reduce standard error, the Treasury bond to be consistent with our choice of a risk-free rate, and the geometric averages to reflect our desire for a risk premium that we can use for longer-term expected returns. On the issue of country premiums, there are some who argue that country risk is diversifiable and that there should be no country risk premium. After looking at the basis for their argument, and then considering the alternative view that there should be a country risk premium, we present approaches for estimating country risk premiums, one based on country bond default spreads and one based on equity market volatility.
Should There Be a Country Risk Premium? Is there more risk in investing in a Malaysian or Brazilian stock than there is in investing in the United States? The answer, to most, seems to be obviously affirmative. That, however, does not answer the question of whether there should be an additional risk premium charged when investing in those markets. Note that the only risk that is relevant for purposes of estimating a cost of equity is market risk or risk that cannot be diversified away. The key question then becomes whether the risk in an emerging market is diversifiable or nondiversifiable risk. If, in fact, the additional risk of investing in Malaysia or Brazil can be diversified away, then there should be no additional risk premium charged. If it cannot, then it makes sense to think about estimating a country risk premium. But diversified away by whom? Equity in a Brazilian or Malaysian firm can be held by hundreds or thousands of investors, some of whom may hold only domestic stocks in their portfolio, whereas others may have more global exposure. For purposes of analyzing country risk, we look at the marginal investor—the investor most likely to be trading on the equity. If that marginal investor is globally diversified, there is at least the potential for global diversification. If the marginal investor does not have a global portfolio, the likelihood of diversifying away country risk declines substantially. Stulz (1999) made a similar point using different terminology. He differentiated between segmented markets, where risk premiums can be different in each market because investors cannot or will not invest outside their domestic markets, and open markets, where investors can invest across markets. In a segmented market, the marginal investor will be diversified only across investments in that market, whereas in an open market, the marginal investor has the opportunity (even if he or she does not take it) to invest across markets. Even if the marginal investor is globally diversified, there is a second test that has to be met for country risk not to matter. All or much of country risk should be country specific. In other words, there should be low correlation across markets. Only then will the risk be diversifiable in a globally diversified portfolio. If, however, the returns across countries have significant positive correlation, country risk has a market risk component, is not diversifiable, and can command a premium. Whether returns across countries are positively correlated is an empirical question. Studies from the 1970s and 1980s suggested that the correlation was low, and this was an impetus for global diversification. Partly because of the success of that sales pitch and partly because economies around the world have become increasingly intertwined over the past decades, more recent studies indicate that the correlation across markets has risen. This is borne out by the speed with which troubles in one market, say Russia, can spread to a market with little or no obvious relationship to it, say Brazil. So where do we stand? We believe that while the barriers to trading across markets have dropped, investors still have a home bias in their portfolios and that markets remain partially segmented. While globally diversified investors are playing an increasing role in the pricing of equities around the world, the resulting increase in correlation across markets has resulted in a portion of country risk being nondiversifiable or market risk. The next section considers how best to measure this country risk and build it into expected returns.
Measuring Country Risk Premiums If country risk matters and leads to higher premiums for riskier countries, the obvious follow-up question becomes how we measure this additional premium. This section looks at three approaches. The first builds on default spreads on country bonds issued by each country, whereas the second uses equity market volatility as its basis. The third is a melded approach that uses both default spreads and equity market volatility.
Default Risk Spreads While there are several measures of country risk, one of the simplest and most easily accessible is the rating assigned to a country’s debt by a ratings agency; Standard & Poor’s (S&P), Moody’s Investors Service, and Fitch all rate countries. These ratings measure default risk (rather than equity risk), but they are affected by many of the factors that drive equity risk—the stability of a country’s currency, its budget and trade balances, and its political stability, for instance.12 The other advantage of ratings is that they can be used to estimate default spreads over the riskless rate. The default spreads are estimated by comparing the dollar- and euro-denominated bonds issued by governments that share a sovereign rating. For instance, a 10-year dollar-denominated bond issued by the Peruvian government, rated Baa by Moody’s, traded at an interest rate of 5.2% in January 2011—a 1.7% spread over the U.S. Treasury bond rate of 3.5% at the time. Since there can be country-specific factors that cause these rates to 126
vary across bonds, we average the spreads within each ratings class to arrive at the average spread for each sovereign rating. A cross six countries rated Baa in January 2011, for example, the average default spread was 2.00%. Thus, any country with a Baa rating would be assigned a spread of 2.00% in January 2011. The perils with sovereign ratings have been documented over the past few years. Specifically, ratings agencies seem to lag markets in responding to changes in country risk. An alternative approach to estimating default spreads is the CDS market. As we noted previously in the section on risk-free rates, in early 2011 there were CDS instruments traded in about 60 countries, providing an updated market measure of default risk. In January 2011, for instance, the CDS spread for Peru was 160 basis points (1.6 percent), close to the default spread estimated from the dollar-denominated bond. Table 7.5 summarizes default spreads estimated for selected countries in January 2011, using both the sovereign rating and CDS approaches. Table 7.5 Default Spreads for Selected Countries—January 2011
Analysts who use default spreads as measures of country risk typically add them on to the cost of both equity and debt of every company traded in that country. For instance, the cos t of equity for a Peruvian company, estimated in U.S. dollars, will be 2.00 percent higher than the cost of equity of an otherwise similar U.S. company. If we assume that the risk premium for the United States and other mature equity markets is 4.31 percent, the cost of equity for a Peruvian company with a beta of 1.2 can be estimated as follows (with a U.S. Treasury bond rate of 3.5 percent).
In some cases, analysts add the default spread to the U.S. risk premium and multiply the total risk premium by the beta. This increases the cost of equity for high-beta companies and lowers it for low-beta firms. While ratings provide a convenient measure of country risk, there are costs associated with using them as the only measure. First, ratings agencies often lag markets when it comes to responding to changes in the underlying default risk. Second, the ratings agency focus on default risk may obscure other risks that could still affect equity markets. What are the alternatives? There are numerical country risk scores that have been developed by some services as much more comprehensive measures of risk. The Economist, for instance, has a score that runs from 0 to 100 (where 0 is no risk, and 100 is most risky) that it uses to rank emerging markets. Alternatively, country risk can be estimated from the bottom up by looking at economic fundamentals in each country. This, of course, requires significantly more information than the other approaches. Finally, default spreads measure the risk associated with bonds issued by countries and not the equity risk in these countries. Since equities in any market are likely to be more risky than bonds, you could argue that default spreads understate equity risk premiums.
Relative Standard Deviations There are some analysts who believe that investors in equity markets choose among these markets based on their assessed riskiness and that the risk premiums should reflect the differences in equity risk. A conventional measure of equity risk is the standard deviation in stock prices; higher standard deviations are generally associated with more risk. If you scale the standard deviation of one market against another, you obtain a measure of relative risk.
127
This relative standard deviation, when multiplied by the premium used for U.S. stocks, should yield a measure of the total risk premium for any market.
Assume, for the moment, that you are using a mature market premium for the United States of 4.31 percent and that the annual standard deviation of U.S. stocks is 20 percent. If the annual standard deviation of Indonesian stocks is 35 percent, the estimate of a total risk premium for Indonesia would be:
The country risk premium can be isolated as follows:
While this approach has intuitive appeal, there are problems with using standard deviations computed in markets with widely different market structures and liquidity. There are very risky emerging markets that have low standard deviations for their equity markets because the markets are illiquid. This approach will understate the equity risk premiums in those markets. The second problem is related to currencies, since the standard deviations are usually measured in local currency terms; the standard deviation in the U.S. market is a dollar standard deviation, whereas the standard deviation in the Indonesian market is a rupiah standard deviation. This is a relatively simple problem to fix, though, since the standard deviations can be measured in the same currency—you could estimate the standard deviation in dollar returns for the Indonesian market.
THE DANGER OF DOUBLE COUNTING RISK When assessing country risk, there is a substantial chance that the same risk may be counted more than once in a valuation. For instance, there are analysts who use the dollar-denominated bonds issued by a country—the Brazilian dollar bond, for instance—as the risk-free rate when estimating cost of equity for Brazilian companies. The interest rate on this bond already incorporates the default spreads discussed in the preceding section. If the risk premium is also adjusted upward to reflect country risk, there has been a double counting of the risk. This effect is made worse when betas are adjusted upward and cash flows are adjusted downward (a process called “haircutting”) because of country risk.
Default Spreads + Relative Standard Deviations The country default spreads that come with country ratings provide an important first step, but still only measure the premium for default risk. Intuitively, we would expect the country equity risk premium to be larger than the country default risk spread. To address the issue of how much higher, one can look at the volatility of the equity market in a country relative to the volatility of the country bond used to estimate the spread. This yields the following estimate for the country equity risk premium:
To illustrate, consider the case of Brazil. In January 2011, Brazil was rated Baa3 by Moody’s, resulting in a default spread of 2.00 percent. The annualized standard deviation in the Brazilian equity index over the previous two years was 17.65 percent, while the annualized standard deviation in the Brazilian dollar- denominated bond was 7.32 percent. The resulting country equity risk premium for Brazil is as follows:
Note that this country risk premium will increase if the country rating drops or if the relative volatility of the equity market increases. Adding this premium to the mature market (U.S.) premium of 4.31 percent would yield a total equity risk premium for Brazil of 9.13 percent. Why should equity risk premiums have any relationship to country bond spreads? A simple explanation is that an investor who can make 11 percent on a dollar-denominated Brazilian government bond would not settle for an expected return of 10.5 percent (in dollar terms) on Brazilian equity. Playing devil’s advocate, however, a critic could argue that the interest rate on a country bond, from which default spreads are extracted, is not really an expected return since it is based on the promised cash flows (coupon and principal) on the bond rather than the expected cash flows. In fact, if we wanted to estimate a risk premium for bonds, we would need to estimate the expected return based on expected cash flows, allowing for the default risk. This would result in a much lower default spread and equity risk premium. 128
Both this approach and the previous one use the standard deviation in equity of a market to make a judgment about country risk premium, but they measure it relative to different bases. This approach uses the country bond as a base, whereas the previous one uses the standard deviation in the U.S. market. This approach assumes that investors are more likely to choose between Brazilian bonds and Brazilian equity, whereas the previous approach assumes that the choice is across equity markets.
Choosing among the Approaches The three approaches to estimating country risk premiums will generally give you different estimates, with the bond default spread and relative equity standard deviation approaches yielding lower country risk premiums than the melded approach that uses both the country bond default spread and the equity standard deviation. We believe that the larger country risk premiums that emerge from the last approach are the most realistic for the immediate future, but that country risk premiums will change over time. Just as companies mature and become less risky over time, countries can mature and become less risky as well. One way to adjust country risk premiums over time is to begin with the premium that emerges from the melded approach and to adjust this premium down toward either the country bond default spread or the country premium estimated from equity standard deviations. Another way of presenting this argument is to note that the differences between standard deviations in equity and bond prices narrow over longer periods, and the resulting relative volatility will generally be smaller.13 Thus, the equity risk premium will converge on the country bond spread as we look at longer-term expected returns. For example, the country risk premium for Brazil would be 4.82 percent for the next year but decline over time to either the 2.00 percent (country default spread) or even lower.
Estimating Asset Exposure to Country Risk Premiums Once country risk premiums have been estimated, the final question that has to be addressed relates to the exposure of individual companies within that country to cou ntry risk. There are three alternative views of country risk: 1. Assume that all companies in a country are equally exposed to country risk. Thus, for Brazil, with its estimated country risk premium of 4.82 percent, each company in the market will have an additional country risk premium of 4.82 percent added to its expected returns. For instance, the cost of equity for Petrobras, an integrated oil company listed in Brazil with a beta of 0.80, in U.S. dollar terms would be (assuming a U.S. Treasury bond rate of 3.50 percent and a mature market [U.S.] risk premium of 4.31 percent):
Note that the risk-free rate used is the U.S. Treasury bond rate and that the 4.31 percent is the equity risk premium for a mature equity market (estimated from historical data in the U.S. market). The biggest limitation of this approach is that it assumes that all firms in a country, no matter what their business or size, are equally exposed to country risk. To convert this dollar cost of equity into a cost of equity in the local currency, all that we need to do is to scale the estimate by relative inflation. To illustrate, if the Brazilian inflation rate is 6 percent and the U.S. inflation rate is 2 percent, the cost of equity for Petrobras in Brazilian real (BR) terms can be written as:
This will ensure consistency across estimates and valuations in different currencies. 2. Assume that a company’s exposure to country risk is proportional to its exposure to all other market risk, which is measured by the beta. For Petrobras, this would lead to a cost of equity estimate of:
This approach does differentiate between firms, but it assumes that betas that measure exposure to all other market risk measure exposure to country risk as well. Thus, low-beta companies are less exposed to country risk than high-beta companies. 3. The most general approach, and our preferred approach, is to allow for each company to have an exposure to country risk that is different from its exposure to all other market risk. Measuring this exposure with λ, the cost of equity for any firm is estimated as follows:
How can we best estimate λ? This question is considered in far more detail in the next chapter, but we would argue that commodity companies that get most of their revenues in U.S. dollars14 by selling into a global market should be less exposed than manufacturing companies that service the local market. Using this rationale, Petrobras, which derives most of its revenues in the global oil market in U.S. dollars, should be less exposed than the typical Brazilian 129
firm to country risk.15 Using a λ of 0.50, for instance, we get a cost of equity in U.S. dollar terms for Petrobras of:
Note that the third approach essentially converts our expected return model to a two-factor model, with the second factor being country risk, with λ measuring exposure to country risk. This approach also seems to offer the most promise in analyzing companies with exposures in multiple countries like Coca-Cola and Nestlé. While these firms are ostensibly developed market companies, they have substantial exposure to risk in emerging markets, and their costs of equity should reflect this exposure. We could estimate the country risk premiums for each country in which they operate and a λ relative to each country, and use these to estimate a cost of equity for either company.
ctryprem.xls: There is a dataset on the Web that contains the updated ratings for countries and the risk premiums associated with each.
Alternative Approach: Implied Equity Premiums There is an alternative to estimating risk premiums that does not require historical data or corrections for country risk, but does assume that the market, overall, is correctly priced. Consider, for instance, a very simple valuation model for stocks:
This is essentially the present value of dividends growing at a constant rate. Three of the four inputs in this model can be obtained externally—the current level of the market (value), the expected dividends next period, and the expected growth rate in earnings and dividends in the long term. The only unknown is then the required return on equity; when we solve for it, we get an implied expected return on stocks. Subtracting the risk-free rate will yield an implied equity risk premium. To illustrate, assume that the current level of the S&P 500 index is 900, the expected dividend yield on the index is 2 percent, and the expected growth rate in earnings and dividends in the long term is 7 percent. Solving for the required return on equity yields the following:
Solving for r,
If the current risk-free rate is 6 percent, this will yield a premium of 3 percent. This approach can be generalized to allow for high growth for a period, and extended to cover cash flow–based, rather than dividend, models. To illustrate this, consider the S&P 500 index as of January 1, 2011. The index was at 1257.64, and the cash flows from dividends and stock buybacks during the course of 2010 were 53.96. In addition, the consensus estimate16 of growth in earnings for companies in the index was approximately 6.95 percent for the next five years. Since this is not a growth rate that can be sustained forever, we employ a two-stage valuation model, where we allow growth to continue at 6.95 percent for five years, and then lower the growth rate to the Treasury bond rate of 3.29 percent after that.17 The following table summarizes the expected cash flows for the next five years of high growth and the first year of stable growth thereafter: Year Cash Flow on Index 1 2
53.96(1.0695) = 57.72 57.72(1.0695) = 61.73
3
61.73(1.0695) = 66.02
4
66.02(1.0695) = 70.60
5
70.60(1.0695) = 75.51
6
75.51(1.0329) = 77.99
If we assume that these are reasonable estimates of the cash flows and that the index is correctly priced, then:
130
Note that the last term in the equation is the terminal value of the index, based on the stable growth rate of 3.29 percent, discounted back to the present. Solving for r in this equation yields us the required return on equity of 8.49 percent. Netting out the Treasury bond rate of 3.29 percent yields an implied equity premium of 5.20 percent. The advantage of this approach is that it is market-driven and current, and does not require any historical data. Thus, it can be used to estimate implied equity premiums in any market. It is, however, bounded by whether the model used for the valuation is the right one and by the availability and reliability of the inputs to that model. For instance, the equity risk premium for the Brazilian market on September 30, 2009, was estimated from the following inputs. The index (Bovespa) was at 61,172, and the aggregate cash flow yield on the index was 4.95 percent. Earnings in companies in the index are expected to grow 6 percent (in U.S. dollar terms) over the next five years, and 3.45 percent thereafter. These inputs yield a required return on equity of 9.17 percent, which when compared to the U.S. Treasury bond rate of 3.45 percent on that day results in an implied equity premium of 5.72 percent. For simplicity, we have used nominal dollar expected growth rates18 and Treasury bond rates, but this analysis could have been done entirely in the local currency. The implied equity premiums change over time as stock prices, earnings, and interest rates change. In fact, the contrast between these premiums and the historical premiums is best illustrated by graphing out the implied premiums in the S&P 500 going back to 1960 in Figure 7.1. In terms of mechanics, smoothed historical growth rates in earnings and dividends were used as projected growth rates, and a two-stage dividend discount model was used. Looking at these numbers, the following conclusions would be drawn: Figure 7.1 Implied Premiums for U.S. Equity Market (1960 to 2010)
The arithmetic averag e historical risk premium, which is used by many practitioners, has been higher than the implied premium over almost the entire 50-year period (with 2009 the only exception). The geometric premium does provide a more interesting mix of results, with implied premiums exceeding historical premiums in the mid-1970s and again following 2008. The implied equity premium did increase during the 1970s, as inflation increased. This does have interesting implications for risk premium estimation. Instead of assuming that the risk premium is a constant and unaffected by the level of inflation and interest rates, which is what we do with historical risk premiums, it may be more realistic to increase the risk premium if expected inflation and interest rates go up. While historical risk premiums have generally drifted down for the last few decades, there is a strong tendency toward mean reversion in implied equity premiums. Thus, the premium, which peaked at 6.5 percent in 1978, moved down toward the average in the 1980s. By the same token, the premium of 2 percent that we observed at the end of the dot-com boom in the 1990s quickly reverted back to the average during the market correction from 2000 to 2003. Given this tendency, it is possible that we can end up with a far better estimate of the implied equity premium by looking at not just the current premium, but also at historical trend lines. Finally, the crisis of 2008 was unprecedented in terms of its impact on equity risk premiums. Implied equity risk premiums rose more during 2008 than in any one of the prior 50 years. Much of that change occurred in the last 15 weeks of the year when the US and other developed markets went through gyrations more typical of emerging markets. A large portion of that increase dissipated in 2009 but returned again in 2010 and 2011. As a final point, there is a strong tendency toward mean reversion in financial markets. Given this tendency, it is possible that we can end up with a far better estimate of the implied equity premium by looking not just at the current premium but also at historical data. There are two ways in which we can do this: 1. We can use the average implied equity premium over longer periods, say 10 to 15 years. Note that we do not 131
need as many years of data here as we did with the historical premium estimate, because the standard errors tend to be smaller. 2. A more rigorous approach would require relating implied equity risk premiums to fundamental macroeconomic data over the period. For instance, given that implied equity premiums tend to be higher during periods with higher inflation rates (and interest rates), we ran a regression of implied equity premiums against Treasury bond rates between 1960 and 2010:
The regression has some explanatory power, with an R-squared of 15 percent, and the t statistics (in brackets under the coefficients) indicate the statistical significance of the independent variable used. Substituting the current Treasury bond rate into this equation should yield an updated estimate19 of the implied equity premium.
histimpl.xls: This dataset on the Web shows the inputs used to calculate the premium in each year for the U.S. market.
HI STORICAL VERSUS IMPLIED EQUITY PREMIUMS: EFFECT OF MARKET VIEWS As you can see from the preceding discussion, historical premiums can be very different from implied equity premiums. At the end of 2000, the historical risk premium for stocks over bonds in the United States was 5.51 percent, whereas the implied equity risk premium was 2.87 percent. In contrast, at the end of 2008, the historical risk premium was 3.88%, whereas the implied premium was 6.43%. When doing discounted cash flow valuation, you have to decide which risk premium you will use in the valuation, and your choice will be determined by both your market views and your valuation mission.
Market Views: If you believe that the market is right in the aggregate, though it may make mistakes on individual stocks, the risk premium you should use is the current implied equity risk premium. If you believe that the market often makes mistakes in the aggregate and that risk premiums in markets tend to move back to historical norms (mean reversion), you should go with the historical premium (4.31 percent at the end of 2000). A way to split the difference is to assume that markets are right across time, though they may make mistakes at individual points in time. If you make this assumption, you should use an average implied equity risk premium over time. The average implied equity risk premium from 1960 to 2010 is 3.95 percent. While this book uses the historical premium a few times in its valuations, it sticks with the average implied premium in most of the valuations. Valuation mission: If your valuation requires you to be market neutral, you should use the current implied equity risk premium. This is often the case if you are an equity research analyst or if you have to value a company for an acquisition.
implprem.xls: This spreadsheet allows you to estimate the implied equity premium in a market.
DEFAULT SPREADS ON BONDS The interest rates on bonds are determined by the default risk that investors perceive in the issuer of the bonds. This default risk is often measured with a bond rating, and the interest rate that corresponds to the rating is estimated by adding a default spread to the riskless rate. In Chapter 4, we examined the process used by rating agencies to rate firms. This chapter considers how to estimate default spreads for a given ratings class and why these spreads may change over time.
Estimating Default Spreads The simplest way to estimate default spreads for each ratings class is to find a sampling of bonds within that ratings class and obtain the current market interest rate on these bonds. Why do we need a sampling rather than just one bond? A bond can be misrated or mispriced. Using a sample reduces or eliminates this problem. In obtaining this sample, you should try to focus on the most liquid bonds with as few special features attached to them as possible. Corporate bonds are often illiquid and the interest rates on such bonds may not reflect current market rates. The presence of special features on bonds such as convertibility can affect the pricing of these bonds and consequently the interest rates estimated on them. Once a sample of bonds within each ratings class has been identified, you need to estimate the interest rate on these bonds. There are two measures that are widely used. The first is the yield on the bond, which is the coupon rate divided by the market price. The second is the yield to maturity on the bond, which is the interest rate that makes the present value of the coupons and face value of the bond equal to the market price. In general, it is the yield to maturity that better measures the market interest rate on the bond. 132
Having obtained the interest rates on the bonds in the sample, you have two decisions to make. The first relates to weighting. You could compute a simple average of the interest rates of the bonds in the sample or a weighted average, with the weights based upon the trading volume—more liquid bonds will be weighted more than less liquid bonds. The second relates to the index Treasury rate, since the average interest rate for a ratings class is compared to this rate to arrive at a default spread. In general, the maturity of the Treasury should match the average maturity of the corporate bonds chosen to estimate the average interest rate. Thus, the average interest rate for five-year BBB-rated corporate bonds should be compared to the average interest rate for five-year Treasuries to derive the spread for the BBB-rated bonds. Publications like Barron’s have historically provided interest rates on at least higher-rated bonds (BBB or higher), an increasing number of online services provide the same information today for all rated bonds. Table 7.6 is extracted from one such online service in early 2011 for 10-year bonds using a 10-year T Bond rate of 3.5% as the riskfree rate. Table 7.6 Default Spreads and Interest Rates — January 2011 Source: BondsOnline.com.
Bond Rating Default Spread Interest Rate on Debt AAA
0.50%
4.00%
AA
0.65%
4.15%
A+
0.85%
4.35%
A
1.00%
4.50%
A–
1.10%
4.60%
BBB
1.60%
5.10%
BB+
3.00%
6.50%
BB
3.35%
6.85%
B+
3.75%
7.25%
B
5.00%
8.50%
B–
5.25%
8.75%
CCC
8.00%
11.50%
CC
10.00%
13.50%
C
12.00%
15.50%
D
15.00%
18.50%
Determinants of Default Spreads Table 7.6 provides default spreads at a point in time, but default spreads not only vary across time, but they also can vary for bonds with the same rating but different maturities. This section considers how default spreads vary across time and for bonds with varying maturities.
Default Spreads and Bond Maturity Empirically, the default spread for corporate bonds of a given ratings class seems to increase with the maturity of the bond. Figure 7.2 presents the default spreads estimated for Aaa, Baa, and Caa-rated bonds for maturities ranging from 1 to 30 years in January 2011. Figure 7.2 Default Spreads by Maturity—January 2011 Source: BondsOnline.com.
133
At least there seems to be no perceptible pattern to default spreads over maturity in 2011. Thus, the default spread on the 10-year bond is not noticeably larger than the default spread on a 1-year bond. This has not always been the case. There have been some p eriods in history where default spreads were an increasing function of maturity and other periods where they were a decreasing function.
Default Spreads over Time The default spreads presented in Table 7.6, after a year of declining markets and a slowing economy, were significantly higher than the default spreads a year earlier. This phenomenon is not new. Historically, default spreads for every ratings class have increased during recessions and decreased during economic booms. Figure 7.3 graphs the spread between 10-year Moody’s Baa-rated bonds and the 10-year Treasury bond rate each year from 1960 to 2010 and contrasts it with the implied equity risk premium each year. The default spreads did increase during periods of low economic growth; note the increase during 1973–1974 and 1979–1981 in particular. Although default spreads and equity risk premiums in most periods have generally moved in tandem, there have been exceptional periods when they moved in different directions. In the late 1990s, for instance, the dot-com boom in stock prices resulted in declining equity risk premiums, while default spreads stayed relatively stable. In contrast, the subprime boom in 2004 to 2007 lowered default spre ads, while equity risk premiums stayed unchanged. Figure 7.3 Baa Bond Default Spread and Implied Equity Risk Premiums: 1960 to 2010
ratings.xls: This dataset on the Web summarizes default spreads by bond ra ting class for the most recent period.
CONCLUSION The risk-free rate is the starting point for all expected return models. For an asset to be risk free, it has to be free of both default and reinvestment risk. Using these criteria, the appropriate risk-free rate to use to obtain expected returns should be a default-free (government) zero coupon rate that is matched up to when the cash flow that is being discounted occurs. In practice, however, it is usually appropriate to match up the duration of the risk-free 134
asset to the duration of the cash flows being analyzed. In valuation, this will lead us toward long-term government bond rates as risk-free rates. It is also important that the risk-f ree rate be consistent with the cash flows being discounted. In particular, the currency in which the risk-free rate is denominated and whether it is a real or nominal risk-free rate should be determined by the currency in which the cash flows are estimated and whether the estimation is done in real or nominal terms. The risk premium is a fundamental and critical component in portfolio management, corporate finance, and valuation. Given its importance, it is surprising that more attention has not been paid in practical terms to estimation issues. This chapter considered the conventional approach to estimating risk premiums, which is to use historical returns on equity and government securities, and evaluated some of its weaknesses. It also examined how to extend this approach to emerging markets, where historical data tends to be both limited and volatile. The alternative to historical premiums is to estimate the equity premium implied by equity prices. This approach does require that we start with a valuation model for equities, and estimate the expected growth and cash flows, collectively, on equity investments. It has the advantages of not requiring historical data and of reflecting current market perceptions.
QUESTIONS AND SHORT PROBLEMS In the problems following, use an equity risk premium of 5.5 percent if none is specified. 1. Assume that you are valuing an Indonesian firm in U.S. dollars. What would you use as the riskless rate? 2. Explain why a six-month Treasury bill rate is not an appropriate riskless rate in discounting a five-year cash flow. 3. You have been asked to estimate a riskless rate in Indonesian rupiah. The Indonesian government has rupiahdenominated bonds outstanding, with an interest rate of 17%. S&P has a rating of BB on these bonds, and the typical spread for a BB-rated country is 5% over a riskless rate. Estimate the rupiah riskless rate. 4. You are valuing an Indian company in rupees. The current exchange rate is Rs 45 per dollar and you have been able to obtain a 10-year forward rate of Rs 70 per dollar. If the U.S. Treasury bond rate is 5%, estimate the riskless rate in Indian rupees. 5. You are attempting to do a valuation of a Chilean company in real terms. While you have been unable to get a real riskless rate in Latin America, you know that inflation-indexed Treasury bonds in the United States are yielding 3%. Could you use this as a real riskless rate? Why or why not? What are the alternatives? 6. Assume you have estimated the historical risk premium, based on 50 years of data, to be 6%. If the annual standard deviation in stock prices is 30%, estimate the standard error in the risk premium estimate. 7. When you use a historical risk premium as your expected future risk premium, what are the assumptions that you are making about investors and markets? Under what conditions would a historical risk premium give you too high a number (to use as an expected premium)? 8. You are trying to estimate a country equity risk premium for Poland. You find that S&P has assigned an A rating to Poland and that Poland has issued euro-denominated bonds that yield 7.6% in the market currently. (Germany, a AAA-rated country, has euro-denominated bonds outstanding that yield 5.1%.) a. Estimate the country risk premium, using the default spread on the country bond as the proxy. b. If you were told that the standard deviation in the Polish equity market was 25% and that the standard deviation in the Polish euro bond was 15%, estimate the country risk premium. 9. The standard deviation in the Mexican Equity Index is 48%, and the standard deviation in the S&P 500 is 20%. You use an equity risk premium of 5.5% for the United States. a. Estimate the country equity risk premium for Mexico using relative equity standard deviations. b. Now assume that you are told that Mexico is rated BBB by Standard & Poor’s and that it has dollar-denominated bonds outstanding that trade at a spread of about 3% above the Treasury bond rate. If the standard deviation in these bonds is 24%, estimate the country risk premium for Mexico. 10. The S&P 500 is at 1,400. The expected dividends and cash flows next year on the stocks in the index are expected to be 5% of the index. If the expected growth rate in dividends and cash flows over the long term is expected to be 6% and the riskless rate is 5.5%, estimate the implied equity risk premium. 11. The Bovespa (Brazilian equity index) is at 15,000. The dividends on the index last year were 5% of the index value, and analysts expect them to grow 15% a year in real terms for the next five years. After the fifth year, the growth is expected to drop to 5% in real terms in perpetuity. If the real riskless rate is 6%, estimate the implied 135
equity risk premium in this market. 12. As stock prices go up, implied equity risk premiums will go down. Is this statement always true? If not, when is it not true? Well-behaved term structures would include a upward-sloping yield curve, where long-term rates are at most 2 to 3 percent higher than short-term rates. 1
In investment analysis, where we look at projects, these durations are usually between 3 and 10 years. In valuation, the durations tend to be much longer, since firms are assumed to have infinite lives. The durations in these cases are often well in excess of 10 years and increase with the expected growth potential of the firm. 2
Ratings agencies provide ratings for a country for borrowings in the local currency and a foreign currency. The latter rating is usually lower (since countries have a greate r chance of defaulting when they borrow in a foreign currency), but the rating that matters for this analysis is the rating in the local currency. If the rating is Aaa (Moody’s) or AAA (S&P), the government bond rate will be the risk-free rate. 3
The default spread for a rating is computed by looking at dollar-denominated bond s issued by other governments with a Ba2 rating and comparing the rates on these bonds to the U.S. Treasury bond rate. 4
In cases where only a one-year forward rate exists, an approximation for the long-term rate can be obtained by first backing out the one-year local currency borrowing rate, taking the spread over the one-year Treasury bill rate, and then adding this spread onto the long-term Treasury bond rate. For insta nce, with a one-year forward rate of 39.95 on the Thai bond, we obtain a one-year Thai baht riskless rate of 9.04 percent (given a one-year Treasury bill rate of 4 percent). Adding the spread of 5.04 percent to the 10-year Treasury bond rate of 5 percent provides a 10-year Thai baht rate of 10.04 percent. 5
6 See Stocks, Bonds, Bills and Inflation, an annual edition that reports on annual returns on stocks, Treasury bonds, and Treasury bills, as well as inflation rates from 1926 to the present (www.ibbotson.com).
For the historical data on stock returns, bond returns, and bill returns, check under “Updated Data” in www.stern.nyu.edu/~adamodar. 7
These estimates of the standard error are probably understated, because they are based on the assumption that annual returns are uncorrelated over time. There is substantial empirical evidence that returns are correlated over time, which would make this standard e rror estimate much larger. 8
The compounded return is computed by taking the value of the investment at the start of the period (Value0) and the value at the end (Value N), and then computing the following: 9
In other words, good years are more likely to be followed by poor years, and vice versa. The evidence on negative serial correlation in stock returns over time is extensive, and can be found in Fama and French (1988). While they find that the one-year correlations are low, the five-year serial correlations are strongly negative for all size classes. 10
This data is from the Credit Suisse Global Investment Returns Sourcebook 2011, updated by Dimson, Marsh and Staunton at the London Business School. 11
The process by which country ratings are obtained is explained on the S&P web site at www.standardandpoors.com. 12
13 Jeremy Siegel reports on the standard deviation in equity markets in his book, Stocks for the Very Long Run: The Definitive Guide to Investment Strategies (New York, McGraw-Hill, 2007), and notes that they tend to decrease wit h time horizon.
While I have categorized the revenues into dollar revenues and revenues in dollars, the analysis can be generalized to look at revenues in stable currencies (e.g., the dollar, euro, etc.) and revenues in r isky currencies. 14
15
λAracruz = % from local marketPR/% from local marketaverage Brazilian firm = 0.20/0.80 = 0.25.
We used the average of the analyst estimates for individual firms (bottom-up). Alternatively, we could have used the top-down estimate for the S&P 500 earnings (from economists). 16
136
The Treasury bond rate is the sum of expected inflation and the expected real rate. If we assume that real growth is equal to the real rate, the long-term stable growth rate should be equal to the Treasury bond rate. 17
The input that is most difficult to estimate for emerging markets is a long-term expected growth rate. For Brazilian stocks, I used the average consensus estimate of growth in earnings for the largest Brazilian companies that have American depositary receipts (ADRs) listed on them. This est imate may be biased as a consequence. 18
On April 30, 2011, for instance, I substituted the Treasury bond rate of 3.5 percent into the regression equation to get: 19
137
CHAPTER 8 Estimating Risk Parameters and Costs of Financing The preceding chapter laid the groundwork for estimating the costs of equity and capital for firms by looking at how best to estimate a riskless rate that operates as a base for all costs, an equity risk premium for estimating the cost of equity, and default spreads for estimating the cost of debt. It did not, however, consider how to estimate the risk parameters for individual firms. This chapter examines the process of estimating risk parameters for individual firms, for estimating both the cost of equity and the cost of debt. For the cost of equity, we look at the standard process of estimating the beta for a firm and consider alternative approaches. For the cost of debt, we examine bond ratings as measures of default risk and the determinants of these ratings. The chapter closes by bringing together the risk parameter estimates for individual firms and the economy-wide estimates of the risk-free rate and risk premiums to estimate a cost of capital for the firm. To do this, the sources of capital have to be weighted by their relative market values.
THE COST OF EQUITY AND CAPITAL Firms raise money from both equity investors and lenders to fund investments. Both groups of investors make their investments expecting to make a return. Chapter 4 argued that the expected return for equity investors would include a premium for the equity risk in the investment. We label this expected return the cost of equity. Similarly, the expected return that lenders hope to make on their investments includes a premium for default risk, and we call that expected return the cost of debt. If we consider all of the financing that the firm takes on, the composite cost of financing will be a weighted average of the costs of equity and debt, and this weighted cost is the cost of capital. The chapter begins by estimating the equity risk in a firm and using the equity risk to estimate the cost of equity, and follows up by measuring the default risk to estimate a cost of debt. It concludes by determining the weights we should attach to each of these costs to arrive at a cost of capital.
COST OF EQUITY The cost of equity is the rate of return investors require on an equity investment in a firm. The risk and return models described in Chapter 4 need a riskless rate and a risk premium (in the CAPM) or premiums (in the APM and multifactor models), which were estimated in the last chapter. They also need measures of a firm's exposure to market risk in the form of betas. These inputs are used to arrive at an expected return on an equity investment:
This expected return to equity investors includes compensation for the market risk in the investment and is the cost of equity. This section concentrates on the estimation of the beta of a firm. While much of the discussion is directed at the CAPM, it can be extended to apply to the arbitrage pricing and multifactor models, as well.
Betas In the CAPM, the beta of an investment is the risk that the investment adds to a market portfolio. In the APM and multifactor model, the betas of the investment relative to each factor have to be measured. There are three approaches available for estimating these parameters: One is to use historical data on market prices for individual investments; the second is to estimate the betas from the fundamental characteristics of the investment; and the third is to use accounting data. All three approaches are described in this section.
Historical Market Betas The conventional approach for estimating the beta of an investment is a regression of returns on the investment against returns on a market index. For firms that have been publicly traded for a length of time, it is relatively straightforward to estimate returns that an investor would have made on investing in the firms' equity in intervals (such as a week or a month) over that period. In theory, these stock returns on the assets should be related to returns on a market portfolio (i.e., a portfolio that includes all traded assets) to estimate the betas of the assets. In practice, we tend to use a stock index such as the S&P 500 as a proxy for the market portfolio, and we estimate betas for stocks against the index.
Regression Estimates of Betas 138
The standard procedure for estimating betas is to regress stock returns (Rj) against market returns (Rm):
where a = Intercept from the regression b = Slope of the regression = Covariance(Rj, Rm)/σ2m The slope of the regression corresponds to the beta of the stock and measures the riskiness of the stock. The intercept of the regression provides a simple measure of performance of the investment during the period of the regression, when returns are measured against the expected returns from the capital asset pricing model. To see why, consider the following rearrangement of the capital asset pricing model:
Compare this formulation of the return on an investment to the return equation from the regression:
Thus, a comparison of the intercept a to Rf(1 – β) should provide a measure of the stock's performance, at least relative to the capital asset pricing model.1 In summary, then:
The difference between a and Rf(1 – β) is called Jensen's alpha2 and provides a measure of whether the investment in question earned a return greater than or less than its required return, given both market performance and risk. For instance, a firm that earned 15 percent during a period when firms with similar betas earned 12 percent will have earned an excess return of 3 percent; its intercept will also exceed Rf(1 – β) by 3 percent. The third statistic that emerges from the regression is the R-squared (R2) of the regression. While the statistical explanation of the R-squ ared is that it provides a measure of the goodness of fit of the regression, the economic rationale is that it provides an estimate of the proportion of the risk of a firm that can be attributed to market risk; the balance (1 – R2) can then be attributed to firm-specific risk. The final statistic worth noting is the standard error of the beta estimate. The slope of the regression, like any statistical estimate, may be different from the true value, and the standard error reveals just how much error there could be in the estimate. The standard error can also be used to arrive at confidence intervals for the “true” beta value from the slope estimate.
ILLUSTRATION 8.1: Estimating a Regression Beta for Boeing Boeing Company is a firm in both the aerospace and defense businesses, and has been traded on the New York Stock Exchange (NYSE) for decades. In assessing risk parameters for Boeing, we compute the returns on the stock and the market index in two steps: 1. The returns to a stockholder in Boeing are computed month by month from January 1996 to December 2000. These returns include both dividends and price appreciation and are defined as follows:
Dividends are added to the returns of the month in which stockholders are entitled to the dividend.3 2. The returns on the S&P 500 market index are computed for each month of the period, using the level of the index at the end of each month and the monthly dividend on stocks in the index.
where indexj is the level of the index at the end of month j and dividendsj is the dividends paid on the index in month j. Although the S&P 500 and the NYSE Composite are the most widely used indexes for U.S. stocks, they are, at best, imperfect proxies for the market portfolio in the CAPM, which is supposed to include all assets. Figure 8.1 graphs monthly returns on Boeing against returns on the S&P 500 index from January 1996 to December 2000.
139
The regression statistics for Boeing are as follows:
(a) Slope of the regression = 0.56. This is Boeing's beta, based on monthly returns from 1996 to 2000. Using a different time period for the regression or different return intervals (weekly or daily) for the same period can re sult in a different beta. (b) Intercept of the regression = 0.54%. This is a measure of Boeing's performance, when it is compared with Rf(1 – β). The monthly riskless rate (since the returns used in the regression are monthly returns) between 1996 and 2000 averaged 0.4%, resulting in the following estimate for the performance:
This analysis suggests that Boeing performed 0.36% better than expected, when expectations are based on the CAPM, on a monthly basis between January 1996 and December 2000. This results in an annualized excess return of approximately 4.41%.
Note, however, that this does not imply that Boeing would be a good investment in the future. The performance measure also does not provide a breakdown of how much of this excess return can be attributed to the performance of the entire sector (aerospace and defense) and how much is specific to the firm. To make that breakdown, we would need to compute the excess over the same period for other firms in the aerospace and defense industry and compare them with Boeing's excess return. The difference would then be attributable to firm-specific actions. In this case, for instance, the average annualized excess return on other aerospace/defense firms between 1996 and 2000 was – 0.85%, suggesting that the firm-specific component of performance for Boeing is actually 5.26% [firm-specific Jensen's alpha = 4.41% – (– 0.85%)].
(c) R-squared of the regression = 9.43%. This statistic suggests that 9.43% of the risk (variance) in Boeing comes from market sources, and that the balance of 90.57% of the risk comes from firm-specific components. The latter risk should be diversifiable and therefore will not be rewarded with a higher expected return. Boeing's R-squared is lower than the median R-squared of companies listed on the New York Stock Exchange, which was approximately 19% in 2000. (d) Standard error of beta estimate = 0.23. This statistic implies that the true beta for Boeing could range from 0.33 to 0.79 (subtracting and adding one standard error to beta estimate of 0.56) with 67% confidence and from 0.10 to 1.02 (subtracting and adding two standard errors to beta estimate of 0.56) with 95% confidence. While these ranges may seem large, they are not unusual for most U.S. companies. This suggests that we should consider estimates of betas from regressions with caution.
Figure 8.1 Boeing versus S&P 500 from 1996 to 2000
Using a Service Beta Most of us who use betas obtain them from an estimation service; Merrill Lynch, Barra, Value Line, Standard & Poor's, Morningstar, and Bloomberg are some of the well-known services. All these services begin with the regression beta just described and adjust them to reflect what they feel are better estimates of future risk. Although many of these services do not reveal their estimation procedures, Bloomberg is an exception. Figure 8.2 is the beta calculation page from Bloomberg for Boeing, using the same period as our regression (January 1996 to December 2000). Figure 8.2 Beta Estimate for Boeing Copyright 2001 Bloomberg LP. Reprinted with permission. All rights reserved.
140
While the time period used is identical to the one used in our earlier regression, there are subtle differences between this regression and the one in Figure 8.1. First, Bloomberg uses price appreciation in the stock and the market index in estimating betas and ignores dividends. 4 The fact that dividends are ignored does not make much difference for a company like Boeing, but it could make a difference for a company that either pays no dividends or pays significantly higher dividends than the market. This explains the mild differences in the intercept (0.50% versus 0.54%) and the beta (0.57 versus 0.56). Second, Bloomberg also computes what it calls an adjusted beta, which is estimated as follows:
These weights (0.67 and 0.33) do not vary across stocks, and this process pushes all estimated betas toward 1. Most services employ similar procedures to adjust betas toward 1. In doing so, they are drawing on empirical evidence that suggests that the betas for most companies, over time, tend to move toward the average beta, which is 1. This may be explained by the fact that firms get more diversified in their product mix and client base as they get larger. While we agree with the notion that betas move toward 1 over time, the weighting process used by most services strikes us as arbitrary and not particularly useful.
Estimation Choices for Beta Estimation There are three decisions that must be made in setting up the regression described earlier. The first concerns the length of the estimation period. Most estimates of betas, including those by Value Line and Standard & Poor's, use five years of data, while Bloomberg uses two years of data. The trade-off is simple: A longer estimation period provides more data, but the firm itself might have changed in its risk characteristics over the time period. Boeing, during the period of our analysis, acquired both Rockwell and McDonnell Douglas, changing its business mix and its basic risk characteristics. The second estimation issue relates to the return interval. Returns on stocks are available on an annual, a monthly, a weekly, a daily, and even an intraday basis. Using daily or intraday returns increases the number of observations in the regression, but it exposes the estimation process to a significant bias in beta estimates related to nontrading.5 For instance, the betas estimated for small firms, which are more likely to suffer from nontrading, are biased downward when daily returns are used. Using weekly or monthly returns can reduce the nontrading bias significantly.6 In this case, using weekly returns for two years yields a beta estimate for Boeing of only 0.88, while the monthly beta estimate over the same period is 0.96. The third estimation issue relates to the choice of a market index to be used in the regression. The standard practice used by most beta estimation services is to estimate the betas of a company relative to the index of the market in which its stock trades. Thus, the betas of German stocks are estimated relative to the Frankfurt DAX, British stocks relative to the FTSE, Japanese stocks relative to the Nikkei, and U.S. stocks relative to the NYSE Composite or the S&P 500. While this practice may yield an estimate that is a reasonable measure of risk for the domestic investor, it may not be the best approach for an international or cross-border investor, who would be better served with a beta estimated relative to an international index. For instance, Boeing's beta between 1996 and 2000 estimated relative to the Morgan Stanley Capital International (MSCI) index that is composed of stocks from different global markets yields a value of 0.82. To the extent that different services use different estimation periods, use different market indexes, and adjust the regression beta differently, they will often provide different beta estimates for the same firm at the same point in time. While these beta differences are troubling, note that the beta estimate delivered by each of these services comes with a standard error, and it is very likely that all the betas reported for a firm fall within the range of 141
standard errors from the regressions.
Historical Beta Estimation for Companies in Smaller (or Emerging) Markets The process for estimating betas in markets with fewer stocks listed on them is no different from the process described earlier, but the estimation choices on return intervals, the market index, and the return period can make a much bigger difference in the estimate.
INDEX DOMINATION AND BETA ESTIMATES There are a number of indexes that are dominated by one or a few stocks. One of the most striking cases was the Helsinki Stock Exchange (HEX) in the late 1990s. Nokia, the telecommunications giant, represented 75 percent of the Helsinki Index in terms of market value. Not surprisingly, a regression of Nokia against the HEX yielded the results shown in Figure 8.3. The regression looks impeccable. In fact, the noise problem that we noted with Boeing, arising from the high standard errors, disappears. The beta estimate has a standard error of 0.03, but the results are deceptive. The low standard error is the result of a regression of Nokia on itself, since it dominates the index. The beta is meaningless to a typical investor in Nokia, who is likely to be diversified, if not globally, at least across European stocks. Worse still, the betas of all other Finnish stocks against the HEX become betas estimated against Nokia. In fact, the beta of every other Finnish stoc k at the time of this regression was less than 1. How is this possible, you might ask, if the average beta is 1? It is the weighted average beta that is 1, and if Nokia (which comprises three-quarters of the index) has a beta greater than 1 (which it does), every other stock in the index could well end up with a beta less than 1.
Figure 8.3 Beta Estimate for Nokia Copyright 2001 Bloomberg LP. Reprinted with permission. All rights reserved.
When liquidity is limited, as it often is in many stocks in emerging markets, the betas estimated using short return intervals tend to be much more biased. In fact, using daily or even weekly returns in these markets will tend to yield betas that are not good measures of the true market risk of the company. In many emerging markets, both the companies being analyzed and the market itself change significantly over short periods of time. Using five years of returns, as we did for Boeing, for a regression may yield a beta for a company (and market) that bears little resemblance to the company (and market) as it exists today. Finally, the indices that measure market returns in many smaller markets tend to be dominated by a few large companies. For instance, the Bovespa (the Brazilian index) was dominated for several years by Telebras, which represented almost half the index. Nor is this just a problem with emerging markets. The DAX, the equity index for German stocks, is dominated by Allianz, Deutsche Bank, Siemens, and Daimler. When an index is dominated by one or a few companies, the betas estimated against that index are unlikely to be true measures of market risk. In fact, the betas are likely to be close to 1 for the large companies that dominate the index and wildly variable for all other companies.
ILLUSTRATION 8.2: Estimating a Beta for Titan Cement Company Titan Cement is a cement and construction company in Greece. Reproduced in Figure 8.4 is the beta estimate for Titan from April 1999 to April 2001 (using weekly returns) obtained from a beta service (Bloomberg). Note that the index used is the Athens Stock Exchange Index. Based on this regression, we arrive at the following equation:
Figure 8.4 Beta Estimate for Titan Cement: Athens Stock Exchange Index
142
The beta for Titan Cement, based upon this regression, is 0.93. The standard error of the estimate, shown in brackets below, is only 0.08, but the caveats about narrow indexes apply to the Athens Stock Exchange Index. Drawing on the arguments in the previous section, if the marginal investor in Titan Cement is, in fact, an investor diversified across European companies, the appropriate index would have been a European stock index. The Bloomberg beta calculation with the MSCI European index is reported in Figure 8.5. Note the decline in beta to 0.33 and the increase in the standard error of the beta estimate.
Figure 8.5 Beta Estimate for Titan Cement: MSCI European Index
In fact, if the marginal investor is globally diversified, Titan Cement's beta (as well as Boeing's beta in Illustration 8.1) should have been estimated against a global index. Using the Morgan Stanley Capital International (MSCI) global index, we get a regression beta of 0.33 in Figure 8.6 . In fact, the beta estimate and the standard error look very similar to the ones estimated against the European index.
Figure 8.6 Beta Estimate for Titan Cement: MSCI Global Index
143
Estimating the Historical Beta for Private Firms The historical approach to estimating beta s works only for assets that have been traded and have market prices. Private companies do not have a market price history. Consequently, we cannot estimate a regression beta for these companies. Nevertheless, we still need estimates of cost of equity and capital for these comp anies. You might argue that this is not an issue if you do not value private companies; but you will still be confronted with this issue even when valuing publicly traded firms. Consider, for instance, the following scenarios: If you have to value a private firm for an initial public offering, you will need to estimate discount rates for the valuation. Even after a firm has gone public, there will be a period of time lasting as long as two years when there will be insufficient data for a regression. If you are called upon to value the division of a publicly traded firm that is up for sale, you will not have past prices to draw on to run a regression. Finally, if the firm has gone through significant restructuring—divestitures or recapitalization—in the recent past, regression betas become meaningless because the company itself has changed its risk characteristics. Thus regression betas are either unavailable or meaningless in a significant number of valuations. Some analysts assume that discounted cash flow valuation is not feasible in these scenarios; instead they use multiples. Others make assumptions about discount rates based on rules of thumb. Neither approach is appealing. The next section develops an approach for estimating betas that is general enough to apply to all of these companies.
risk.xls. This spreadsheet allows you to run a regression of stock returns against market returns and estimate risk parameters.
The Limitations of Regression Betas Much of what has been presented in this section represents an indictment of regression betas. In the case of Boeing, the biggest problem was that the beta had high standard error. In fact, this is not a problem unique to Boeing. Figure 8.7 presents the distribution of standard errors on beta estimates for U.S. companies. Figure 8.7 Distribution of Standard Errors on Beta—U.S. Firms from 2008 to 2010
144
With the Nokia regression, we seem to cure the standard error problem but at a very large cost. The low standard errors reflect the domination of the index by a stock and result in betas that may be precise but bear no resemblance to true risk. Changing the market index, the return period, and the return interval offers no respite. If the index becomes a more representative index, the standard errors on betas will increase, reflecting the fact that more of the risk in the stock is firm-specific. If the beta changes as the return period or interval changes, it creates more uncertainty about the true beta of the company. In short, regression betas will almost always be either too noisy or skewed by estimation choices to be useful measures of the equity risk in a company. The cost of equity is far too important an input into a discounted cash flow valuation to be left to statistical chance.
Fundamental Betas A second way to estimate betas is to look at the fundamentals of the business. The beta for a firm may be estimated from a regression, but it is determined by decisions the firm has made on what business to be in and how much operating leverage to use in the business, and by the degree to which the firm uses financial leverage. This section examines an alternative way of estimating betas, where we are less reliant on historical betas and more cognizant of their fundamental determinants.
Determinants of Betas The beta of a firm is determined by three variables: (1) the type of business or businesses the firm is in, (2) the degree of operating leverage of the firm, and (3) the firm's financial leverage. Although we will use these determinants to find betas in the capital asset pricing model, the same analysis can be used to calculate the betas for the arbitrage pricing and the multifactor models as well.
Type of Business Since betas measure the risk of a firm relative to a market index, the more sensitive a business is to market conditions, the higher its beta. Thus, other things remaining equal, cyclical firms can be expected to have higher betas than noncyclical firms. Companies involved in housing and automobiles, two sectors of the economy that are very sensitive to economic conditions, should have higher betas than companies in food processing and tobacco, which are relatively insensitive to business cycles. This view can be extended to a company's products. The degree to which a product's purchase is discretionary will affect the beta of the firm manufacturing the product. Firms whose products are much more discretionary to their customers—they can defer or delay buying these products—should have higher betas than firms whose products are viewed as necessary or less discretionary. Thus, the beta of Procter & Gamble, which sells diapers and daily household products, should be lower than the beta of Gucci, which manufactures luxury products.
Degree of Operating Leverage The degree of operating leverage is a function of the cost structure of a firm and is usually defined in terms of the relationship between fixed costs and total costs. A firm that has high fixed costs relative to total costs is said to have high operating leverage. A firm with high operating leverage will also have higher variability in operating income than would a firm producing a similar product with low operating leverage. Other things remaining equal, 145
the higher variance in operating income will lead to a higher beta for the firm with high operating leverage. Can firms change their operating leverage? While some of a firm's cost structure is determined by the business it is in (an energy utility has to build expensive power plants, and airlines have to buy or lease expensive planes), firms in the United States have become increasingly inventive in lowering the fixed cost component in their total costs. For instance, firms have made cost structures more flexible by: Negotiating labor contracts that emphasize flexibility and allow the firm to make its labor costs more sensitive to its financial success. Entering into joint venture agreements, where the fixed costs are borne by someone else. Subcontracting manufacturing and outsourcing, which reduce the need for expensive plant and equipment. While the arguments for such actions may be couched in terms of competitive advantage and flexibility, they do also reduce the operating leverage of the firm and its exposure to market risk. While operating leverage affects betas, it is difficult to measure the operating leverage of a firm, at least from the outside, since fixed and variable costs are often aggregated in income statements. It is possible to get an approximate measure of the operating leverage of a firm by looking at changes in operating income as a function of changes in sales.
For firms with high operating leverage, operating income should change more than proportionately when sales change.
SIZE, GROWTH, AND BETAS Generally, smaller firms with higher growth potential are viewed as riskier than larger, more stable firms. While the rationale for this argument is clear when talking about total risk, it becomes more difficult to see when looking at market risk or betas. Should a smaller software firm have a higher beta than a larger software firm? One reason to believe that it should is operating leverage. If there is a setup cost associated with investing in infrastructure or economies of scale, smaller firms will have higher fixed costs than larger firms, leading in turn to higher betas for these firms. With growth firms, the argument for higher betas rests on the notion of discretionary versus nondiscretionary purchases. For a high-growth firm to deliver on its growth, new customers have to adopt the product or existing customers have to buy more of the product. Whether they do so will depend, in large part, on how well-off they feel. This, in turn, will make the profits of high-growth firms much more dependent on how well the economy is doing, thus increasing their betas.
Degree of Financial Leverage Other things remaining equal, an increase in financial leverage will increase the beta of the equity in a firm. Intuitively, we would expect that the fixed interest payments on debt result in increasing income in good times and decreasing income in bad times. Higher leverage increases the variance in net income and makes equity investment in the firm riskier. If all the firm's risk is borne by the stockholders (i.e., the beta of debt is zero),7 and debt has a tax benefit to the firm, then,
where βL = Levered beta for equity in the firm βu = Unlevered beta of the firm (i.e., the beta of the firm without any debt) t = Marginal tax rate D/E = Debt-to-equity ratio (market value) Intuitively, we expect that as leverage increases (as measured by the debt-to-equity ratio), equity investors bear increasing amounts of market risk in the firm, leading to higher betas. The tax factor in the equation captures the tax benefits that accrue from interest payments. The unlevered beta of a firm is determined by the nature of its products and services (cyclicality, discretionary nature) and its operating leverage. It is often also referred to as the asset beta, since it is determined by the assets owned by the firm. Thus, the levered beta, which is also the beta for an equity investment in a firm, is determined both by the riskiness of the business it operates in and by the amount of financial leverage risk it has taken on. Since financial leverage multiplies the underlying business risk, it stands to reason that firms that have high business risk should be reluctant to take on financial leverage. It also stands to reason that firms that operate in stable businesses should be much more willing to take on financial leverage. Utilities, for instance, have historically had high debt ratios but have not had high betas, mostly because their underlying businesses have been stable and fairly predictable.
ILLUSTRATION 8.3: Effects of Leverage on Betas: Boeing 146
From the regression for the period from 1996 to 2000, Boeing had a historical beta of 0.56. Since this regression uses stock prices of Boeing over this period, we began by estimating the average debt-to-equity ratio between 1996 and 2000, using market values for debt and equity.
The beta over the 1996–2000 period reflects this average leverage. To estimate the unlevered beta over the period, a marginal tax rate of 35% is used:
The unlevered beta for Boeing over the 1996–2000 period is 0.51. The levered beta at different levels of debt can then be estimated:
For instance, if Boeing were to increase its debt equity ratio to 10%, its equity beta will be:
If the debt equity ratio were raised to 25%, the equity beta would be:
The following table summarizes the beta estimates for different levels of financial leverage ranging from 0% to 90% debt.
As Boeing's financial leverage increases, the beta increases concurrently.
levbeta.xls. This spreadsheet allows you to estimate the unlevered beta for a firm and compute the betas as a function of the leverage of the firm.
Bottom-Up Betas Breaking down betas into their business risk and financial leverage components provides us with an alternative way of estimating betas, in which we do not need past prices on an individual firm or asset to estimate its beta. To develop this alternative approach, we need to introduce an additional property of betas that proves invaluable. The beta of two assets put together is a weighted average of the individual asset betas, with the weights based on market value. Consequently, the beta for a firm is a weighted average of the betas of all the different businesses it is in. We can estimate the beta for a firm in five steps:
Step 1: Identify the business or businesses the firm operates in. Step 2: Find other publicly traded firms in each business and obtain their regression betas, which we use to compute an average beta for the firms. Step 3: Estimate the average unlevered beta for the business by unlevering the average (or median) beta for the firms by their average (or median) debt-to-equity ratio. Alternatively, we could estimate the unlevered beta for each firm and then compute the average of the unlevered betas. The first approach is preferable because unlevering an erroneous regression beta is likely to compound the error.
Step 4: Estimate an unlevered beta for the firm being analyzed, taking a weighted average of the unlevered betas for the businesses it operates in, using the proportion of firm value derived from each business as the weights. If values are not available, use operating income or revenues as weights. This weighted average is called the bottom-up unlevered beta. 147
where the firm is assumed to operating in k different businesses.
Step 5: Finally, estimate the current market values of debt and equity at the firm and use this debt-to-equity ratio to estimate a levered beta. The betas estimated using this processs are called bottom-up betas.
The Case for Bottom-Up Betas At first sight, the use of bottom-up betas may seem to leave us exposed to all of the problems noted with regression betas. After all, the betas for other publicly traded firms in the business are obtained from regressions. Notwithstanding this, bottom-up betas represent a significant improvement on regression betas for the following reasons: Although each regression beta is estimated with standard error, the average across a number of regression betas has much lower standard error. The intuition is simple. A high standard error on a beta estimate indicates that it can be significantly higher or lower than the true beta. Averaging across these individual regression betas results in an average beta that is far more precise than the individual betas that went into it. In fact, if the estimation errors on individual firm betas are uncorrelated across firms, the savings in standard error can be stated as a function of the average standard error or beta estimates and the number of firms in the sample.
where n is the number of firms in the sample. Thus, if the average standard error in beta estimates for software firms is 0.50 and the number of software firms is 100, the standard error of the average beta is only 0.05 (0.50/√100). A bottom-up beta can be adapted to reflect actual changes in a firm's business mix and expected changes in the future. Thus if a firm divested a major portion of its operations last week, the weights on the businesses can be modified to reflect the divestiture. The same can be done with acquisitions. In fact, a firm's strategic plans to enter new businesses in the future can be brought into the beta estimates for future periods. Firms do change their debt ratios over time. Although regression betas reflect the average debt-to-equity ratio maintained by the firm during the regression period, bottom-up betas use the current debt-to-equity ratio. If a firm plans to change its debt-to-equity ratio in the future, the beta can be adjusted to show these changes. Finally, bottom-up betas wean us from our dependence on historical stock prices. While we do need these prices to get betas for comparable firms, all we need for the firm being analyzed is a breakdown of the businesses it is in. Thus, bottom-up betas can be estimated for private firms, divisions of businesses, and stocks that have just started trading in financial markets.
Computational Details Although the idea behind bottom-up betas is fairly simple, there are several computational details that deserve attention.
Defining comparable firms. First, we have to decide how narrowly we want to define a business. Consider, for instance, a firm that manufactures entertainment software. We could define the business as entertainment software and consider only companies that primarily manufacture entertainment software to be comparable firms. We could go even further and define comparable firms as firms manufacturing entertainment software with revenues similar to that of the company being analyzed. While there are benefits to narrowing the comparable firm definition, there is a cost. Each additional criterion added to the definition of comparable will mean that fewer firms make the list, and the savings in standard error that comprise the biggest benefit to bottom-up betas become smaller. A commonsense principle should therefore come into play. If there are hundreds of firms in a business, as there are in the software sector, you can afford to be more selective. If there are relatively few firms, not only do you have to become less selective, you might have to broaden the definition of comparable to bring other firms into the mix. Estimating betas. Once the comparable firms in a business have been defined, you have to estimate the betas for these firms. Although it would be best to estimate the beta for each of these firms against a common and well-diversified equity index, it is usually easier to use service betas that are available for each of these firms. These service betas may be estimated against different indexes. For instance, if you define your business to be global telecommunications and obtain betas for global telecom firms from Bloomberg, 148
these betas will be estimated against the local indexes. This is usually not a fatal problem, especially with large samples, since errors in the estimates tend to average out. Averaging method. The average beta for the firms in the sector can be computed in one of three ways. We could use market-weighted averages, but the savings in standard error that touted in the earlier section will be muted, especially if there are one or two very large firms in the sample. We could estimate the simple average of the betas of the companies, thus weighting all betas equally. The process weights the smallest firms in the sample disproportionately (to their market value), but the savings in standard error are likely to be maximized. If the data being average (betas, debt to equity ratios) have large outliers, we can use the median values. Controlling for differences. In essence, when we use betas from comparable firms, we are assuming that all firms in the business are equally exposed to business risk and have similar operating leverage. Note that the process of levering and unlevering of betas allows us to control for differences in financial leverage. If there are significant differences in operating leverage—cost structure—across companies, the differences in operating leverage can be controlled for as well. This would require estimation of a business beta, where the effects of operating leverage are taken out from the unlevered beta:
Note the similarity to the adjustment for financial leverage; the only difference is that both fixed and variable costs are eligible for the tax deduction, and the tax rate is therefore no longer a factor. The business beta can then be relevered to reflect the differences in operating leverage across firms.
CASH AND BETAS In the process for estimating bottom up betas, we suggested a two step process: getting a weighted average of the betas of the businesses that a firm is in, using the sector-average betas of other publicly traded firms in each business and then adjusting for the debt to equity ratio of the firm in question. In making these adjustments, though, we have to deal that a firm may have a significant portion of its assets as cash. Since cash is usually invested in close to riskless, liquid investments, it should have a beta of zero. So, how does cash enter the computation? It does so in two places. When we computed the sector-average beta, we suggesting unlevering the average regression beta for the sector, using the average debt to equity ratio and marginal tax rate for the sector. Thus, with an average levered beta of 1.30, an average debt to equity ratio of 50% and an average tax rate of 40%, we estimate a sector-average unlevered beta of 1.00 for the entertainment business:
However, this is the unlevered beta for companies in this business and these companies will generally have some of these value in cash balances. Assume, for instance, that the average cash balance of entertainment firms in the sector is 10%. The unlevered beta for the entertainment business alone can then be computed as follows: Unlevered beta for entertainment (.90) + Beta for cash (.10) = 1.00 Plugging in a beta of zero for cash, we get a beta for just the entertainment business: Unlevered beta for entertainment business = 1.00/.90 = 1.11 We call this the beta for the sector, corrected for cash, and use it in the computation of bottom up betas. The second place it shows up is when we compute the bottom up beta for a company. To estimate the bottom up beta for just the operating assets of a company like Boeing, we would take a weighted average of the cash corrected unlevered betas of the aerospace and defense businesses. This is the beta we would use to compute the cost of equity and cost of capital. To get a bottom up beta for Boeing as a company, we would then bring in the cash holdings as a separate asset and give it a beta of zero. This beta would then be a beta for all of Boeing's assets and for Boeing's equity in those assets. Each beta has a use in valuation.
betas.xls: This dataset on the Web has updated betas and unlevered betas by business sector in the United States.
ILLUSTRATION 8.4: Estimating a Bottom-Up Beta for Vans Shoes—January 2001 Vans Shoes is a shoe manufacturing firm with a market capitalization of $191 million. To estimate the bottom-up beta for Vans Shoes, consider the betas of all publicly traded shoe companies in the following table:
149
In addition to the beta for each firm, the table reports the market debt-to-equity ratio, the effective tax rate, and a measure of operating leverage obtained by dividing selling, general, and administrative (SG&A) expenses (which we consider fixed) by other operating expenses (which we consider variable). We can estimate the unlevered beta for the business using the averages for these values:
Using the average tax rate of 25.95%, we can estimate the unlevered beta.
The beta for Vans Shoes can then be obtained using the firm's marginal tax rate of 34.06% and its market debt-to-equity ratio of 9.41%.
This levered beta is based on the implicit assumption that all shoe manufacturers have similar operating leverage. In fact, we could adjust the unlevered beta for the average fixed cost/variable cost ratio for the business and then relever back at the operating leverage for Vans Shoes:
We can then use Vans' fixed cost/variable cost ratio of 31.16% to estimate an adjusted unlevered and levered beta.
By having a debt-to-equity ratio, and operating leverage, that is lower than the average for the industry, Vans Shoes ends up with a beta much lower than that of the industry.
ILLUSTRATION 8.5: Estimating a Bottom-Up Beta for Boeing—September 2000 Boeing has undergone a significant change in both its business mix and its financial leverage over the past five years. Not only did it acquire Rockwell and McDonnell Douglas, giving it a major foothold in the defense business, but it borrowed substantial amounts to make these acquisitions. Since these events have occurred over time, the historical regression beta does not fully reflect the effects of these changes. To estimate Boeing's beta in 2000, we broke its business into two areas: 1. Commercial aircraft, which is Boeing's core business of manufacturing commercial jet aircraft and providing related support services. 2. Information, space, and defense systems (ISDS), which include research, development, production, and support of military aircraft, helicopters, and missile systems. Each of these areas of business has very different risk characteristics, and the unlevered beta for each business was estimated by looking at comparable firms in each business. The following table summarizes these estimates.
150
For commercial aircraft there are no truly comparable firms. We looked at Boeing's own beta prior to its expansion in the defense business and computed the unlevered beta using this estimate. For ISDS, we used 17 firms that derived the bulk of their revenues from defense contracting, and computed the average beta and debt-to-equity ratio for these firms. The unlevered beta was computed using these averages. The values for each of the divisions were estimated using the revenues from each segment8 and a typical revenue multiple9 for that type of business. The unlevered beta for Boeing as a company in 2000 can be estimated by taking a value-weighted average of the betas of each of the different business areas. This is reported in the last column to be 0.8774. The equity beta can then be estimated using the debt-to-equity ratio for Boeing in 2000. Combining the market value of equity of $55.20 billion and the value of debt of $7.85 billion, and using a 35% tax rate for the firm, we arrive at the beta for Boeing.
This is very different from the historical beta of 0.56 that we obtained from the regression, but it is, in our view, a much truer reflection of the risk in Boeing in 2000.
ILLUSTRATION 8.6: Estimating a Bottom-Up Beta for Titan Cement—January 2000 To estimate a beta for Titan Cement, we began by defining comparable firms as other cement companies in Greece but found only one comparable firm. When we expanded the list to include cement companies across Europe, we increased our sample to nine firms. Since we did not see any reason to restrict our comparison to just European firms, we decided to look at the average beta for cement companies globally. There were 108 firms in this sample, with an average beta of 0.99, an average tax rate of 34.2%, and an average debt-to-equity ratio of 27.06%. We used these numbers to arrive at an unlevered beta of 0.84.
We then used Titan's market values of equity (566.95 million Gdr) and debt (13.38 million Gdr) to estimate a levered beta for its equity:
We used Titan's marginal tax rate of 24.14% in this calculation.
HOW WELL DO BETAS TRAVEL? Often, when analyzing firms in small or emerging markets, we have to estimate betas by looking at firms in the same business but traded on other markets. This is what we did when estimating the beta for Titan Cement. Is this appropriate? Should the beta for a steel company in the United States be comparable to that of a steel company in Indonesia? We see no reason why it should not be. But the company in Indonesia has much more risk, you might argue. We do not disagree, but the fact that we use similar betas does not mean that we believe that the costs of equity are identical across all steel companies. In fact, using the approach described in the preceding chapter, the risk premium used to estimate the cost of equity for the Indonesian company will incorporate a country risk premium, whereas the cost of equity for the U.S. company will not. Thus, even if the betas used for the two companies are identical, the cost of equity for the Indonesian company will be much higher. There are a few exceptions to this proposition. Recall that one of the key determinants of betas is the degree to which a product or service is discretionary. It is entirely possible that products or services that are discretionary in one market (and command high betas) may be nondiscretionary in another market (and have low betas). For instance, phone service is viewed as a nondiscretionary product in most developed markets, but is a discretionary product in emerging markets. Consequently, the average beta estimated by looking at telecom firms in developed markets will understate the true beta of a telecom firm in an emerging market. For the latter beta, the comparable firms should be restricted to include only telecom firms in emerging markets.
Calculating Betas after a Major Restructuring The bottom-up process of estimating betas provides a solution when firms go through major restructurings that change both their business mix and their leverage. In these cases, the regression betas are misleading because they do not reflect fully the effects of these changes. Boeing's beta estimated using the bottom-up approach is likely to provide a more precise estimate than the historical beta from a regression of Boeing's stock prices, given Boeing's acquisitions of Rockwell and McDonnell Douglas and its increase in leverage. In fact, a firm's beta can be estimated using the bottom-up approach even before the restructuring becomes effective. Illustration 8.7, for instance, estimates Boeing's beta just before and after its acquisition of McDonnell Douglas, allowing for the changes in both the business mix and the leverage.
ILLUSTRATION 8.7: Beta of a Firm after an Acquisition: Boeing and McDonnell Douglas In 1997, Boeing announced that it was acquiring McDonnell Douglas, another company involved in the aerospace and defense business. At the time of the acquisition, the two firms had the following market values and betas:
Note that the market values of equity used for the two firms reflect the market values after the acquisition announcement and reflect the acquisition price agreed on for McDonnell Douglas shares. In order to evaluate the effects of the acquisition on Boeing's beta, we first examine the effects of the merger on the business risk of the combined firm by estimating the unlevered betas of the two companies and calculating the combined firm's unlevered beta. (We used a
151
marginal tax rate of 35% for both firms.)
The unlevered beta for the combined firm can be calculated as the weighted average of the two unlevered betas, with the weights based on the market values of the two firms.
Boeing's acquisition of McDonnell Douglas was accomplished by issuing new stock in Boeing to cover the value of McDonnell Douglas' equity of $12,555 million. Since no new debt was used to finance the deal, the debt outstanding in the firm after the acquisition is just the sum of the debt outstanding at the two companies before the acquisition.
The debt/equity ratio can then be computed as follows:
This debt/equity ratio in conjunction with the new unlevered beta for the combined firm yields a new beta of:
Accounting Betas A third approach is to estimate the market risk parameters from accounting earnings rather than from traded prices. Thus, changes in earnings at a division or a firm, on a quarterly or an annual basis, can be related to changes in earnings for the market, in the same periods, to arrive at an estimate of a accounting beta to use in the CAPM. While the approach has some intuitive appeal, it suffers from three potential pitfalls. First, accounting earnings tend to be smoothed out relative to the underlying value of the company, as accountants spread expenses and income over multiple periods. This results in betas that are “biased down,” especially for risky firms, or “biased up” for safer firms. In other words, betas are likely to be closer to 1 for all firms using accounting data. Second, accounting earnings can be influenced by nonoperating factors, such as changes in depreciation or inventory methods, and by allocations of corporate expenses at the divisional level. Finally, accounting earnings are measured, at most, once every quarter, and often only once every year, resulting in regressions with few observations and not much explanatory power (low R-squared, high standard errors).
ILLUSTRATION 8.8: Estimating Accounting Betas: Defense Division of Boeing—1995 Having operated in the defense business for decades, Boeing has a record of its profitability. These profits are reported in the following table, together with earnings changes for companies in the S&P 500 from 1980 to 1994.
Year
S&P 500
Boeing's Defense Business
1980
–2.10%
–12.70%
1981
–6.70%
–35.56%
1982
–45.50%
27.59%
1983
37.00%
159.36%
1984
41.80%
13.11%
1985
–11.80%
–26.81%
1986
7.00%
–16.83%
1987
41.50%
20.24%
1988
41.80%
18.81%
1989
2.60%
–29.70%
1990
–18.00%
–40.00%
1991
–47.40%
–35.00%
1992
64.50%
10.00%
1993
20.00%
–7.00%
1994
25.30%
11.00%
152
Copyright 2001 Bloomberg LP. Reprinted with permission. All rights reserved. Regressing the changes in profits in the defense division (Δ Earningsdefense) against changes in profits for the S&P 500 (Δ EarningsS&P) yields the following:
Based on this regression, the beta for the defense division is 0.65.
accbeta.xls: This spreadsheet allows you to estimate the accounting beta on a division or firm.
spearn.xls: This dataset on the Web has earnings changes, by year, for the S&P 500 going back to 1960.
Market, Bottom-Up, and Accounting Betas: Which One Do We Use? For most publicly traded firms, betas can be estimated using accounting data or market data or from the bottom-up approach. Since the betas will almost never be the same using these different approaches, the question is, which one do we use? We would almost never use accounting betas, for all the reasons specified earlier. We are almost as reluctant to use historical market betas for individual firms because of the standard errors in beta estimates, the failures of the local indexes (as is the case with most emerging market companies), and the inability of these regressions to reflect the effects of major changes in the business mix and financial risk at the firm. Bottom-up betas, in our view, provide us with the best beta estimates for three reasons: 1. They allow us to consider changes in business and financial mix, even before they occur. 2. They use average betas across large numbers of firms, which tend to be less noisy than individual firm betas. 3. They allow us to calculate betas by area of business for a firm, which is useful both in the context of investment analysis and in valuation.
Measuring Country Risk Exposure (Lambda) Chapter 7 introduced the concept of country risk exposure and the notion of lambda—a measure of a company's exposure to country risk. In this section, we want to consider intuitively what factors determine this exposure and how best to estimate lambda. A company's exposure to country risk is affected by almost every aspect of its operations, beginning with where its factories are located and who its customers are and continuing with what currency its contracts are denominated in and how well it manages its exposure to exchange rate risk. Much of this information, however, is internal information and not available to someone valuing the firm from the outside. As a practical matter, then, we can estimate lambda using one of the following approaches:
Revenue breakdown. The simplest way of estimating lambda is to use the proportion of a firm's revenues that are generated in a country and scale this to the proportion of the revenues generated by the average firm in that country.
Consider Embraer, a Brazilian aerospace company that in 2008 derived about 9 percent of its revenues from Brazil. If the average Brazilian company generates 60 percent of its revenues in Brazil, this would translate into a lambda of 0.15 (0.09/.060). Note, though, that if Embraer gets any of its remaining revenue in other risky emerging markets, you would have to compute lambdas against these markets as well. Regression versus country bond. A second approach to estimating lambdas would be to run regressions of stock returns for each firm in the emerging market against the returns on the country bond. In effect, we are assuming that returns on the country bond are reflections of changes in country risk: Country bond prices increase when country risk decreases and decrease when country risk increases. When we run this regression, we are measuring how sensitive a company's stock price is to changes in country risk perceptions. To provide an illustration: Regressing the stock prices of Embraer against the dollardenominated Brazilian government bond from 2006 to 2008 yields a slope (lambda) of 0.27. Put in intuitive 153
terms, Embraer's returns moved 0.27 percent for every 1 percent change in returns on the bond. That would be our estimate of lambda for the company.
LAMBDAS: WORTH THE TROUBLE? The intuition behind the use of lambdas is that a company's risk exposure should be based on where it does business and not where it is incorporated. Thus, an emerging market company that gets the bulk of its revenues in developed markets should be less exposed to the country risk in that emerging market. By the same token, a developed market company that gets large portions of its revenues in emerging markets should see its cost of equity increase because of that exposure. Having said this, it is often difficult to obtain the information needed to estimate lambdas. The lambda for a company should depend not only on where it gets its revenues, but where it produces its goods and the degree to which it insures against country risk using derivatives or conventional insurance. For most companies, the information on these inputs is either unavailable or incomplete. Thus, any benefits from estimating lambdas may be drowned out by the estimation error in those lambdas. For firms that have revenue exposures that are similar to those of other firms in the market, it may make sense to stick with the standard approach of using beta to capture company risk. There are two scenarios where it does make sense to estimate lambdas: 1. Emerging market companies with disproportionately large developed market exposures. In almost every emerging market, there are a few companies that are incorporated in that market that derive the bulk of their revenues outside the market: Tata Consulting Services in India and Embraer in Brazil are good examples, deriving less than 10% of their revenues in the domestic market. For these companies, we would expect low lambdas against their local markets, reflecting their lighter domestic risk exposure. 2. Developed market companies with large revenues from risky emerging markets. Note that it is not revenues from foreign markets per se that create a problem but revenues from markets that have significant political and economic risk. Nestle and Coca-Cola, for instance, are developed market companies that have significant revenues from Asia and Latin America. For these companies, we should be adjusting the cost of equity for the additional risk exposure from emerging market countries.
From Betas to Cost of Equity Having estimated the riskless rate and the risk premium(s) in Chapter 7 and the beta(s) in this chapter, we can now estimate the expected return from investing in equity at any firm. In the CAPM, this expected return can be written as:
where the riskless rate would be the rate on a long-term government bond; the beta would be either the historical, fundamental, or accounting betas described earlier; and the risk premium would be either the historical premium or an implied premium. In the arbitrage pricing and multifactor model, the expected return would be written as follows:
where the riskless rate is the long-term government bond rate; βj is the beta relative to factor j, estimated using historical data or fundamentals; and risk premiumj is the risk premium relative to factor j, estimated using historical data. The expected return on an equity investment in a firm, given its risk, has implications for both equity investors in the firm and the managers of the firm. For equity investors, it is the rate they need to earn to be compensated for the risk they have taken in investing in the equity of the firm. If, after analyzing an investment, they conclude they cannot make this return, they would not buy this investment; alternatively, if they decide they can make a higher return, they would make the investment. For managers in the firm, the return investors need to make to break even on their equity investments becomes the return they have to try to deliver to keep these investors from becoming restive and rebellious. Thus, it becomes the rate they have to beat in terms of returns on their equity investments in projects. In other words, this is the cost of equity to the firm.
ILLUSTRATION 8.9: Estimating the Cost of Equity for Boeing—December 2000 Now that we have an estimate of beta of 0.9585 for Boeing, based on the bottom-up estimates, we can estimate its cost of equity. To make the estimate, we used the prevailing Treasury bond rate of 5% and a historical risk premium of 5.51%.
There are two point to make about this estimate. The first is that the cost of equity would have been significantly lower if we had chosen to use the implied equity premium on December 31, 2000, which was about 2.87% (see Chapter 7).
The second point is that we are not considering the exposure that Boeing has to emerging market risk from its business. If the exposure is significant, we should be adding a country risk premium to the cost of equity estimate.
154
ILLUSTRATION 8.10: Estimating the Cost of Equity for Embraer—March 2008 Embraer is a Brazilian aerospace firm. To estimate its cost of equity, we first estimated the unlevered beta by looking at aerospace firms globally.
Embraer's debt-to-equity ratio at the time of this analysis was 26.84%,10 resulting in a levered beta for Embraer (with a marginal tax rate of 34% for Brazil):
To estimate the cost of equity for Embraer in U.S. dollar terms, we began with the Treasury bond rate of 3.8% at the time of the analysis, but incorporated the country risk associated with Brazil into the risk premium. Using the approach described in Chapter 7, we estimated a country risk premium of 3.66% in March 2008; the default spread for Brazil at the time was 2.00% and the Bovespa was approximately 1.83 times more volatile than the Brazilian government bond. In conjunction wit h a mature market risk premium of 4% estimated for the United States at the time, this yields a cost of equity of 10.54%.
Again, there are several points that are worth making on this estimate. The first is that this cost of equity can be expected to change over time as Brazil matures as a market and country risk declines. The second is that we have assumed that betas measure exposure to country risk. A company like Embraer that derives the bulk of its revenues outside Brazil could argue that it is less exposed to country risk. In the preceding section, we introduced the concept of lambda and came up with two estimates for Embraer: 0.15, using revenue exposure, and 0.27, using the regression of Embraer stock returns against the Brazilian dollar bond. We will use the latter to compute the cost of equity:
The final point is that the cost of equity in dollar terms can be converted into a nominal Brazilian real (BR) cost of equity fairly simply by considering the differences in expected inflation rates in Brazil and the United States. For instance, if the expected inflation rate in Brazil is 6% and the expected inflation rate in the United States is 2%, the cost of equity in nominal BR is as follows:
Implicitly, we assume that BR risk-free rates around the world are the same with this approach and that the risk premium scales up with inflation as well. The alternative is to estimate a cost of equity from scratch, beginning with a nominal BR risk-free rate (which was 8% at the time of this analysis) and adding the premiums from before:
Substituting a real risk-free rate in the equation would yield a real cost of equity. Thus, if we assume that the real risk-free rate is 1.5% (the rate on an inflation-adjusted TIPS bond), the real cost of equity would have been:
COST OF EQUITY AND A SMALL FIRM PREMIUM Chapter 6 presented evidence of a small firm premium—small market-cap stocks earn higher returns than large market-cap stocks with equivalent betas. The magnitude and persistence of the small firm premium can be viewed as evidence that the capital asset pricing model understates the risk of smaller companies, and that a cost of equity based purely on a CAPM beta will therefore yield too low a number for these firms. There are some analysts who argue that you should therefore add a premium to the estimated cost of equity for smaller firms. Since small cap stocks have earned about 4 percent more than large cap stocks over the past few decades, you could consider this a reasonable estimate for the small firm premium. To estimate the cost of equity for a small cap stock with a beta of 1.2 (assuming a risk-free rate of 3.5 percent and a market risk premium of 4 percent), for instance, you would do the following:
We would introduce four notes of caution with this approach. First, this opens the door to a series of adjustments that you could make to the cost of equity, reflecting the numerous inefficiencies cited in Chapter 6. For instance, you could estimate a low PE premium, a low price-to-book premium, and a high dividend yield premium and add them all to the cost of equity. If our objective in valuation is to uncover market mistakes, it would be a mistake to start off with the presumption that markets are right in their assessments in the first place. Second, a better way of incorporating the small firm premium would be to identify the reasons for the premium and then develop more direct measures of risk. For instance, assume that the higher risk of small cap stocks comes from the higher operating leverage that these firms have, relative to their larger competitors. You could adjust the betas for operating leverage (as we did a few pages ago for Vans Shoes) and use the higher betas for small firms. Third, the small cap premium of 4 percent that we estimated from historical data comes with a significant standard error (of
155
approximately 2 percent). Thus, the true small cap premium can be 8 percent or 0 percent. Fourth, even if your company is a small company today and deserves a small cap premium, assuming a high growth rate for your firm will make it a large cap firm eventually. It follows that you would expect the small cap premium to fade over time.
FROM COST OF EQUITY TO COST OF CAPITAL Although equity is undoubtedly an important and indispensable ingredient of the financing mix for every business, it is but one ingredient. Most businesses finance some or much of their operations using debt or some security that is a combination of equity and debt. The costs of these sources of financing are generally very different from the cost of equity, and the cost of financing for a firm should reflect their costs as well, in proportion to their use in the financing mix. Intuitively, the cost of capital is the weighted average of the costs of the different components of financing—including debt, equity, and hybrid securities—used by a firm to fund its financial requirements. This section examines the process of estima ting the cost of financing other than equity, and the weights for computing the cost of capital.
Calculating the Cost of Debt The cost of debt measures the current cost to the firm of borrowing funds to finance projects. In general terms, it is determined by the following variables:
The riskless rate. As the riskless rate increases, the cost of debt for firms will also increase. The default risk (and associated default spread) of the company. As the default risk of a firm increases, the cost of borrowing money will also increase. Chapter 7 looked at how the default spread has varied across time and can vary across maturity. The tax advantage associated with debt. Since interest is tax deductible, the after-tax cost of debt is a function of the tax rate. The tax benefit that accrues from paying interest makes the after-tax cost of debt lower than the pretax cost. Furthermore, this benefit increases as the tax rate increases.
This section focuses on how best to estimate the default risk in a firm and to convert that default risk into a default spread that can be used to come up with a cost of debt.
Estimating the Default Risk and Default Spread of a Firm The simplest scenario for estimating the cost of debt occurs when a firm has long-term bonds outstanding that are widely traded. The market price of the bond in conjunction with its coupon and maturity can serve to compute a yield that is used as the cost of debt. For instance, this approach works for a firm that has dozens of outstanding bonds that are liquid and trade frequently. Some firms have bonds outstanding that do not trade on a regular basis. Since these firms are usually rated, we can estimate their costs of debt by using their ratings and associated default spreads. Thus, a firm with an A rating can be expected to have a cost of debt approximately 1.00 percent higher than the Treasury bond rate, since this is the spread typically paid by AA-rated firms. Many companies choose not to get rated and smaller firms and most private businesses fall into this category. Although ratings agencies have sprung up in many emerging markets, there are still a number of markets where companies are not rated on the basis of default risk. When there is no rating available to estimate the cost of debt, there are two alternatives: 1. Recent borrowing history. Many firms that are not rated still borrow money from banks and other financial institutions. By looking at the most recent borrowings made by a firm, we can get a sense of the types of default spreads being charged the firm and use these spreads to come up with a cost of debt. 2. Estimate a synthetic rating. An alternative is to play the role of a ratings agency and assign a rating to a firm based on its financial ratios; this rating is called a synthetic rating. To make this assessment, we begin with rated firms and examine the financial characteristics shared by firms within each ratings class. To illustrate, Table 8.1 lists the range of interest coverage ratios for small (less than $5 billion in market cap) nonfinancial service firms in each S&P ratings class.11 Table 8.1 Interest Coverage Ratios and Ratings: Low Market Cap Firms Source for raw data: Capital IQ, BondsOnline.com.
Interest Coverage Ratio Rating Spread
156
More than 12.5
AAA
0.50%
9.5 to 12.5
AA
0.65%
7.5 to 9.5
A+
0.85%
6 to 7.5
A
1.00%
4.5 to 6
A–
1.10%
3.5 to 4.5
BBB
1.60%
3 to 3.5
BB
3.35%
2.5 to 3
B+
3.75%
2 to 2.5
B
5.00%
1.5 to 2
B–
5.25%
1.25 to 1.5
CCC
8.00%
0.8 to 1.25
CC
10.00%
0.5 to 0.8
C
12.00%
Less than 0.5
D
15.00%
Now consider a small firm that is not rated but has an interest coverage ratio of 6.15. Based on this ratio, a synthetic rating of A would be assessed for the firm, and a default spread of 1.00% would be added to the risk-free rate to arrive at the pretax cost of debt. The interest coverage ratios tend to be lower for larger (market cap greater than $5 billion) firms for any given rating. Table 8.2 summarizes these ratios. Table 8.2 Interest Coverage Ratios and Ratings: High Market Cap Firms Source: Capital IQ, BondsOnline.com.
Interest Coverage Ratio Rating Spread More than 8.5
AAA
0.50%
6.5 to 8.5
AA
0.65%
5.5 to 6.5
A+
0.85%
4.25 to 5.5
A
1.00%
3 to 4.25
A–
1.10%
2.5 to 3
BBB
1.60%
2 to 2.5
BB
3.35%
1.75 to 2
B+
3.75%
1.5 to 1.75
B
5.00%
1.25 to 1.5
B–
5.25%
0.8 to 1.25
CCC
8.00%
0.65 to 0.8
CC
10.00%
0.2 to 0.65
C
12.00%
Less than 0.2
D
14.00%
This approach can be expanded to allow for multiple ratios and qualitative variables as well. Once a synthetic rating is assessed, it can be used to estimate a default spread, which when added to the risk-free rate yields a pretax cost of debt for the firm.
EXTENDING THE SYNTHETIC RATINGS APPROACH By basing the rating on the interest coverage ratio alone, we run the risk of missing the information that is available in the other financial ratios used by ratings agencies. The approach can be extended to incorporate other ratios. The first step would be to develop a score based on multiple ratios. For instance, the Altman Z score, which is used as a proxy for default risk, is a function of five financial ratios that are weighted to generate a Z score. The ratios used and their relative weights are usually estimated by looking at past defaults. The second step is to relate the level of the score to a bond rating, much as is done in Tables 8.1 and 8.2 with interest coverage ratios. In making this extension, though, note that complexity comes at a cost. While credit or Z scores may, in fact, yield better estimates of synthetic ratings than those based on interest coverage ratios, changes in ratings arising from these scores are much more difficult to explain than those based on interest coverage ratios. That is a reason to prefer the flawed but simpler ratings derived from interest coverage ratios.
157
Estimating a Tax Rate To estimate the after-tax cost of debt, consider the fact that interest expenses are tax deductible to the firm. While the computation is fairly simple and requires that the pretax cost be multiplied by (1 – tax rate), the question of what tax rate to use can be a difficult one to answer, because there are so many choices. For instance, firms often report an effective tax rate, estimated by dividing the taxes due by the taxable income. The effective tax rate, though, is usually very different from the marginal or statutory tax rate, which is the rate at which the last dollar of income is taxed. Since interest expenses save you taxes at the margin (they are deducted from your last dollar of income), the right tax rate to use is the marginal tax rate. The other caveat to keep in mind is that interest creates a tax benefit only if a firm has enough income to cover the interest expenses. Firms that have operating losses will not get a tax benefit from interest expenses, at least in the year of the loss. The after-tax cost of debt will be equal to the pretax cost of debt in that year. If you expect the firm to make money in future years, you would need to adjust the after-tax cost of debt for taxes in those years. We return to this issue and examine it in more detail in Chapter 10, where we look at the same issue in the context of estimating after-tax cash flows.
ILLUSTRATION 8.11: Estimating the Cost of Debt: Boeing in December 2000 Boeing was rated AA by S&P. Using the typical default spreads for AA-rated firms in December 2000, we could estimate the pretax cost for Boeing by adding the default spread of 1.00%12 to the riskless rate of 5%.
Boeing has an effective tax rate of 27%, but we use a marginal tax rate of 35% to estimate the after-tax cost of debt for Boeing.
Note that we will attach this after-tax cost of debt to all of Boeing's debt, short-term or long-term. While that may seem unfair, since Boeing could have borrowed short term a t lower rates, we are assuming that the rollover cost of short-term debt will approximate to the cost of long term debt. Furthermore, we do not want to systematically reward companies with short-term debt by giving them lower costs of capital. One final point about ratings. The ratings agencies rate both individual bond issues and entire companies. The rating used for the pretax cost of debt should be the rating for the company and not for an individual bond. Even a risky company can structure and issue a safe bond, and estimating a cost of debt based on that bond's rating will underestimate the overall cost of debt.
Estimating the Cost of Debt for an Emerging Market Firm In general, there are three problems that we run into when assessing the cost of debt for emerging market firms. The first is that most of these firms are not rated, leaving us with no option but to estimate the synthetic rating (and associated costs). The second is that the synthetic ratings may be skewed by differences in interest rates between the emerging market and the United States. Interest coverage ratios will usually decline as interest rates increase, and it may be far more difficult for a company in an emerging market to achieve the interest coverage ratios of companies in developed markets. Finally, the existence of country default risk hangs over the cost of debt of firms in that market. The second problem can be fixed fairly simply by either modifying the tables developed using U.S. firms or restating the interest expenses (and interest coverage ratios) in dollar terms. The question of country risk is a thornier one. Conservative analysts often assume that companies in a country cannot borrow at a rate lower than the country itself can borrow at. With this reasoning, the cost of debt for an emerging market company will include the country default spread for the country.
The counter to this argument is that companies may be safer than the countries in which they operate, and that they bear only a portion or perhaps even none of the country default spread.
ILLUSTRATION 8.12: Estimating the Cost of Debt: Embraer in March 2008 To estimate Embraer's cost of debt, we first estimated a synthetic rating for the firm. Based on its operating income of $527 million and interest expenses of $176 million in 2007, we arrived at an interest coverage ratio of 2.99 and a BBB rating. While the default spread for BBB rated bonds was only 1.50% at the time, there is the added consideration that Embraer is a Brazilian firm. Since the Brazilian dollar-denominated government bond had a default spread of 2.00% at the time of the analysis, you could argue that every Brazilian company should pay this premium in addition to its own default spread. With this reasoning, the pretax cost of debt for Embraer in U.S. dollars (assuming a Treasury bond rate is 3.8%) can be calculated:
158
Using a marginal tax rate of 34%, we can estimate an after-tax cost of debt for Embraer:
With this approach, the cost of debt for a firm can never be lower than the cost of debt for the country in which it operates. Note, though, that Embraer gets a significant portion of its revenues in dollars from contracts with non-Brazilian airlines. Consequently, it could reasonably argue that it is less exposed to risk than is the Brazilian government and should therefore command a lower cost of debt. Put differently, there are some companies (generally large companies with significant foreign operations) to which, rather than add the entire default spread for the country to the cost of debt, we may add only a portion.
ratings.xls: This spreadsheet allows you to estimate the synthetic rating and cost of debt for any firm.
Calculating the Cost of Hybrid Securities While debt and equity represent the fundamental financing choices available for firms, there are some types of financing that share characteristics with both debt and equity. These are called hybrid securities. This section considers how best to estimate the costs of such securities.
Cost of Preferred Stock Preferred stock shares some of the characteristics of debt (the preferred dividend is prespecified at the time of the issue and is paid out before the common dividend) and some of the characteristics of equity (the preferred dividend is not tax deductible). If preferred stock is viewed as perpetual (as it usually is), the cost of preferred stock can be written as follows:
This approach assumes the dividend is constant in dollar terms forever and that the preferred stock has no special features (convertibility, callability, etc.). If such special features exist, they will have to be valued separately to estimate the cost of preferred stock. In terms of risk, preferred stock is safer than common equity, because preferred dividends are paid before dividends on common equity. It is, however, riskier than debt since interest payments are made prior to preferred dividend payments. Consequently, on a pretax basis, it should command a higher cost than debt and a lower cost than equity.
ILLUSTRATION 8.13: Calculating the Cost of Preferred Stock: Ford in 2011 In April 2011, Ford Motor Company had preferred stock that paid a dividend of $1.875 annually and traded at $26.475 per share. The cost of preferred stock can be estimated as follows:
At the same time, Ford's cost of equity, using an estimated beta of 1.40, a risk-free rate of 3.5% and an equity risk premium of 5%, was 10.5%; its pretax cost of debt, based on its S&P rating of B+, was 8.50%, and its after-tax cost of debt was 5.10%. Not surprisingly, its preferred stock was less expensive than equity, but much more expensive than debt.
Calculating the Cost of Other Hybrid Securities A convertible bond is a bond that can be converted into equity at the option of the bondholder. A convertible bond can be viewed as a combination of a straight bond (debt) and a conversion option (equity). Instead of trying to calculate the cost of these hybrid securities individually, we can break down hybrid securities into their debt and equity components and treat the components separately.
ILLUSTRATION 8.14: Breaking Down a Convertible Bond into Debt and Equity Components: MGM Resorts In 2010, MGM Resorts, the casino company, issued 5-year convertible bonds with a coupon rate of 4.25% and a 10-year maturity. Since the firm was losing money, it was rated CCC+ by S&P and would have had to pay 10% if it had issued straight bonds at the same time. A year later, the bonds were trading at a price that was 112% of par, and the total par value of the convertible bond issue was $1.15 billion. The convertible bond can be broken down into straight bond and conversion option components.
The straight bond component of $818 is treated as debt, and has the same cost as the rest of debt. The conversion option of $302 is treated as equity, with the same cost of equity as other equity issued by the firm. For the entire bond issue of $1,150 million, with an overall market value
159
of $1,288 million, the value of debt is $916 million, and the value of equity is $372 million.
Calculating the Weights of Debt and Equity Components Now that we have the costs of debt, equity, and hybrid securities, we have to estimate the weights that should be attached to each. Before we discuss how best to estimate weights, we define what we include in debt. We then make the argument that weights used should be based on market value and not book value. This is so because the cost of capital measures the cost of issuing securities—stocks as well as bonds—to finance projects, and these securities are issued at market value, not at book value.
What Is Debt? The answer to this question may seem obvious since the balance sheet for a firm shows the outstanding liabilities of the firm. There are, however, limitations with using these liabilities as debt in the cost of capital computation. The first is that some of the liabilities on a firm's balance sheet, such as accounts payable and supplier credit, are not interest-bearing. Consequently, applying an after-tax cost of debt to these items can provide a misleading view of the true cost of capital for a firm. The second is that there are items off the balance sheet that create fixed commitments for the firm and provide the same tax deductions that interest payments on debt do. The most prominent of these off-balance sheet items are operating leases. Chapter 3 contrasted operating and capital leases and noted that operating leases are treated as operating expenses rather than financing expenses. Consider, though, what an operating lease involves. A retail firm leases a store space for 12 years and enters into a lease agreement with the owner of the space agreeing to pay a fixed amount each year for that period. We do not see much difference between this commitment and borrowing money from a bank and agreeing to pay off the bank loan over 12 years in equal annual installments. There are therefore two adjustments we will make when we estimate how much debt a firm has outstanding. 1. We will consider only interest-bearing debt rather than all liabilities. We would include both short-term and long-term borrowings in debt. 2. We will also capitalize operating leases and treat them as debt.
Capitalizing Operating Leases Converting operating lease expenses into a debt equivalent is straightforward. The operating lease commitments in future years, which are revealed in the footnotes to the financial statements for U.S. firms, should be discounted back at a rate that reflects their status as unsecured and fairly risky debt. As an approximation, using the firm's current pretax cost of borrowing as the discount rate yields a good estimate of the value of operating leases. There are still some countries where companies do not have to reveal their operating lease commitments to investors. When this is the case, you can get a reasonably close estimate of the debt value of operating leases by estimating the present value of an annuity equal to the current year's payment for a period that reflects a typical lease period (8 to 10 years). There is one final issue relating to capitalization. Earlier in this chapter it was stated that the interest coverage ratio could be used to estimate a synthetic rating for a firm that is not rated. For firms with little in terms of conventional debt and substantial operating leases, the interest coverage ratio used to estimate a synthetic rating has to be adapted to include operating lease expenses.
This ratio can then be used in conjunction with Tables 8.1 and 8.2 to estimate a synthetic rating.
ILLUSTRATION 8.15: The Debt Value of Operating Leases: Boeing in December 2000 Boeing has both conventional debt and operating lease commitments. This illustration estimates the “debt value” of Boeing's operating leases by taking the present value of operating lease expenses over time. To compute the present value of operating leases in the following table (in $millions), we use the pretax cost of borrowing for the firm, estimated in Illustration 8.11 to be 6%.
Year
Operating Lease Expense Present Value at 6%
1
$205
$193.40
2
$167
$146.83
3
$120
$100.75
4
$ 86
$ 68.12
5
$ 61
$ 45.58
160
6 to 15
$—
Present value of operating lease expenses
$ 0.00 $556.48
Thus, Boeing has $556 million more in debt than is reported in the balance sheet.
Oplease.xls: This spreadsheet allows you to convert operating lease expenses into debt.
Book Value versus Market Value Debt Ratios There are three standard arguments against using market value, and none of them is convincing. First, there are some financial managers who argue that book value is more reliable than market value because it is not as volatile. While it is true that book value does not change as much as market value, this is more a reflection of book value's weakness rather than its strength, since the true value of the firm changes over time as both firm-specific and market information is revealed. We would argue that market value, with its volatility, is a much better reflection of true value than is book value.13 Second, the defenders of book value also suggest that using book value rather than market value is a more conservative approach to estimating debt ratios. This assumes that market value debt ratios are always lower than book value debt ratios, an assumption not based on fact. Furthermore, even if the market value debt ratios are lower than the book value ratios, the cost of capital calculated using book value ratios will be lower than those calculated using market value ratios, making it a less conservative estimate, not more. To illustrate this point, assume that the market value debt ratio is 10 percent, while the book value debt ratio is 30 percent, for a firm with a cost of equity of 15 percent and an after-tax cost of debt of 5 percent. The cost of capital can be calcul ated as follows:
Third, it is claimed that lenders will not lend on the basis of market value, but this claim again seems to be based more on perception than on fact. Any homeowner who has taken a second mortgage on a house that has appreciated in value knows that lenders do lend on the basis of market value. It is true, however, that the greater the perceived volatility in the market value of an asset, the lower the borrowing potential on that asset.
Estimating the Market Values of Equity and Debt The market value of equity is generally the number of shares outstanding times the current stock price. If there are other equity claims in the firm such as warrants and management options, these should also be valued and added to the value of the equity in the firm. The market value of debt is usually more difficult to obtain directly, since very few firms have all their debt in the form of bonds outstanding trading in the market. Many firms have nontraded debt, such as bank debt, which is specified in book value terms but not market value terms. A simple way to convert book value debt into market value debt is to treat the entire debt on the books as one coupon bond, with a coupon set equal to the interest expenses on all the debt and the maturity set equal to the face-value weighted average maturity of the debt, and then to value this coupon bond at the current cost of debt for the company. Thus, the market value of $1 billion in debt, with interest expenses of $60 million (annually) and a maturity of six years, when the current cost of debt is 7.5 percent, can be estimated as follows:
ILLUSTRATION 8.16: Difference between Market Value and Book Value Debt Ratios: Boeing in June 2000 This illustration contrasts the book values of debt and equity with the market values. For debt, we estimate the market value of debt using the book value of debt, the interest expense on the debt, the average maturity of the debt, and the pretax cost of debt for each firm. For Boeing, the book value of debt is $6,972 million, the interest expense on the debt is $453 million, the average maturity of the debt is 13.76 years, and the pretax cost of debt is 6%. The estimated market value is:
161
To this, we need to add the present value of operating leases of $556 million to arrive at a total market value for debt of $7,847 million. The book value of equity for Boeing was $12,316 million while the market value of equity was $55,197 million. The debt ratios in market value and book value terms are computed as follows: Debt to equity
Market Value
Book Value
7,847/55,197 = 14.22%
6,972/12,316 = 56.61%
Debt/(Debt + Equity) 7,847/(7,847 + 55,197) = 12.45% 6,972/(6,972 + 12,316) = 36.15% The market debt ratio is significantly lower than the book debt ratio.
GROSS DEBT VERSUS NET DEBT Gross debt refers to all debt outstanding in a firm. Net debt is the difference between gross debt and the cash balance of the firm. For instance, a firm with $1.25 billion in interest-bearing debt outstanding and a cash balance of $1 billion has a net debt balance of $250 million. The practice of netting cash against debt is common in both Latin America and Europe, and debt ratios are usually estimated using net debt. It is generally safer to value a firm based on gross debt outstanding and to add the cash balance outstanding to the value of operating assets to arrive at the firm value. The interest payment on total debt is then entitled to the tax benefits of debt, and we can assess the effect of whether the company invests its cash balances efficiently on value. In some cases, especially when firms maintain large cash balances as a matter of routine, analysts prefer to work with net debt ratios. If you choose to use net debt ratios, you have to be consistent all the way through the valuation. To begin, the beta for the firm should be estimated using a net debt-to-equity ratio rather than a gross debt-to-equity ratio. The cost of equity that emerges from the beta estimate can be used to estimate a cost of capital, but the market value weight on debt should be based on net debt. Once you discount the cash flows of the firm at the cost of capital, you should not add back cash. Instead, you should subtract the net debt outstanding to arrive at the estimated value of equity. Implicitly, when you net cash against debt to arrive at net debt ratios, you are assuming that cash and debt have roughly similar risk. While this assumption may not be outlandish when analyzing highly rated firms, it becomes much shakier when debt becomes riskier. For instance, the debt in a BB-rated firm is much riskier than the cash balance in the firm, and netting out one against the other can provide a misleading view of the firm's default risk. In general, using net debt ratios will overstate the value of riskier firms.
wacccalc.xls: This spreadsheet allows you to convert book values of debt into market values.
Estimating the Cost of Capital Since a firm can raise its money from three sources—equity, debt, and preferred stock—the cost of capital is defined as the weighted average of each of these costs. The cost of equity (ke) reflects the riskiness of the equity investment in the firm, the after-tax cost of debt (kd) is a function of the default risk of the firm, and the cost of preferred stock (kps) is a function of its intermediate standing in terms of risk between debt and equity. The weights on each of these components should reflect their market value proportions, since these proportions best measure how the existing firm is being financed. Thus if E, D, and PS are the market values of equity, debt, and preferred stock respectively, the cost of capital can be written as follows:
ILLUSTRATION 8.17: Estimating Cost of Capital: Boeing in December 2000 Having estimated the costs of debt and equity in earlier illustrations, and the market value debt ratio in Illustration 8.16, we can put them together to arrive at a cost of capital for Boeing.
ILLUSTRATION 8.18: Estimating Cost of Capital: Embraer in March 2008 To estimate a cost of capital for Embraer, we again draw on the estimates of cost of equity and cost of debt we obtained in prior illustrations. The cost of capital will be estimated using gross debt ratios first in U.S. dollars:
162
The cost of capital for Embraer is estimated as follows:
To convert this into a nominal BR cost of capital, we would apply the differential inflation rates (6% in Brazil and 2% in the United States).
To estimate the cost of capital using the net debt ratio, we bring in the cash balance of 4,437 million BR that Embraer had at the time of the analysis:
Intuitively, the levered beta here is lower than the unlevered beta because we are incorporating the cash into the beta computation (with the assumption that cash is riskless).
Note that the weight on equity is greater than 100% (112.57%) and the weight on debt is negative (–12.57%) because net debt is negative. Notwithstanding these disconcerting inputs, the cost of capital is close to the cost of capital using the standard debt ratio approach, and the difference can be attributed to the fact that the net debt approach nets out the tax benefit of debt against the tax costs of earning interest income on cash.
BEST PRACTICES AT FIRMS We have spent this chapter discussing what firms should do when it comes to estimating the cost of capital. What do they actually do? Bruner, Eades, Harris, and Higgins (1998) surveyed 27 well-regarded corporations, and their findings are summarized in Table 8.3. Table 8.3 Current Practices for Estimating Cost of Capital Cost of Capital Item
Curre nt Practices 81% of firms used the capital asset pricing model to estimate the cost of equity, 4% used a modified Cost of equity capi tal asset pricing model, and 15% were uncertain about how they estimated the cost of equity. 70% of firms used 10-year Treasuries or longer as the riskless rate, 7% used 3- to 5-year Treasuries, and 4% used the Treasury bill rate. 52% used a published source for a beta estimate, while 30% estimated it themselves. There was wide variation in the market risk premium used, with 37% using a premium between 5% and 6%. 52% of firms used a marginal borrowing rate and a marginal tax rate, while 37% used the current Cost of debt average borrowing rate and the effective tax rate. 59% used market value weights for debt and equity in the cost of capital, 15% used book value weights, Weights for debt and equity and 19% were uncertain about what weights they used.
Source: Bruner, Eades, Harris, and Higgins (1998).
CONCLUSION When we analyze the investments of a firm or assess its value, we need to know the cost that the firm faces in raising equity, debt, and capital. The risk and return models described in earlier chapters can be used to estimate the costs of equity and capital for a firm. Building on the premise that the cost of equity should reflect the riskiness of equity to investors in the firm, there are three basic inputs we need to estimate the cost of equity for any firm. The riskless rate is the expected return on an investment with no default risk and no reinvestment risk. Since much of the analysis in corporate finance is long term, the riskless rate should be the interest rate on a long-term government bond. The risk p remium measures what investors demand as a premium for investing in risky investments instead of riskless investments. This risk premium, which can vary across investors, can be estimated either by looking at past returns on stocks and government securities or by looking at how the market prices stocks currently. The beta for a firm is conventionally measured using a regression of returns on the firm's stock against returns on a market index. This approach yields imprecise beta estimates, and we are better off estimating betas by examining the betas of the 163
businesses that the firm operates in. The cost of capital is a weighted average of the costs of the different components of financing, with the weights based on the market values of each component. The cost of debt is the market rate at which the firm can borrow, adjusted for any tax advantages of borrowing. The cost of preferred stock, however, is the preferred dividend yield. The cost of capital is useful at two levels. On a composite basis, it is what these firms have to make collectively on their investments to break even. It is also the appropriate discount rate to use to discount expected future cash flows to arrive at an estimate of firm value.
QUESTIONS AND SHORT PROBLEMS In the problems following, use an equity risk premium of 5.5 percent if none is specified and a tax rate of 40 percent where no tax rate is provided. 1. In December 1995, Boise Cascade's stock had a beta of 0.95. The Treasury bill rate at the time was 5.8%, and the Treasury bond rate was 6.4%. The firm had debt outstanding of $1.7 billion and a market value of equity of $1.5 billion; the corporate marginal tax rate was 36%. (The historical risk premium for stocks over Treasury bills is 8.5% and the risk premium for stocks over Treasury bonds is 5.5%.) a. Estimate the expected return on the stock for a short-term investor in the company. b. Estimate the expected return on the stock for a long-term investor in the company. c. Estimate the cost of equity for the company. 2. Continuing problem 1, Boise Cascade also had debt outstanding of $1.7 billion and a market value of equity of $1.5 billion; the corporate marginal tax rate was 36%. a. Assuming that the current beta of 0.95 for the stock is a reasonable one, estimate the unlevered beta for the company. b. How much of the risk in the company can be attributed to business risk and how much to financial leverage risk? 3. Biogen Inc., a biotechnology firm, had a beta of 1.70 in 1995. It had no debt outstanding at the end of that year. a. Estimate the cost of equity for Biogen if the Treasury bond rate is 6.4%. b. What effect will an increase in long-term bond rates to 7.5% have on Biogen's cost of equity? c. How much of Biogen's risk can be attributed to business risk? 4. Genting Berhad is a Malaysian conglomerate with holdings in plantations and tourist resorts. The beta estimated for the firm relative to the Malaysian stock exchange is 1.15, and the long-term government borrowing rate in Malaysia is 11.5%. (The Malaysian risk premium is 12% and the default spread on Malaysian local currency debt is 2%.) a. Estimate the expected return on the stock. b. If you were an international investor, what concerns, if any, would you have about using the beta estimated relative to the Malaysian Index? If you do have concerns, how would you modify the beta? 5. You have just done a regression of monthly stock returns of HeavyTech Inc., a manufacturer of heavy machinery, on monthly market returns over the past five years and have come up with the following regression:
The variance of the stock is 50%, and the variance of the market is 20%. The current T-bill rate is 3% (it was 5% one year ago). The stock is currently selling for $50, down $4 over the past year; it has paid a dividend of $2 during the past year and expects to pay a dividend of $2.50 over the next year. The NYSE Composite has gone down 8% over the past year, with a dividend yield of 3%. HeavyTech Inc. has a tax rate of 40%. a. What is the expected return on HeavyTech over the next year? b. What would you expect HeavyTech's price to be one year from today? 164
c. What would you have expected HeavyTech's stock returns to be over the past year? d. What were the actual returns on HeavyTech over the past year? e. HeavyTech has $100 million in equity and $50 million in debt. It plans to issue $50 million in new equity and retire $50 million in debt. Estimate the new beta. 6. Safecorp, which owns and operates grocery stores across the United States, currently has $50 million in debt and $100 million in equity outstanding. Its stock has a beta of 1.2. It is planning a leveraged buyout (LBO), where it will increase its debt-to-equity ratio of 8. If the tax rate is 40%, what will the beta of the equity in the firm be after the LBO? 7. Novell, which had a market value of equity of $2 billion and a beta of 1.50, announced that it was acquiring WordPerfect, which had a market value of equity of $1 billion and a beta of 1.30. Neither firm had any debt in its financial structure at the time of the acquisition, and the corporate tax rate was 40%. a. Estimate the beta for Novell after the acquisition, assuming that the entire acquisition was financed with equity. b. Assume that Novell had to borrow the $1 billion to acquire WordPerfect. Estimate the beta after the acquisition. 8. You are analyzing the beta for Hewlett Packard (HP) and have broken down the company into four broad business groups, with market values and betas for each group. Business Group
Market Value of Equity Beta
Mainframes
$2.0 billion
1.10
Personal computers $2.0 billion
1.50
Software
$1.0 billion
2.00
Printers
$3.0 billion
1.00
a. Estimate the beta for Hewlett Packard as a company. Is this beta going to be equal to the beta estimated by regressing past returns on HP stock against a market index? Why or why not? b. If the Treasury bond rate is 7.5%, estimate the cost of equity for Hewlett Packard. Estimate the cost of equity for each division. Which cost of equity would you use to value the printer division? c. Assume that HP divests itself of the mainframe business and pays the cash out as a dividend. Estimate the beta for HP after the divestiture. (HP had $1 billion in debt outstanding.) 9. The following table summarizes the percentage changes in operating income, percentage changes in revenue, and betas for four pharmaceutical firms.
a. Calculate the degree of operating leverage for each of these firms. b. Use the operating leverage to explain why these firms have different betas. 10. A prominent beta estimation service reports the beta of Comcast Corporation, a major cable TV operator, to be 1.45. The service claims to use weekly returns on the stock over the prior five years and the NYSE Composite as the market index to estimate betas. You replicate the regression using weekly returns over the same period and arrive at a beta estimate of 1.60. How would you reconcile the two estimates? 11. Battle Mountain is a mining company with gold, silver, and copper in mines in South America, Africa, and Australia. The beta for the stock is estimated to be 0.30. Given the volatility in commodity prices, how would you explain the low beta? 12. You have collected returns on AnaDone Corporation (AD Corp.), a large, diversified manufacturing firm, and the NYSE index for five years: Year AD Corp. NYSE 1981 10%
5%
1982 5%
15%
1983 –5%
8%
165
1984 20%
12%
1985 –5%
–5%
a. Estimate the intercept (alpha) and slope (beta) of the regression. b. If you bought stock in AD Corp. today, how much would you expect to make as a return over the next year? (The six-month T-bill rate is 6%.) c. Looking back over the past five years, how would you evaluate AD Corp.'s performance relative to the market? d. Assume now that you are an undiversified investor and that you have all of your money invested in AD Corp. What would be a good measure of the risk that you are taking on? How much of this risk would you be able to eliminate if you diversify? e. AD Corp. is planning to sell off one of its divisions. The division under consideration has assets that comprise half of the book value of AD Corp. and 20% of the market value. Its beta is twice the average beta for AD Corp. (before divestment). What will the beta of AD Corp. be after divesting this division? 13. You run a regression of monthly returns of Mapco Inc., an oil- and gas-producing firm, on the S&P 500 index, and come up with the following output for the period 1991 to 1995:
There are 20 million shares outstanding, and the current market price is $2 per share. The firm has $20 million in debt outstanding. (The firm has a tax rate of 36%.) a. What would an investor in Mapco's stock require as a return if the T-bond rate is 6%? b. What proportion of this firm's risk is diversifiable? c. Assume now that Mapco has three divisions of equal size (in market value terms). It plans to divest itself of one of the divisions for $20 million in cash and acquire another for $50 million (it will borrow $30 million to complete this acquisition). The division it is divesting is in a business line where the average unlevered beta is 0.20, and the division it is acquiring is in a business line where the average unlevered beta is 0.80. What will the beta of Mapco be after this acquisition? 14. You have just run a regression of monthly returns of American Airlines (AMR Corporation) against the S&P 500 over the past five years. You have misplaced some of the output and are trying to derive it from what you have. a. You know the R-squared of the regression is 0.36, and that your stock has a variance of 67%. The market variance is 12%. What is the beta of AMR? b. You also remember that AMR was not a very good investment during the period of the regression and that it did worse than expected (after adjusting for risk) by 0.39% a month for the five years of the regression. During this period, the average risk-free rate was 4.84%. What was the intercept on the regression? c. You are comparing AMR to another firm, which also has an R-squared of 0.48. Will the two firms have the same beta? If not, why not? 15. You have run a regression of monthly returns on Amgen, a large biotechnology firm, against monthly returns on the S&P 500 index, and come up with the following output:
The current one-year Treasury bill rate is 4.8% and the current 30-year bond rate is 6.4%. The firm has 265 million shares outstanding, selling for $30 per share. a. What is the expected return on this stock over the next year? b. Would your expected return estimate change if the purpose was to get a discount rate to value the company? c. An analyst has estimated, correctly, that the stock did 51.1% better than expected, annually, during the period of the regression. Can you estimate the annualized risk-free rate that she used for her estimate? 166
d. The firm has a debt/equity ratio of 3% and faces a tax rate of 40%. It is planning to issue $2 billion in new debt and acquire a new business for that amount, with the same risk level as the firm's existing business. What will the beta be after the acquisition? 16. You have just run a regression of monthly returns on MAD Inc., a newspaper and magazine publisher, against returns on the S&P 500, and arrived at the following result:
The regression has an R-squared of 22%. The current T-bill rate is 5.5%, and the current T-bond rate is 6.5%. The risk-free rate during the period of the regression was 6%. Answer the following questions relating to the regression: a. Based on the intercept, how well or badly did MAD do, relative to expectations, during the period of the regression? b. You now realize that MAD Inc. went through a major restructuring at the end of last month (which was the last month of your regression), and made the following changes: The firm sold off its magazine division, which had an unlevered beta of 0.6, for $20 million. It borrowed an additional $20 million, and bought back stock worth $40 million. After the sale of the division and the share repurchase, MAD Inc. had $40 million in debt and $120 million in equity outstanding. If the firm's tax rate is 40%, reestimate the beta after these changes. 17. Time Warner Inc., the entertainment conglomerate, had a beta of 1.61 in 1995. Part of the reason for the high beta was the debt left over from the leveraged buyout of Time by Warner in 1989, which amounted to $10 billion in 1995. The market value of equity at Time Warner in 1995 was also $10 billion. The marginal tax rate was 40%. a. Estimate the unlevered beta for Time Warner. b. Estimate the effect of reducing the debt ratio by 10% each year for the next two years on the beta of the stock. 18. Chrysler, the automotive manufacturer, had a beta of 1.05 in 1995. It had $13 billion in debt outstanding in that year, and 355 million shares trading at $50 per share. The firm had a cash balance of $8 billion at the end of 1995. The marginal tax rate was 36%. a. Estimate the unlevered beta of the firm. b. Estimate the effect of paying out a special dividend of $5 billion on this unlevered beta. c. Estimate the beta for Chrysler after the special dividend. 19. You are trying to estimate the beta of a private firm that manufactures home appliances. You have managed to obtain betas for publicly traded firms that also manufacture home appliances.
The private firm has a debt equity ratio of 25% and faces a tax rate of 40%. The publicly traded firms all have marginal tax rates of 40% as well. a. Estimate the beta for the private firm. b. What concerns, if any, would you have about using betas of comparable firms? 20. As the result of stockholder pressure, RJR Nabisco is considering spinning off its food division. You have been asked to estimate the beta for the division, and decide to do so by obtaining the beta of comparable publicly traded firms. The average beta of comparable publicly traded firms is 0.95,and the average debt-to-equity ratio of these firms is 35%. The division is expected to have a debt ratio of 25%. The marginal corporate tax rate is 36%. a. What is the beta for the division? b. Would it make any difference if you knew that RJR Nabisco had a much higher fixed cost structure than the 167
comparable firms used here? 21. Southwestern Bell, a phone company, is considering expanding its operations into the media business. The beta for the company at the end of 1995 was 0.90, and the debt-to-equity ratio was 1. The media business is expected to be 30% of the overall firm value in 1999, and the average beta of comparable firms is 1.20; the average debt-to-equity ratio for these firms is 50%. The marginal corporate tax rate is 36%. a. Estimate the beta for Southwestern Bell in 1999, assuming that it maintains its current debt-to-equity ratio. b. Estimate the beta for Southwestern Bell in 1999, assuming that it decides to finance its media operations with a debt-to-equity ratio of 50%. 22. The chief financial officer of Adobe Systems, a growing software manufacturing firm, has approached you for some advice regarding the beta of his company. He subscribes to a service that estimates Adobe Systems' beta each year, and he has noticed that the beta estimates have gone down every year since 1991—from 2.35 in 1991 to 1.40 in 1995. He would like the answers to the following questions: a. Is this decline in beta unusual for a growing firm? b. Why would the beta decline over time? c. Is the beta likely to keep decreasing over time? 23. You are analyzing Tiffany & Company, an upscale retailer, and find that the regression estimate of the firm's beta is 0.75; the standard error for the beta estimate is 0.50. You also note that the average unlevered beta of comparable specialty retailing firms is 1.15. a. If Tiffany has a debt/equity ratio of 20%, estimate the beta for the company based on comparable firms. (The tax rate is 40%.) b. Estimate a range for the beta from the regression. c. Assume that Tiffany is rated BBB and that the default spread for BBB-rated firms is 1% over the Treasury bond rate. If the Treasury bond rate is 6.5%, estimate the cost of capital for the firm. 24. You have been asked to estimate the cost of capital for NewTel, a telecom firm. The firm has the following characteristics: There are 100 million shares outstanding, trading at $250 per share. The firm has a book value of debt with a maturity of six years of $10 billion, and interest expenses of $600 million on the debt. The firm is not rated, but it had operating income of $2.5 billion last year. (Firms with an interest coverage ratio of 3.5 to 4.5 were rated BBB, and the default spread was 1%.) The tax rate for the firm is 35%. The Treasury bond rate is 6%, and the unlevered beta of other telecom firms is 0.80. a. Estimate the market value of debt for this firm. b. Based on the synthetic rating, estimate the cost of debt for this firm. c. Estimate the cost of capital for this firm. The regression is sometimes calculated using returns in excess of the riskless rate for both the stock and the market. In that case, the intercept of the regression should be zero if the actual returns equal the expected returns from the CAPM, greater than zero if the stock does better than expected, and less than zero if it does worse than expected. 1
The terminology is confusing, since the intercept of the regression is sometimes also called the alpha and is sometimes compared to zero as a measure of risk-adjusted performance. The intercept can be compared to zero only if the regression is run with excess returns for both the stock and the index; the riskless rate has to be subtracted from the raw return in each month for both. 2
The stock has to be bought by a day called the ex-dividend day in order for investors to be entitled to dividends. The returns in a period include dividends if the ex-dividend day is in that period. 3
4
This is done purely for computational convenience.
The nontrading bias arises because the returns in nontrading periods are zero (even though the market may have moved up or down significantly in those periods). Using these nontrading period returns in th e regression will 5
168
reduce the correlation between stock returns and market returns and the beta of the stock. The bias can also be reduced using statistical techniques suggested by Dimson (1979) and S choles, and Williams (1977). 6
This formula was origi nally developed by Hamada in 1972. There are two common modifications. One is to ignore the tax effects and compute the levered beta as: 7
If debt has market risk (i.e., its beta is greater than zero), the original formula can be modified to take this into account. If the beta of debt is β D, the beta of equity can be written as:
Note that Boeing breaks its business down in its financial statements into these two segments. We could have used operating income or EBITDA and a typical multiple to arrive at value. 8
To estimate these multiples, we looked at the market value of publicly traded firms relative to their revenues. This is a ratio of enterprise value to revenues. 9
We used total debt in making this estimate. We discuss the alternate practice of using net debt ratios (obtai ned by netting cash out against debt) later in this chapter. 10
This table was updated in early 2011 by listing out all rated firms with market capitalization lower than $5 billion and t heir interest coverage ratios, and then sorting firms based on their bond ratings. The ranges were adjusted to eliminate outliers and to prevent overlapping ranges. 11
12
The default spread was obtained from the rating/spread table in 2000.
There are some who argue that stock prices are much more volatile than the underlying true valu e. Even if this argument is justified (and it has not conclusively been shown to be so), the difference between market value and true value is likely to be much smaller than the difference between book value and true value. 13
169
CHAPTER 9 Measuring Earnings To estimate cash flows, we usually begin with a measure of earnings. Free cash flows to the firm, for instance, are based on after-tax operating earnings. Free cash flows to equity estimates, on the other hand, commence with net income. While we obtain measures of operating and net income from accounting statements, the accounting earnings for many firms bear little or no resemblance to the true earnings of the firm. This chapter begins by considering the philosophical difference between the accounting and financial views of firms. We then consider how the earnings of a firm, at least as me asured by accountants, have to be adjusted to get a measure of earnings that is more appropriate for valuation. In particular, we examine how to treat operating lease expenses, which we argue are really financial expenses, and research and development expenses, which we consider to be capital expenses. The adjustments affect not only our measures of earnings but our estimates of book value of capital. We also look at extraordinary items (both income and expenses) and one-time charges, the use of which has expanded significantly in recent years as firms have shifted toward managing earnings more aggressively. The techniques used to smooth earnings over periods and beat analyst estimates can skew reported earnings, and, if we are not careful, the values that emerge from them.
ACCOUNTING VERSUS FINANCIAL BALANCE SHEETS When analyzing a firm, what are the questions to which we would like to know the answers? A firm, as defined here, includes both investments already made—assets in place—and investments yet to be made—growth assets. In addition, a firm can either borrow the funds it needs to make these investments, in which case it is using debt, or raise it from its owners in the form of equity. Figure 9.1 summarizes this description of a firm in the form of a financial balance sheet. Figure 9.1 A Financial Balance Sheet
Note that while this summary does have some similarities with the accounting balance sheet, there are key differences. The most important one is that here we explicitly consider growth assets when we look at what a firm owns. When doing a financial analysis of a firm, we would like to be able to answer a number of questions relating to each of these items. Figure 9.2 lists the questions. As we see in this chapter, accounting statements allow us to acquir e some information about each of these questions, but they fall short in terms of both the timeliness with which they provide it and the way in which they measure asset value, earnings, and risk. Figure 9.2 Key Financial Questions
ADJUSTING EARNINGS The income statement for a firm provides measures of both the operating and equity income of the firm in the form o f the earnings before interest and taxes (EBIT) and net income. When valuing firms, there are two important considerations in using these measures. One is to obtain as updated an estimate as possible, given how much firms change over time. The second is that reported earnings at these firms may bear little resemblance to true earnings 170
because of limitations in accounting rules and the fi rms' own actions.
Importance of Updating Earnings Firms reveal their earnings in their financial statements and annual re ports to stockholders. Annual reports are released only at the end of a firm's financial year, but you are often required to value firms all through the year. Consequently, the last annual report that is available for a firm being valued can contain information that is several months old. In the case of firms that are changing rapidly over time, it is dangerous to base value estimates on information that is this old. Instead, use more recent information. Since firms in the United States are required to file quarterly reports with the Securities and Exchange Commission (10-Qs) and reveal these reports to the public, a more recent estimate of key items in the financial statements can be obtained by aggregating the numbers over the most recent four quarters. The estimates of revenues and earnings that emerge from this exercise are called trailing 12-month revenues and earnings and can be very different from the values for the same variables in the most recent annual report. There is a price paid for the updating. Unfortunately, not all items in the annual report are updated in the quarterly reports. You have to either use the numbers in the last annual report (which does lead to inconsistent inputs) or estimate their values at the end of the last quarter (which leads to estimation error). For example, many firms do not reveal details about options outstanding (issued to managers and employees) in quarterly reports, while they do reveal them in annual reports. Since you need to value these options, you can use the options outstanding as of the last annual report, or assume that the options outstanding today have changed to reflect changes in the other variables. (For instance, if revenues have doubled, you can assume that the options have doubled as well.) For younger firms, it is critical that you stay with the most updated numbers you can find, even if these numbers are estimates. These firms are often growing exponentially, and using numbers from the last financial year will lead to misleading estimates of value. Even firms that are not growing are changing substantially from quarter to quarter, and updated information might give you a chance to capture these changes. There are several financial markets where firms still file financial reports only once a year, thus denying us the option of using quarterly updates. When valuing firms in these markets, you may have to draw on unofficial sources to update their valuations.
ILLUSTRATION 9.1: Updated Earnings for Apple—April 2011 Assume that you were valuing Apple in April 2011. The last 10-K was as of September 2010 and the firm had released two quarterly reports (10-Qs), one ending in December 2010 and one ending in March 2011. To illustrate how much the fundamental inputs to the valuation have changed in the six months, the information in the last 10-K is compared to the trailing 12-month information in the latest 10-Q for revenues, operating income, R&D expenses, and net income (in millions of dollars).
The trailing 12-month revenues are 40 percent higher than the revenues reported in the latest 10-K, and the firm's operating income and net income have both increased substantially. Apple in April 2011 was a much more profitable firm than Apple in September 2010. Note that these are not the only inputs that will change. For younger firms, the number of shares outstanding can also change dramatically from period to period. Using the most updated numbers will give you a more realistic valuation.
Correcting Earnings Misclassification Companies have three types of expenses: 1. Operating expenses are expenses that generate benefits for the firm only in the current period. For instance, the fuel used by an airline in the course of its flights is an operating expense, as is the labor cost for an automobile company associated with producing vehicles. 2. Capital expenses are expenses that generate benefits over multiple periods. For example, the expense associated with building and outfitting a new factory for an automobile manufacturer is a capital expense, since it will generate several years of revenues. 3. Financial expenses are expenses associated with nonequity capital raised by a firm. Thus, the interest paid on a bank loan would be a financial expense. The operating income for a firm, measured correctly, should be equal to its revenues less its operating expenses. Neither financial nor capital expenses should be included in the operating expenses in the year that they occur, though capital expenses may be depreciated or amortized over the periods that the firm obtains benefits from the 171
expenses. The net income of a firm should be its revenues less both its operating and financial expenses. The accounting measures of earnings can be misleading because operating, capital, and financial expenses are sometimes misclassified. This section considers the two most common misclassifications and how to correct for them. The first is the inclusion of capital expenses such as research and development (R&D) in operating expenses, which skews the estimation of both operating and net income. The second adjustment is for financial expenses such as operating lease expenses that are treated as operating expenses. This affects the measurement of operating income and free cash flows to the firm. The other factor to consider is the effect of the phenomenon of so-called managed earnings at these firms. Firms sometimes use accounting techniques to post earnings that beat analyst estimates, resulting in misleading measures of earnings.
Capital Expenses Treated as Operating Expenses While in theory operating income in after only operating expenses, the reality is that there are a number of capital expenses that are treated as operating expenses. For instance, a significant shortcoming of accounting statements is the way in which they treat research and development expenses. Using the rationale that the products of research are too uncertain and difficult to quantify, accounting standards have generally required that all R&D expenses be expensed in the period in which they occur. This has several consequences, but one of the most profound is that the value of the assets created by research does not show up on the balance sheet as part of the total assets of the firm. This, in turn, creates ripple effects for the measurement of capital and profitability ratios for the firm. We consider how to capitalize R&D expenses in the first part of the section and extend the argument to other capital expenses in the second part of the section.
Capitalizing R&D Expenses Research expenses, notwithstanding the uncertainty about future benefits, should be capitalized. To capitalize and value research assets, we make an assumption about how long it takes for research and development to be converted, on average, into commercial products. This is called the amortizable life of these assets. This life will vary across firms and reflect the time involved in converting research into products. To illustrate, research and development expenses at a pharmaceutical company should have fairly long amortizable lives, since the approval process for new drugs is long. In contrast, research and development expenses at a software firm, where products tend to emerge from research much more quickly, should be amortized over a shorter period. Once the amortizable life of research and development expenses has been estimated, the next step is to collect data on R&D expenses over past years ranging back over the amortizable life of the research asset. Thus, if the research asset has an amortizable life of five years, the R&D expenses in each of the five years prior to the current one have to be obtained. For simplicity, it can be assumed that the amortization is uniform over time, which leads to the following estimate of the residual value of the research asset today:
Thus, in the case of the research asset with a five-year life, you cumulate one-fifth of the R&D expenses from four years ago, two-fifths of the R&D expenses from three years ago, three-fifths of the R&D expenses from two years ago, four-fifths of the R&D expenses from last year, and this year's entire R&D expense to arrive at the value of the research asset. This augments the value of the assets of the firm and, by extension, the book value of equity.
Finally, the operating income is adjusted to reflect the capitalization of R&D expenses. First, the R&D expenses that were subtracted out to arrive at the operating income are added back to the operating income, reflecting their recategorization as capital expenses. Next, the amortization of the research asset is treated the same way that depreciation is and netted out to arrive at the adjusted operating income:
This adjustment will generally increase operating income for firms that have R&D expenses that are growing over time. The net income will also be affected by this adjustment:
While we would normally consider only the after-tax portion of this amount, the fact that R&D is entirely tax deductible eliminates the need for this adjustment.1 172
R&DConv.xls: This spreadsheet allows you to convert R&D expenses from operating to capital expenses.
ILLUSTRATION 9.2: Capitalizing R&D Expenses: Amgen in March 2009 Amgen is a biotechnology firm. Like mo st pharmaceutical firms, it has a substantial amount of R&D expenses, and we attempt to capitalize it in this section. The first step in this conversion is determining an amortizable life for R&D expenses. How long will it take, on an expected basis, for research to pay off at Amgen? Given the length of the approval process for new drugs by the Food and Drug Administration, we assume that this amortizable life is 10 years. The second step in the analysis is collecting research and development expenses from prior years, with the number of years of historical data being a function of the amortizable life. The following table provides this information (in millions of dollars) for each of the years:
Year
R&D Expenses
Current (2008) $3,030 –1
$3,266
–2
$3,366
–3
$2,314
–4
$2,028
–5
$1,655
–6
$1,117
–7
$ 864
–8
$ 845
–9
$ 823
–10
$ 663
The current year's information reflects the R&D in the most recent financial year (which was calendar year 2008). The portion of the expenses in prior years that would have been amortized already and the amortization this year from each of these expenses is considered. To make estimation simpler, these expenses are amortized linearly over time; with a 10-year life, 10% is amortized each year. This allows you to estimate the value of the research asset created at each of these firms, and the amortization of R&D expenses in the current year. The procedure is illustrated in the following table:
Note that none of the current year's expenditure has been amortized because it is assumed to occur at the end of the year (which is assumed to be right now) but that 50 percent of the expense from five years ago has been amortized. The sum of the dollar values of unamortized R&D from prior years is $13.284 billion. This can be viewed as the value of Amgen's research asset and would be also added to the book value of equity for computing return on equity and capital measures. The sum of the amortization in the current year for all prior year expenses is $1,694 million. The final step in the process is the adjustment of the operating income to reflect the capitalization of research and development expenses. We make the adjustment by adding back current year's R&D expenses to the operating income (to reflect its reclassification as a capital expense) and subtracting out the amortization of the research asset, estimated in the last step. For Amgen, which reported operating income of $5,594 million in its income statement for 2008, the adjusted operating earnings would be:
The stated net income of $4,196 million can be adjusted similarly.
In the last section, we explained why there is no tax effect to consider.
173
Both the book value of equity and capital are augmented by the value of the research asset. Since measures of return on capital and equity are based on the prior year's values, we computed the value of the research asset at the end of 2007, using the same approach that we used in 2008.
The returns on equity and capital are reported with both the unadjusted and adjusted numbers:
Unadjusted Return on equity
Adjusted for R&D
4,196/17,869 = 23.48% 5,532/29,817 = 18.55%
Pretax return on capital 5,594/21,985 = 25.44% 6,930/33,843 = 20.48% While the profitability ratios for Amgen remain impressive even after the adjustment, they decline from the unadjusted numbers. This is likely to happen for most firms that earn high returns on equity and capital and have substantial R&D expenses.2
Capitalizing Other Operating Expenses While R&D expenses are the most prominent example of capital expenses being treated as operating expenses, there are other operating expenses that arguably should be treated as capital expenses. Consumer product companies such as Gillette and Coca-Cola could argue that a portion of advertising expenses should be treated as capital expenses, since they are designed to augment brand name value. For a consulting firm, the cost of recruiting and training its employees could be considered a capital expense, since the consultants who emerge are likely to be the heart of the firm's assets and provide benefits over many years. For some technology firms, including e -tailers such as Amazon.com, the biggest operating expense item is selling, general, and administrative expenses (SG&A). These firms could argue that a portion of these expenses should be treated as capital expenses, since they are designed to increase brand name awareness and bring in new customers. While this argument has some merit, you should remain wary about using it to justify capitalizing these expenses. For an operating expense to be capitalized, there should be substantial evidence that the benefits from the expense accrue over multiple periods. Does a customer who is enticed to buy from Amazon, based on an advertisem*nt or promotion, continue as a customer for the long term? There are some analysts who claim that this is indeed the case, and attribute significant value added to each new customer.3 It would be logical, under those circ*mstances, to capitalize these expenses using a procedure similar to that used to capitalize R&D expenses. Determine the period over which the benefits from the operating expense (such as SG&A) will flow. Estimate the value of the asset (similar to the research asset) created by these expenses. If the expenses are SG&A expenses, this would be the SG&A asset. Adjust the operating income for the expense and the amortization of the created asset.
A similar adjustment has to be made to net income:
As with the research asset, the capitalization of these expenses will create an asset that augments the book value of equity (and capital).
ILLUSTRATION 9.3: Should You Capitalize SG&A Expense? Analyzing Amazon.com Let us consider SG&A expenses at Amazon. To make a judgment on whether you should capitalize these expenses, you need to get a sense of what these expenses are and how long the benefits accruing from these expenses last. For instance, assume that an Amazon promotion (the expense of which would be included in SG&A) attracts new customers to the web site, and that customers, once they try Amazon, continue, on average, to be customers for three years. You would then use a three year amortizable life for SG&A expenses, and capitalize them the same way you capitalized R&D: by collecting historical information on SG&A expenses, amortizing them each year, estimating the value of the selling asset, and then adjusting operating income. We do believe, on balance, that selling, general, and administrative expenses should continue to be treated as operating expenses and not capitalized for Amazon for two reasons. First, retail customers are difficult to retain, especially online, and Amazon faces serious competition not only from other online retailers but also from traditional retailers like Wal Mart setting up their online operations. Consequently, the customers that Amazon might attract with its advertising or sales promotions are unlikely to stay for an extended period just because of the initial inducements. Second, as the company has become larger, its selling, general, and administrative expenses seem increasingly directed toward generating revenues in current periods rather than future periods.
174
ILLUSTRATION 9.4: Capitalizing Recruitment and Training Expenses: Cyber Health Consulting Cyber Health Consulting (CHC) is a firm that specializes in offering management consulting services to health-care firms. CHC reported operating income (EBIT) of $51.5 million and net income of $23 million in the most recent year. However, the firm's expenses include the cost of recruiting new consultants ($5.5 million) and the cost of training ($8.5 million). A consultant who joins CHC stays with the firm, on average, four years. To capitalize the cost of recruiting and training, we obtained these costs from each of the prior four years. The following table reports on these human capital expenses, and amortizes each of these expenses over four years.
The adjustments to operating and net income are as follows:
As with R&D expenses, the fact that training and recruiting expenses are fully tax deductible dispenses with the need to consider the tax effect when adjusting net income.
Adjustments for Financing Expenses The second adjustment is for financing expenses that accountants treat as operating expenses. The most significant example is operating lease expenses, which are treated as operating expenses, in contrast to capital leases, which are presented as debt.
Converting Operating Leases into Debt In Chapter 8, the basic approach for converting operating leases into debt was presented. You discount future operating lease commitments back at the firm's pretax cost of debt. The present value of the operating lease commitments is then added to the conventional debt of the firm to arrive at the total debt outstanding.
Once operating leases are recategorized as debt, the operating incomes can be adjusted in two steps. First, the operating lease expense is added back to the operating income, since it is a financial expense. Next, the depreciation on the leased asset is subtracted out to arrive at adjusted operating income:
If you assume that the depreciation on the leased asset approximates the principal portion of the debt being repaid, the adjusted operating income can be computed by adding back the imputed interest expense on the debt value of the operating lease expense:
ILLUSTRATION 9.5: Adjusting Operating Income for Operating Leases: The Gap in 2011 As a specialty retailer, the Gap has hundreds of stores that are leased, with the leases being treated as operating leases. For the most recent financial year (2010), the Gap has operating lease expenses of $1,129 million. The following table presents the operating lease commitments for the firm over the next five years and the lump sum of commitments beyond that point in time.
Year
Commitment
1
$ 997
2
$ 841
3
$ 710
175
4
$ 602
5
$ 483
6 and beyond $1,483 The Gap, based on its S&P bond rating of BB+, has a pretax cost of debt of 5.5%. To compute the present value of the commitments, you have to make a judgment on the lump sum commitment in year 6. Based on the average annual lease commitment over the first five years ($727 million), we arrive at an annuity of two years:4
The present values of the commitments at the 5.5% pretax cost of debt are estimated in the following table:
Year
Commitment Present Value
1
$997
$ 945.02
2
$841
$ 755.60
3
$710
$ 604.65
4
$602
$ 485.94
5
$483
$ 369.56
6 and 7
$741.40
$1,047.50
Debt value of leases
$4,208.28
The present value of operating leases is treated as the equivalent of debt, and is added onto the conventional debt of the firm. The Gap has no interest-bearing debt on its balance sheet. The cumulated debt for the firm is:
To adjust the operating income for the Gap, we first use the full adjustment. To compute depreciation on the leased asset, we assume straightline depreciation over the lease life5 (7 years) on the value of the leased asset, which is equal to the debt value of the lease commitments:
The Gap's stated operating income of $1,968 million is adjusted as follows:
The approximate adjustment is also estimated as follows, where we add the added imputed interest expense using the pretax cost of debt:
Oplease.xls: This spreadsheet allows you to convert operating lease expenses into debt.
WHAT ABOUT OTHER COMMITMENTS? The argument made about leases can be made about other long-term commitments where a firm has no escape hatches or cancellation options, or where the payment is not connected to performance/earnings. For instance, consider a professional sports team that signs a star player to a 10-year contract, agreeing to pay $5 million a year. If the payment is not contingent on performance, this firm has created the equivalent of debt by signing this contract. The upshot of this argument is that firms that have no debt on their balance sheet may still be highly levered and subject to default risk as a consequence. For instance, Mario Lemieux, a star player for the Pittsburgh Penguins, the professional ice hockey team, was given partial ownership of the team because of its failure to meet contractual commitments it had made to him.
Accounting Earnings and True Earnings Firms have become particularly adept at meeting and beating analyst estimates of earnings each quarter. While beating earnings estimates can be viewed as a positive development, some firms adopt accounting techniques that are questionable to accomplish this objective. When valuing these firms, you have to correct operating income for these accounting manipulations to arrive at the correct operating income.
The Phenomenon of Managed Earnings 176
In the 1990s, firms like Microsoft and Intel set the pattern for technology firms. In fact, Microsoft beat analyst estimates of earnings in 39 of the 40 quarters during the decade, and Intel posted a record almost as impressive. As the market values of these firms skyrocketed, other technology firms followed in their footsteps in trying to deliver earnings that were higher than analyst estimates by at least a few pennies. The evidence is overwhelming that the phenomenon is spreading. For an unprecedented 18 quarters in a row from 1996 to 2000, more firms beat consensus earnings estimates than missed them.6 In another indication of the management of earnings, the gap between the earnings reported by firms to the Internal Revenue Service and that reported to equity investors has been growing over the past decade. Given that these analyst estimates are expectations, what does this tell you? One possibility is that analysts consistently underestimate earnings and never learn from their mistakes. While this is a possibility, it seems extremely unlikely to persist over an entire decade. The other is that technology firms particularly have far more discretion in how they measure and report earnings and are using this discretion to beat estimates. In particular, the treatment of research expenses as operating expenses gives these firms an advantage when it comes to managing earnings. Does managing earnings really increase a firm's stock price? It might be possible to beat analysts quarter after quarter, but are markets as gullible? They are not, and the advent of so-called whispered earnings estimates is in reaction to the consistent delivery of earnings that are above expectations. What are whispered earnings? Whispered earnings are implicit earnings estimates that firms have to beat to surprise the market, and these estimates are usually a few cents higher than analyst estimates. For instance, on April 10, 1997, Intel reported earnings per share of $2.10 per share, higher than analyst estimates of $2.06 per share, but saw its stock price drop 5 points because the whispered earnings estimate had been $2.15. In other words, markets had built into expectations the amount by which Intel had beaten earnings estimates historically.
Why Do Firms Manage Earnings? Firms generally manage earnings because they believe that they will be rewarded by markets for delivering earnings that are smoother and come in consistently above analyst estimates. As evidence, they point to the success of firms like Microsoft and Intel, and the brutal punishment meted out for firms that do not meet expectations. Many financial managers also seem to believe that investors take earnings numbers at face value, and the managers work at delivering bottom lines that reflect this belief. This may explain why any efforts by the Financial Accounting Standards Board (FASB) to change the way earnings are measured are fought with vigor, even when the changes make sense. For instance, any attempts by FASB to value the options granted by firms to their managers at a fair value and charge them against earnings or change the way mergers are accounted for have been consistently opposed by technology firms. It may also be in the best interests of the managers of firms to manage earnings. Managers know that they are more likely to be fired when earnings drop significantly relative to prior periods. Furthermore, there are firms where managerial compensation is still built around profit targets, and meeting these targets can lead to lucrative bonuses.
Techniques for Managing Earnings How do firms manage earnings? One aspect of good earnings management is the care and nurturing of analyst expectations, a practice that Microsoft perfected during the 1990s. Executives at the firm monitored analyst estimates of earnings, and stepped in to lower expectations when they believed that the estimates were too high.7 There are several other techniques that are used, and some of the most common are considered in this section. Not all the techniques are hurtful to the firm, and some may indeed be considered prudent management.
Planning ahead. Firms can plan investments and asset sales to keep earnings rising smoothly. Revenue recognition. Firms have some leeway when it comes to when revenues have to be recognized. As an example, Microsoft, in 1995, adopted an extremely conservative approach to accounting for revenues from its sale of Windows 95, and chose not to show large chunks of revenues that it was entitled (though not obligated) to show.8 In fact, the firm had accumulated $1.1 billion in unearned revenues by the end of 1996 that it could borrow on to supplement earnings in a weaker quarter. Booking revenues early. In an opposite phenomenon, firms sometimes ship products during the final days of a weak quarter to distributors and retailers and record the revenues. Consider the case of MicroStrategy, a technology firm that went public in 1998. In the last two quarters of 1999, the firm reported revenue growth of 20 percent and 27 percent respectively, but much of that growth was attributable to large deals announced just days after each quarter ended, with some revenues attributed to the just-ended quarter.9 In a more elaborate variant of this strategy, two technology firms, both of which need to boost revenues, can enter into a transaction swapping revenues. Capitalizing operating expenses. Just as with revenue recognition, firms are given some discretion in whether they classify expenses as operating or capital expenses, especially for items like software R&D. AOL's practice of capitalizing and writing off the cost of the CDs and disks it provided with magazines, for instance, allowed it to report positive earnings through much of the late 1990s. 177
Write-offs. A major restructuring charge can result in lower income in the current period, but it provides two benefits to the firm taking it. Since operating earnings are reported both before and after the restructuring charge, it allows the firm to separate the expense from operations. It also makes beating earnings easier in future quarters. To see how restructuring can boost earnings, consider the case of IBM. By writing off old plants in the year they are closed, IBM was able to drop depreciation expenses to 5 percent of revenue in 1996 from an average of 7 percent in 1990–1994. The difference, in 1996 revenue, was $1.64 billion, or 18 percent of the company's $9.02 billion in pretax profit that year. Technology firms have been particularly adept at writing off a large portion of acquisition costs as “in-process R&D” to register increases in earnings in subsequent quarters. Deng and Lev (1998) studied 389 firms that wrote off in-process R&D between 1990 and 1996;10 these write-offs amounted, on average, to 72 percent of the purchase price on these acquisitions, and increased the acquiring firm's earnings 22 percent in the fourth quarter after the acquisition. Use of reserves. Firms are allowed to build up reserves for bad debts, product returns, and other potential losses. Some firms are conservative in their estimates in good years, and use the excess reserves that they have built up during these years to smooth out earnings in other years. Income from investments. Firms with substantial holdings of marketable securities or investments in other firms often have these investments recorded on their books at values well below their market values. Thus, liquidating these investments can result in large capital gains, which can boost income in the period.
Adjustments to Income To the extent that firms manage earnings, you have to be cautious about using the current year's earnings as a base for projections. This section consider a series of adjustments that we might need to make to stated earnings before using the number as a basis for projections. We begin by considering the often subtle differences between one-time, recurring, and unusual items. We follow up by examining how best to deal with the debris left over by acquisition accounting. Then we consider how to deal with income from holdings in other companies and investments in marketable securities. Finally, we look at a series of tests that may help us gauge whether the reported earnings of a firm are reliable indicators of its true earnings.
Extraordinary, Recurring, and Unusual Items The rule for estimating both operating and net income is simple. The operating income that is used as a base for projections should reflect continuing operations and should not include any items that are one-time or extraordinary. Putting this statement into practice is often a challenge because there are four types of extraordinary items: 1. One-time expense or income that is truly one-time. A large restructuring charge that has occurred only once in the past 10 years would be a good example. These expenses can be backed out of the analysis and the operating and net income calculated without them. 2. Expenses and income that do not occur every year but seem to recur at regular intervals. Consider, for instance, a firm that has taken a restructuring charge every 3 years for the past 12 years. While not conclusive, this would suggest that the extraordinary expenses are really ordinary expenses that are being bundled by the firm and taken once every three years. Ignoring such an expense would be dangerous because the expected operating income in future years would be overstated. What would make sense would be to take the expense and spread it out on an annual basis. Thus, if the restructuring expense every three years has amounted to $1.5 billion, on average, the operating income for the current year should be reduced by $0.5 billion to reflect the annual charge due to this expense. 3. Expenses and income that recur every year but with considerable volatility. The best way to deal with such items is to normalize them by averaging the expenses across time and reducing this year's income by this amount. 4. Items that recur every year that change signs—positive in some years and negative in others. Consider, for instance, the effect of foreign currency translations on income. For a firm in the United States, the effect may be negative in years in which the dollar gets stronger and positive in years in which the dollars gets weaker. The most prudent thing to do with these expenses would be to ignore them for cash flow purposes; you may or may not adjust discount rates for the risk created by the variability. To differentiate between these items requires that you have access to a firm's financial history. For young firms, this may not be available, making it more difficult to draw the line between expenses that should be ignored, expenses that should be normalized, and expenses that should be considered in full.
Adjusting for Acquisitions and Divestitures Acquisition accounting can wreak havoc on reported earnings for years after an acquisition. The most common byproduct of acquisitions is the amortization of goodwill. This amortization can reduce reported income in subsequent periods. Should we consider amortization to be an operating expense? We think not, since it is both a noncash and often a non-tax-deductible charge. The safest route to follow with goodwill amortization is to look at earnings prior 178
to the amortization. Technology companies have used an unusual ploy to get the goodwill created when a premium is paid over book value off their books. Using the argument that the bulk of the market value paid for technology companies comes from the value of the research done by the firm over time, they have written off what they called in-process R&D to preserve consistency. After all, they argue, the R&D they do internally is expensed. As with amortization of goodwill, writing off in-process R&D creates a noncash and non-tax-deductible charge and we should look at earnings prior to their write-off. When firms divest assets, they can generate income in the form of capital gains. Infrequent divestitures can be treated as one-time items and ignored, but some firms divest assets on a regular basis. For such firms, it is best to ignore the income associated with the divestiture, but to consider the cash flows associated with divestiture, net of capital gains taxes, when estimating net capital expenditures. For instance, a firm with $500 million in capital expenditures, $300 million in depreciation, and $120 million in divestitures every year would have a net capital expenditure of $80 million.
Income from Investments and Cross Holdings Investments in marketable securities generate two types of income. The first takes the form of interest or dividends and the second is the capital gains (or losses) associated with selling securities at prices that are different from their cost bases. In the 1990s, when the stock market was booming, several technology firms used the latter to augment income and beat analyst estimates. In our view, neither type of income should be considered part of the earnings used in valuation for any firm other than a financial service firm that defines its business as the buying and selling of securities (such as a hedge fund). The interest earned on marketable securities should be ignored when valuing the firm, since it is far easier to add the market value of these securities at the end of the process rather than mingle them with other assets. For instance, assume that you have a firm that generates $100 million in after-tax cash flows, but also assume that 20 percent of these cash flows come from holdings of marketable securities with a current market value of $500 million. The remaining 80 percent of the cash flows comes from operating assets; these cash flows are expected to grow at 5 percent a year in perpetuity, and the cost of capital (based on the risk of these assets) is 10 percent. The value of this firm can be most easily estimated as follows: Value of operating assets of the firm = $80(1.05)/(.10 – .05) $1,680 million Value of marketable securities
$ 500 million
Value of firm
$2,180 million
If we had chosen to discount the entire after-tax cash flow of $100 million, we would have had to adjust the cost of capital (to reflect the risk of the marketable securities). The adjustment, done right, should yield the same value as that estimated.11 The capital gain or loss from the sale of marketable securities should be ignored for a different reason. If you incorporate this gain into your income and use it in your forecasts, not only are you counting on being able to sell your securities for higher prices each period in the future but you risk double counting the value of these securities if you are adding them to the value of the operating assets to arrive at an estimate of value. Firms that have a substantial number of cross holdings in other firms will often report increases or decreases to earnings reflecting income or losses from these holdings. The effect on earnings will vary depending on how the holding is categorized. Chapter 3 differentiated among three classifications: 1. A minority passive holding, where only the dividends received from the holding are recorded in income. 2. A minority active interest, where the portion of the net income (or loss) from the subsidiary is shown in the income statement as an adjustment to net income (but not to operating income). 3. A majority active interest, where the income statements are consolidated and the entire operating income of the subsidiary (or holding) are shown as part of the operating income of the firm. In such cases, the net income is usually adjusted for the portion of the subsidiary owned by others (minority interests). The safest route to take with the first two types of holdings is to ignore the income shown from the holding when valuing a firm, to value the holding separately and to add it to the value obtained for the other assets. As a simple example, consider a firm (Holding Inc.) that generates $100 million in after-tax cash flows from its operating assets, and assume that these cash flows will grow at 5 percent a year forever. In addition, assume that the firm owns 10 percent of another firm (Subsidiary Inc.) with after-tax cash flows of $50 million growing at 4 percent a year forever. Finally, assume that the cost of capital for both firms is 10 percent. The firm value for Holding Inc. can be estimated as follows: Value of operating assets of Holding Inc. = 100(1.05)/(.10 – .05)
$2,100 million
Value of operating assets of Subsidiary Inc. = 50(1.04)/(.10 – .04) $ 867 million
179
Value of Holding Inc. = $2,100 + .10(867)
$2,187 million
When earnings are consolidated, you can value the combined firm with the consolidated income statement and then subtract out the value of the minority holdings. To do this, though, you have to assume that the two firms are in the same business and are of equivalent risk since the same cost of capital will be applied to both firm's cash flows. Alternatively, you can strip the entire operating income of the subsidiary from the consolidated operating income and follow the process just laid out to value the holding. We will return to examine this issue is more detail in chapter 16.
ILLUSTRATION 9.6: Adjusting Earnings for One-Time Charges Between 1997 and 1999, Xerox's reported earnings included a significant number of one-time, extraordinary, and unusual items. The summary of the earnings is provided in the following table:
There are a few obvious adjustments to income that represent one-time charges and a host of other issues. Let us consider first the obvious adjustments: The inventory charge and restructuring charges seem to represent one-time charges, though there is the possibility that they represent more serious underlying problems that can create charges in future periods. The charge for discontinued operations also affects only one year's income. These expenses should be added back to arrive at adjusted operating income and net income. The other (net) expenses line item is a recurring but volatile item. We would average this expense when forecasting future income. To arrive at adjusted net income we would also reverse the last two adjustments by subtracting out the equity in net income of subsidiaries (reflecting Xerox's minority holdings in other firms) and adding back the earnings in minority interests (reflecting minority interests in Xerox's majority holdings). The following table adjusts the net income in each of the years for the changes suggested:
The restructuring and inventory charges were tax deductible and the after-tax portion was added back; the tax rate was computed based on taxes paid and taxable income for that year.
We also add back the after-tax portion of the other expenses (net) and subtract out the average annual expense over the three years:
Similar adjustments would need to be made to operating income. Xerox nets out interest expenses against interest income on its Capital subsidiary to report finance income. You would need to separate interest expenses from interest income to arrive at an estimate of operating income for the firm. What are the other issues? The plethora of one-time charges suggests that there may be ongoing operational problems at Xerox that may cause future charges. In fact, it is not surprising that Xerox had to delay its 10-K filing for 2000 because of accounting issues.
CONCLUSION 180
Financial statements remain the primary source of information for most investors and analysts. There are differences, however, in how accounting and financial analysts approach answering a number of key questions about the firm. This chapter begins our analysis of earnings by looking at the accounting categorization of expenses into operating, financing, and capital expenses. While operating and financing expenses are shown in income statements, capital expenditures are spread over several time periods and take the form of depreciation and amortization. Accounting standards misclassify operating leases and research and development expenses as operating expenses (when the former should be categorized as financing expenses and the latter as capital expenses). We s uggest ways in which earnings can be corrected to better measure the impact of these items. In the second part of the chapter, we consider the effect of one-time, nonrecurring, and unusual items on earnings. While the underlying principle is that earnings should include only normal expenses, this is put to the test by the attempts on the part of companies to move normal operating expenses into the nonrecurring column and nonoperating income into operating earnings.
WARNING SIGNS IN EARNINGS REPORTS The most troubling thing about earnings reports is that we are often blindsided not by the items that get reported (such as extraordinary charges) but by the items that are hidden in other categories. We would suggest the following checklist that should be reviewed about any earnings report to gauge the possibility of such shocks: Is earnings growth outstripping revenue growth by a large magnitude year after year? This may well be a sign of increased efficiency, but when the differences are large and continue year after year, you should wonder about the source of these efficiencies. Do one-time or nonoperating charges to earnings occur frequently? The charge itself might be categorized differently each year—an inventory charge one year, a restructuring charge the next, and so on. While this may be just bad luck, it may also reflect a conscious effort by a company to move regular operating expenses into these nonoperating items. Do any of the operating expenses, as a percent of revenues, swing wildly from year to year? This may suggest that this expense item (say SG&A) includes nonoperating expenses that should really be stripped out and reported separately. Does the company manage to beat analyst estimates quarter after quarter by a cent or two? Not every company is a Microsoft. Companies that beat estimates year after year probably are involved in earnings management and are moving earnings across time periods. As growth levels off, this practice can catch up with them. Does a substantial proportion of the revenues come from subsidiaries or related holdings? While the sales may be legitimate, the prices set may allow the firm to move earnings from one unit to the other and give a misleading view of true earnings at the firm. Are accounting rules for valuing inventory or depreciation changed frequently? Are acquisitions followed by miraculous increases in earnings? It is difficult to succeed with an acquisition strategy in the long term. A firm that claims instant success from such as strategy requires scrutiny. Is working capital ballooning out as revenues and earning surge? This can sometimes let us pinpoint those firms that generate revenues by lending to their own customers. None of these factors, by themselves, suggest that we distrust earnings for these firms, but combinations of the factors can be viewed as a warning signal that the earnings statement needs to be held up to higher scrutiny.
QUESTIONS AND SHORT PROBLEMS In the problems following, use an equity risk premium of 5.5 percent if none is specified. 1. Derra Foods is a specialty food retailer. In its balance sheet, the firm reports $1 billion in book value of equity and no debt, but it has operating leases on all its stores. In the most recent year, the firm made $85 million in operating lease payments, and its commitments to make lease payments for the next five years and beyond are: Year Operating Lease Expense 1
$90 million
2
$90 million
3
$85 million
4
$80 million
5
$80 million
6–10 $75 million annually
If the firm's current cost of borrowing is 7%, estimate the debt value of operating leases. Estimate the book value debt-to-equity ratio. 2. Assume that Derra Foods, in the preceding problem, reported earnings before interest and taxes (with operating leases expensed) of $200 million. Estimate the adjusted operating income, assuming that operating leases are capitalized. 3. FoodMarkets Inc. is a grocery chain. It reported a book debt-to-capital ratio of 10% and a return on capital of 25% on a book value of capital invested of $1 billion. Assume that the firm has significant operating leases. If the operating lease expense in the current year is $100 million and the present value of lease commitments is $750 million, reestimate FoodMarkets' debt to capital and return on capital. (You can assume a pretax cost of debt of 181
8%.) 4. Zif Software is a firm with significant research and development expenses. In the most recent year, the firm had $100 million in R&D expenses. R&D expenses are amortizable over five years, and over the past five years they are: Year
R&D Expenses
–5
$ 50 million
–4
$ 60 million
–3
$ 70 million
–2
$ 80 million
–1
$ 90 million
Current year $100 million
Assuming a linear amortization schedule (over five years), estimate: a. The value of the research asset. b. The amount of R&D amortization this year. c. The adjustment to operating income. 5. Stellar Computers has a well-earned reputation for earning a high return on capital. The firm had a return on capital of 100% on capital invested of $1.5 billion last year. Assume that you have estimated the value of the research asset to be $1 billion. In addition, the R&D expense this year is $250 million, and the amortization of the research asset is $150 million. Reestimate Stellar Computers' return on capital. If only amortization were tax deductible, the tax benefit from R&D expenses would be:
1
This extra tax benefit we get from the entire R&D being tax deductible is as follows: If we subtract out (R&D – Amortization)(1 – Tax rate) and then add the differential tax benefit that is computed above, (1 – Tax rate) drops out of the equation. If the return on capital earned by a firm is well below the cost of capital, the adju stment could result in a higher return. 2
As an example, Jamie Kiggen, an equity research analyst at Donaldson, Lufkin & Jenrette, valued an Amazon customer at $2,400 in an equity research report in 1999. This value was based on the assumption that the customer would continue to buy from Amazon.com and on an expected profit margin from such sales. 3
The value is rounded to the nearest integer.
4
The lease life is computed by adding the estimated annuity life of two years for the lump sum to the initial five years. 5
I/B/E /S estimates.
6
Microsoft preserved its credibility with analysts by also letting them know when their estimates were too low. Firms that are consi stently pessimistic in their analyst presentations lose their credibility and consequently their effectiveness in managing earnings. 7
Firms that bought Windows 95 in 1995 also bought the right to upgrades and support in 1996 and 1997. Microsoft could have shown these as revenues in 1995. 8
9 Forbes magazine carried an article on M arch 6, 2000, on MicroStrategy, with this excerpt: “On Oct. 4 MicroStrategy and NCR announced what they described as a $52.5 million licensing and technology agreement. NCR agreed to pay MicroStrategy $27.5 million to license its software. MicroStrategy bought an NCR unit which had been a competitor for what was then $14 million in stock, and agreed to pay $11 million in cash for a data warehousing system. MicroStrategy reported $17.5 million of the licensing money as revenue in the third quarter, which had closed four days earlier.” 10
Only three firms wrote off in-process R&D during the prior decade (1980 to 1989). 182
This will happen only if the marketable securities are fairly priced and you are earning a fair market return on them. If they are not, you can get different values from the approaches. 11
183
CHAPTER 10 From Earnings to Cash Flows The value of an asset comes from its capacity to generate cash flows. When valuing a firm, the se cash flows should be after taxes, prior to debt payments, and after reinvestment needs. When valuing equity, the cash flows should be after debt payments. There are thus three basic steps to estimating these cash flows. The first is to estimate the earnings generated by a firm on its existing assets and investments, a process we examined in the preceding chapter. The second step is to estimate the portion of this income that would go toward paying taxes. The third is to develop a measure of how much a firm is reinvesting back for future growth. This chapter examines the last two steps. It will begin by investigating the difference between effective and marginal taxes, as well as the effects of substantial net operating losses (NOLs) carried forward. To examine how much a firm is reinvesting, we will break it down into reinvestment in tangible and long-lived assets (net capital expenditures) and short-term assets (working capital). We will use a much broader definition of reinvestment to include investments in research and development (R&D) and acquisitions as part of capital expenditures.
THE TAX EFFECT To compute the after-tax operating income, you multiply the earnings before interest and taxes by an estimated tax rate. This simple procedure can be complicated by three issues that often arise in valuation. The first is the wide differences you observe between effective and marginal tax rates for these firms, and the choice you face between the two in valuation. The second issue arises usually with firms with large losses, leading to net operating losses that are carried forward and can save taxes in future years. The third issue arises from the capitalizing of research and development and other expenses. The fact that these expenditures can be expensed immediately leads to much higher tax benefits for the firm.
Effective versus Marginal Tax Rate You are faced with a choice of several different tax rates. The most widely reported tax rate in financial statements is the effective tax rate, which is computed from the reported income statement as follows:
The second choice on tax rates is the marginal tax rate, which is the tax rate the firm faces on its last dollar of income. This rate depends on the tax code and reflects what firms have to pay as taxes on their marginal income. In the United States, for instance, the federal corporate tax rate on marginal income is 35 percent; with the addition of state and local taxes, most profitable firms face a marginal corporate tax rate of 40 percent or higher. While the marginal tax rates for most firms in the United States should be fairly similar, there are wide differences in effective tax rates across firms. Figure 10.1 provides a distribution of effective tax rates for moneymaking firms in the United States in January 2011. Note that a number of firms report effective tax rates of less than 10 percent as well as that a few firms have effective tax rates that exceed 50 percent. In addition, it is worth noting that this figure does not include about 2,000 firms that did not pay taxes during the most recent financial year (and most of these are money-losing companies) or that have a negative effective tax rate.1 Figure 10.1 Effective Tax Rates for U.S. Companies—January 2011 Source: Value Line.
184
Reasons for Differences between Marginal and Effective Tax Rates Given that most of the taxable income of publicly traded firms is at the highest marginal tax bracket, why would a firm's effective tax rate be different from its marginal tax rate? There are at least four reasons: 1. Many firms, at least in the Unite d States, follow different accounting standards for tax and for reporting purposes. For instance, firms often use straight-line depreciation for reporting purposes and accelerated depreciation for tax purposes. As a consequence, the reported income is significantly higher than the taxable income, on which taxes are based.2 2. Firms sometimes use tax credits to reduce the taxes they pay. These credits, in turn, can reduce the effective tax rate below the marginal tax rate. 3. Firms can sometimes defer taxes on income to future periods. If firms defer taxes, the taxes paid in the current period will be at a rate lower than the marginal tax rate. In a later period, however, when the firm pays the deferred taxes, the effective tax rate will be higher than the marginal tax rate. 4. Firms that generate substantial income for foreign domiciles with lower tax rates do not have to pay domest ic taxes until that income is repatriated back to the domestic country.
Marginal Tax Rates for Multinationals When a firm has global operations, its income is taxed at different rates in different locales. When this occurs, what is the marginal tax rate for the firm? There are three ways in which we can deal with different tax rates. 1. The first is to use a weighted average of the marginal tax rates, with the weights based on the income derived by the firm from each of these countries. The problem with this approach is that the weights will change over time, if income is growing at different rates in different countries. 2. The second is to use the marginal tax rate of the country in which the company is incorporated, with the implicit assumption being that the income generated in other countries will eventually have to be repatriated to the country of origin, at which point the firm will have to pay the marginal tax rate. 3. The third approach is to keep the income from each country separate and apply a different marginal tax rate to each income stream.
Effects of Tax Rate on Value In valuing a firm, should you use the marginal or the effective tax rates. If the same tax rate has to be applied to earnings every period, the safer choice is the marginal tax rate, because none of the four reasons noted can be sustained in perpetuity. As new capital expenditures taper off, the difference between reported and tax income will narrow; tax credits are seldom perpetual and firms eventually do have to pay their deferred taxes. There is no reason, however, why the tax rates used to compute the after-tax cash flows cannot change over time. Thus, in valuing a firm with an effective tax rate of 24 percent in the current period and a marginal tax rate of 35 percent, you can estimate the first year's cash flows using the marginal tax rate of 24 percent and then increase the tax rate to 35 percent over time. It is good practice to assume that the tax rate used in perpetuity to compute the terminal value be the marginal tax rate. 185
When valuing equity, we often start with net income or earnings per share, which are after-tax earnings. Although it looks as though we can avoid dealing with the estimating of tax rates when using after-tax earnings, appearances are deceptive. The current after-tax earnings of a firm reflect the taxes paid this year. To the extent that tax planning or deferral caused this payment to be very low (low effective tax rates) or very high (high effective tax rates), we run the risk of assuming that the firm can continue to do this in the future if we do not adjust the net income for changes in the tax rates in future years.
ILLUSTRATION 10.1: Effect of Tax Rate Assumptions on Value Convoy Inc. is a telecommunications firm that generated $150 million in pretax operating income and reinvested $30 million in the most recent financial year. As a result of tax deferrals, the firm has an effective tax rate of 20%, while its marginal tax rate is 40%. Both the operating income and the reinvestment are expected to grow 10% a year for five years, and 5% thereafter. The firm's cost of capital is 9% and is expected to remain unchanged over time. We will estimate the value of Convoy using three different assumptions about tax rates—the effective tax rate forever, the marginal tax rate forever, and an approach that combines the two rates. Approach 1: Effective Tax Rate Forever We first estimate the value of Convoy assuming that the tax rate remains at 20% forever:
This value is based on the implicit assumption that deferred taxes will never have to be paid by the firm. Approach 2: Marginal Tax Rate Forever We next estimate the value of Convoy assuming that the tax rate is the marginal tax rate of 40% forever:
This firm value is based on the implicit assumption that the firm cannot defer taxes from this point on. In fact, an even more conservative reading would suggest that we should reduce this value by the amount of the cumulated deferred taxes from the past. Thus, if the firm has $200 million in deferred taxes from prior years, and expects to pay these taxes over the next four years in equal annual installments of $50 million, we would first compute the present value of these tax payments: Present value of deferred tax payments = $50 million(PV of annuity, 9%, 4 years) = $161.99 million The value of the firm would then be $1,794.95 million.
Approach 3: Blended Tax Rates In the final approach, we will assume that the effective tax will remain 20% for five years and we will use the marginal tax rate to compute the terminal value:
Note, however, that the use of the effective tax rate for the first five years will increase the deferred tax liability to the firm. Assuming that the firm ended the current year with a cumulated deferred tax liability of $200 million, we can compute the deferred tax liability by the end of the fifth year:
186
We will assume that the firm will pay this deferred tax liability after year 5, but spread the payments over 10 years, leading to a present value of $167.45 million.
Note that the payments do not start until the sixth year, and hence get discounted back an additional five years. The value of the firm can then be estimated:
taxrate.xls: This dataset on the Web summarizes average effective tax rates by industry group in the United States for the most recent quarter.
Effect of Net Operating Losses For firms with large net operating losses carried forward or continuing operating losses, there is the potential for significant tax savings in the first few years that they generate positive earnings. There are two ways of capturing this effect. One is to change tax rates over time. In the early years, these firms will have a zero tax rate, as losses carried forward offset income. As the net operating losses decrease, the tax rates will climb toward the marginal tax rate. As the tax rates used to estimate the after-tax operating income change, the rates used to compute the after-tax cost of debt in the cost of capital computation also need to change. Thus, for a firm with net operating losses carried forward, the tax rate used for both the computation of after-tax operating income and cost of capital will be zero during the years when the losses shelter income. The other approach is often used when valuing firms that already have positive earnings but have a large net operating loss carried forward. Analysts will value the firm ignoring the tax savings generated by net operating losses, and then add to this amount the expected tax savings from net operating losses. Often, the expected tax savings are estimated by multiplying the tax rate by the net operating loss. The limitation of doing this is that it assumes that the tax savings are both guaranteed and instantaneous. To the extent that firms have to generate earnings to create these tax savings and there is uncertainty about earnings, it will overestimate the value of the tax savings. There are two final points that need to be made about operating losses. To the extent that a potential acquirer can claim the tax savings from net operating losses sooner than the firm generating these losses, there can be potential for tax synergy that we examine in chapter 26, when we look at acquisitions. The other is that there are countries where there are significant limitations on how far forward operating losses can be taken. If this is the case, the value of these net operating losses may be reduced.
ILLUSTRATION 10.2: The Effect of Net Operating Loss on Value: Tesla Motors This illustration considers the effect of both net operating losses (NOLs) carried forward and expected losses in future periods on the tax rate for Tesla Motors, the electric car company, in 2011. Tesla reported an operating loss of $65.5 million in 2010, on revenues of $116.74 million, and had an accumulated net operating loss of $140.64 million by the end of that year. While things do look bleak for the firm, we will assume that revenues will grow significantly over the next decade and that the firm's operating margin will converge on the industry average of 10% for mature and healthy automobile firms. The following table summarizes our projections of revenues and operating income for Tesla for the next 10 years:
Note that Tesla continues to lose money over the next five years, and adds to its net operating losses. In years 6, 7, and 8, its operating income is positive but it still pays no taxes because of its accumulated net operating losses from prior years. In year 9, it is able to reduce its taxable
187
income by the remaining net operating loss ($79 million), but it begins paying taxes for the first time. We will assume a 40% tax rate and use this as our marginal tax rate after year 9. The benefits of the net operating losses are thus built into the cash flows and the value of the firm.
The Tax Benefits of R&D Expensing The preceding chapter argued that R&D expenses should be capitalized. If we decide to do so, however, there is a tax benefit that we might be missing. Firms are allowed to deduct their entire R&D expense for tax purposes. In contrast, they are allowed to deduct only the depreciation on their capital expenses. To capture the tax benefit, therefore, you would add the tax savings on the difference between the entire R&D expense and the amortized amount of the research asset to the after-tax operating income of the firm:
A similar adjustment would need to be made for any other operating expense that you choose to capitalize. In Chapter 9, we noted that the adjustment to pretax operating income from capitalizing R&D is:
To estimate the after-tax operating income, we would multiply this value by (1 – Tax rate) and add on the additional tax benefit from before:
In other words, the tax benefit from R&D expensing allows us to add the difference between R&D expense and amortization directly to the after-tax operating income (and to net income).
ILLUSTRATION 10.3: Tax Benefit from Expensing: Amgen in 2011 In Chapter 9, we capitalize, R&D expenses for Amgen and estimated the value of the research asset to Amgen and adjusted operating income. Reviewing Illustration 9.2, we see the following adjustments:
To estimate the tax benefit from expensing for Amgen, first assume that the tax rate for Amgen is 35% and note that Amgen can deduct the entire $3,030 million for tax purposes:
If only the amortization had been eligible for a tax deduction in 2010, the tax benefit would have been:
By expensing instead of capitalizing, Amgen was able to derive a much larger tax benefit. The differential tax benefit can be written as:
Thus, Amgen derives a tax benefit of $468 million because it can expense R&D expenses rather than capitalize them. Completing the analysis, we computed the adjusted after-tax operating income for Amgen. Note that in Illustration 9.2, we estimated the adjusted pretax operating income to be the following:
You could convert this pretax operating income into an after-tax value and add back the tax benefit from R&D:
You can also arrive at the same answer by computing the unadjusted after-tax operating income and adjusting it for R&D:
188
Tax Books and Reporting Books It is no secret that many firms in the United States maintain two sets of books—one for tax purposes and one for reporting purposes—and that this practice not only is legal but is also widely accepted. While the details vary from company to company, the income reported to stockholders generally is much higher than the income reported for tax purposes. When valuing firms, we generally have access to only the former and not the latter, and this can affect our estimates in a number of ways: Dividing the taxes payable, which is computed on the taxable income, by the reported income, which is generally much higher, will yield a tax rate that is lower than the true tax rate. If we use this tax rate as the forecasted tax rate, we could over value the company. This is another reason for shifting to marginal tax rates in future periods. If we base the projections on the reported income, we will overstate expected future income. The effect on cash flows is likely to be muted. To see why, consider one very common difference between reporting and tax income: Straight-line depreciation is used to compute the former and accelerated depreciation is used for the latter. Since we add depreciation back to after-tax income to get to cash flows, the drop in depreciation will offset the increase in earnings. The problem, however, is that we understate the tax benefits from depreciation. Some companies capitalize expenses for reporting purposes (and for depreciating them in subsequent periods) but expense them for tax purposes. Here again, using the income and the capital expenditures from reporting books will result in an understatement of the tax benefits from the expensing. Thus the problems created by firms having different standards for tax and accounting purposes are much greater if we focus on reported earnings (as is the case when we use earnings multiples like PE or EBITDA multiples) than when we use cash flows. If we did have a choice, however, we would base our valuations on the tax books rather than the reporting books.
DEALING WITH TAX SUBSIDIES Firms sometimes obtain tax subsidies from the government for investing in specified areas or types of businesses. These tax subsidies can take the form of either reduced tax rates or tax credits. Either way, these subsidies should increase the value of the firm. The question, of course, is how best to build in the effects into the cash flows. Perhaps the simplest approach is to first value the firm, ignoring the tax subsidies, and to then add on the value increment from the subsidies. For instance, assume that you are valuing a pharmaceutical firm with operations in Puerto Rico, which entitle the firm to a tax break in the form of a lower tax rate on the income generated from these operations. You could value the firm using its normal marginal tax rate, and then add to that value the present value of the tax savings that will be generated by the Puerto Rican operations. There are three advantages with this approach: 1. It allows you to isolate the tax subsidy and consider it only for the period over which you are entitled to it. When the effects of these tax breaks are consolidated with other cash flows, there is a danger that they can be viewed as perpetuities. 2. The discount rate used to compute the tax breaks can be different from the discount rate used on the other cash flows of the firm. Thus, if the tax break is a guaranteed tax credit by the government, you could use a much lower discount rate to compute the present value of the cash flows. 3. Building on the theme that there are few free lunches, it can be argued that governments provide tax breaks for investments only because firms are exposed to higher costs or more risk in these investments. By isolating the value of the tax breaks, firms can then consider whether the trade-off operates in their favor. For example, assume that you are a sugar manufacturer that is offered a tax credit by the government for being in the business. In return, the government imposes sugar price controls. The firm can compare the value created by the tax credit with the value lost because of the price controls and decide whether it should fight to preserve its tax credit.
REINVESTMENT NEEDS The cash flow to the firm is computed after reinvestments. Two components go into estimating reinvestment. The first is net capital expenditures, which is the difference between capital expenditures and depreciation. The other is investments in non-cash working capital.
Net Capital Expenditures In estimating net capital expenditures, we generally deduct depreciation from capital expenditures. The rationale is that the positive cash flows from depreciation pay for at least a portion of capital expenditures, and that it is only the excess that represents a drain on the firm's cash flows. Whereas information on capital spending and depreciation are usually easily accessible in most financia l statements, forecasting these expenditures can be difficult for three reasons. The first is that firms often incur capital spending in chunks—a large investment in one year can be followed by small investments in subsequent years. The second is that the accounting definition of capital spending does not incorporate those capital expenses that are treated as operating expenses such as R&D expenses. The third is that acquisitions are not classified by accountants as capital expenditures. For firms that grow primarily through acquisition, this will result in an understatement of the net capital expenditures.
189
Lumpy Capital Expenditures and the Need for Smoothing Firms seldom have smooth capital expenditure streams. Firms can go through periods when capital expenditures are very high (as is the case when a new product is introduced or a new plant built), followed by periods of relatively light capital expenditures. Consequently, when estimating the capital expenditures to use for forecasting future cash flows, you should normalize capital expenditures. There are at least two ways in which you can accomplish this objective. The simplest normalization technique is to average capital expenditures over a number of years. For instance, you could estimate the average capital expenditures over the past four or five years for a manufacturing firm and use that number rather the capital expenditures from the most recent year. By doing so, you could capture the fact that the firm may invest in a new plant every four years. If instead you had used the capital expenditures from the most recent year, you would have either overestimated capital expenditures (if the firm built a new plant that year) or underestimated them (if the plant had been built in an earlier year). There are two measurement issues that you will need to confront. One relates to the number of years of history that you should use. The answer will vary across firms and will depend on how infrequently the firm makes large investments. The other is on the question of whether averaging capital expenditures over time requires us to average depreciation as well. Since depreciation is already spread out over time, the need for normalization should be much smaller. In addition, the tax benefits received by the firm reflect the actual depreciation in the most recent year, rather than an average depreciation over time. Unless depreciation is as volatile as capital expenditures, it makes more sense to leave depreciation untouched. For firms with a limited history or firms that have changed their business mix over time, averaging over time either is not an option or will yield numbers that are not indicative of its true capital expenditure needs. For these firms, industry averages for capital expenditures are an alternative. Since the sizes of firms can vary across an industry, the averages are usually computed with capital expenditures as a percent of a base input—revenues and total assets are common choices. We prefer to look at capital expenditures as a percent of depreciation, and to average this statistic for the industry. In fact, if there are enough firms in the sample, you could look at the average for a subset of firms that are at the same stage of the life cycle as the firm being analyzed.
ILLUSTRATION 10.4: Estimating Normalized Net Capital Expenditures: Reliance Industries Reliance Industries is one of India's largest firms and is involved in a multitude of businesses ranging from chemicals to textiles. The firm makes substantial investments in these businesses, and the following table summarizes the capital expenditures and depreciation for the period 1997 to 2000:
The firm's capital expenditures have been volatile, but its depreciation has been trending upward. There are two ways in which we can normalize the net capital expenditures. One is to take the average net capital expenditure over the four-year period, which would result in net capital expenditures of INR 13,639 million. The problem with doing this, however, is that the depreciation implicitly being used in the calculation is INR 8,027 million, which is well below the actual depreciation of INR 12,784. A better way to normalize capital expenditures is to use the average capital expenditure over the four-year period (INR 21,166) and depreciation from the year 2000 (INR 12,784) to arrive at a normalized net capital expenditure value:
Note that the normalization did not make much difference in this case because the actual net capital expenditures in 2000 amounted to INR 8,334 million.
Capital Expenses Treated as Operating Expenses In Chapter 9, we discussed the capitalization of expenses such as R&D and personnel training, where the benefits last over multiple periods, and examined the effects on earnings. There should also clearly be an impact on our estimates of capital expenditures, depreciation, and, consequently, net capital expenditures. If we decide to recategorize some operating expenses as capital expenses, we should treat the current period's value for this item as a capital expenditure. For instance, if we decide to capitalize R&D expenses, the amount spent on R&D in the current period has to be added to capital expenditures.
Since capitalizing an operating expense creates an asset, the amortization of this asset should be added to 190
depreciation for the current period. Thus, capitalizing R&D creates a research asset, which generates an amortization in the current period.
If we are adding the current period's expense to the capital expenditures and the amortization of the asset to the depreciation, the net capital expenditures of the firm will increase by the difference between the two:
Note that the adjustment that we make to net capital expenditure mirrors the adjustment we make to operating income. Since net capital expenditures are subtracted from after-tax operating income, we are, in a sense, nullifying the impact on cash flows of capitalizing R&D.
ILLUSTRATION 10.5: Effect of Capitalizing R&D: Amgen in March 2009 In Illustration 9.2 we capitalized Amgen's R&D expense and created a research asset. In Illustration 10.3 we considered the additional tax benefit generated by the fact that a company can expense the entire amount. In this illustration, we complete the analysis by looking at the impact of capitalization on net capital expenditures. Reviewing the numbers again, Amgen had an R&D expense of $3,030 million in 2010. Capitalizing the R&D expenses, using an amortizable life of 10 years, yields a value for the research asset of $13,283 million and an amortization for the current year (2008) of $1,694 million. In addition, note that Amgen reported capital expenditures of $1,646 million in 2008 and depreciation and amortization amounting to $1,073 million. The adjustments to capital expenditures, depreciation, and amortization and net capital expenditures are:
Viewed in conjunction with the adjustment to after-tax operating income in Illustration 10.2, the change in net capital expenditure (an increase of $1336 million from $573 million to $1,929 million) is exactly equal to the change in after-tax operating income. Capitalizing R&D thus has no effect on the free cash flow to the firm. Though the bottom-line cash flow does not change, the capitalization of R&D significantly changes the estimates of earnings and reinvestment. Thus it helps us better understand how profitable a firm is and how much it is reinvesting for future growth.
Acquisitions In estimating capital expenditures, we should not distinguish between internal investments (which are usually categorized as capital expenditures in cash flow statements) and external investments (which are acquisitions). The capital expenditures of a firm, therefore, need to include acquisitions. Since firms seldom make acquisitions every year, and each acquisition has a different price tag, the point about normalizing capital expenditures applies even more strongly to this item. The capital expenditure projections for a firm that makes an acquisition of $100 million approximately every five years should therefore include about $20 million, adjusted for inflation, every year. Should you distinguish between acquisitions funded with cash versus those funded with stock? We do not believe so. While there may be no cash spent by a firm in the latter case, the firm is increasing the number of shares outstanding. In fact, one way to think about stock-funded acquisitions is that the firm has skipped a step in the funding process. It could have issued the stock to the public, and used the cash to make the acquisitions. Another way of thinking about this issue is that a firm that uses stock to fund acquisitions year after year and is expected to continue to do so in the future will increase the number of shares outstanding. This, in turn, will dilute the value per share to existing stockholders.
ILLUSTRATION 10.6: Estimating Net Capital Expenditures: Cisco Systems in 1999 Cisco Systems increased its market value a hundredfold during the 1990s, largely based on its capacity to grow revenues and earnings at an annual rate of 60% to 70%. Much of this growth was created by acquisitions of small companies with promising technologies and Cisco's ability to convert to them into commercial successes. To estimate net capital expenditures for Cisco in 1999, we begin with the estimates of capital expenditure ($584 million) and depreciation ($486 million) in the 10-K. Based on these numbers, we would have concluded that Cisco's net capital expenditures in 1999 were $98 million. The first adjustment we make to this number is to incorporate the effect of research and development expenses. We use a five-year amortizable life and estimate the value of the research asset and the amortization in 1999 in the following table:
191
The net capital expenditures for Cisco were adjusted by adding back the R&D expenses in the most recent financial year ($1,594 million) and subtracting the amortization of the research asset ($485 million). The second adjustment is to bring in the effect of acquisitions that Cisco made during the last financial year. The following table summarizes the acquisitions made during the year and the prices paid on these acquisitions:
Acquired
Method of Acquisition Price Paid
GeoTel
Pooling
$1,344
Fibex
Pooling
318
Sentient Pooling American Purchase Internet Corporation Summa Four Purchase
103
Clarity Wireless Selsius Systems
Purchase
153
Purchase
134
PipeLinks Purchase Amteva Purchase Technologies
118
Total
$2,516
58
129
159
Dollars in millions. Note that both purchase and pooling transactions are included, and that the sum total of these acquisitions is added to net capital expenditures in 1999. We are assuming, given Cisco's track record, that its acquisitions in 1999 are not unusual and reflect Cisco's reinvestment policy. The amortization associated with these acquisitions is already included as part of depreciation by the firm.3 The following table summarizes the final net capital expenditures for Cisco in 1999. Capital expenditures
$ 584.00
– Depreciation
$ 486.00
Net cap ex (from financials) $ 98.00 + R&D expenditures
$1,594.00
– Amortization of R&D
$ 484.60
+ Acquisitions
$2,516.00
Adjusted net cap ex
$3,723.40
IGNORING ACQUISITIONS IN VALUATION: A POSSIBILITY? Incorporating acquisitions into net capital expenditures and value can be difficult, and especially so for firms that do large acquisitions infrequently. Predicting whether there will be acquisitions, how much they will cost, and what they will deliver in terms of higher growth can be close to impossible. There is one way in which you can ignore acquisitions, but it does come with a cost. If you assume that firms pay a fair price on acquisitions (i.e., a price that reflects the fair value of the target company) and you assume that the target company stockholders claim any or all synergy or control value, acquisitions have no effect on value no matter how large they might be and how much they might seem to deliver in terms of higher growth. The reason is simple: A fair-value acquisition is an investment that earns its required return—a zero net present value investment. If you choose not to consider acquisitions when valuing a firm, you have to remain internally consistent. The portion of growth that is due to acquisitions should not be considered in the valuation. A common mistake that is made in valuing companies that have posted impressive historic growth numbers from an acquisition-based strategy is to extrapolate from this growth and ignore acquisitions at the same time. This will result in an overvaluation of your firm, since you have counted the benefits of the acquisitions but have not paid for them. What is the cost of ignoring acquisitions? Not all acquisitions are fairly priced, and not all synergy and control value ends up with the target company stockholders. Ignoring the costs and benefits of acquisitions will result in an misvaluation of a firm that has established a reputation for growing through acquisitions. We undervalue firms that create value by making good acquisitions and overvalue firms that destroy value by overpaying on acquisitions.
192
capex.xls: This dataset on the Web summarizes capital expenditures, as a percent of revenues and firm value, by industry group in the United States for the most recent quarter.
Investment in Working Capital The second component of reinvestment is the cash that needs to be set aside for working capital needs. Increases in working capital tie up more cash and hence generate negative cash flows. Conversely, decreases in working capital release cash and positive cash flows.
Defining Working Capital Working capital is usually defined to be the difference between current assets and current liabilities. However, we will modify that definition when we measure working capital for valuation purposes. We will back out cash and investments in marketable securities from current assets. This is because cash, especially in large amounts, is invested by firms in Treasury bills, short-term government securities, or commercial paper. Although the return on these investments may be lower than what the firm may make on its real investments, they represent a fair return for riskless investments. Unlike inventory, accounts receivable, and other current assets, cash then earns a fair return and should not be included in measures of working capital. Are there exceptions to this rule? When valuing a firm that has to maintain a large cash balance for day-to-day operations or a firm that operates in a market in a poorly developed banking system, you could consider the cash needed for operations as a part of working capital but only if that cash is wasting cash, i.e., cash that is not earning a fair market return. As we shift from a cash-based economy, this operating cash requirement has become smaller and smaller. We will also back out all interest-bearing debt—short-term debt and the portion of long-term debt that is due in the current period—from the current liabilities. This debt will be considered when computing cost of capital and it would be inappropriate to count it twice. The noncash working capital varies widely across firms in different sectors and often across firms in the same sector. Figure 10.2 shows the distribution of noncash working capital as a percent of revenues for U.S. firms in January 2011. Figure 10.2 Noncash Working Capital as Percent of Revenues—U.S. firms in 2011 Source: Value Line.
ILLUSTRATION 10.7: Working Capital versus Noncash Working Capital: Marks and Spencer Marks and Spencer operates retail stores in the Unite d Kingdom and has substantial holdings in retail firms in other parts of the world. The following table breaks down the components of working capital for the firm for 1999 and 2000 and reports both the total workin g capital and noncash working capital in each year:
1999 2000 Cash and near cash
$ 282
$ 301
Marketable securities
$ 204
$ 386
193
Trade debtors (accounts receivable) $1,980 $2,186 Stocks (inventory)
$ 515
$ 475
Other current assets
$ 271
$ 281
Total current assets
$3,252 $3,629
Noncash current assets
$2,766 $2,942
Trade creditors (accounts payable)
$ 215
$ 219
Short-term debt
$ 913
$1,169
Other short-term liabilities
$ 903
$ 774
Total current liabilities
$2,031 $2,162
Nondebt current liabilities
$1,118 $ 993
Working capital
$1,221 $1,467
Noncash working capital
$1,648 $1,949
The noncash working capital is substantially higher than the working capital in both years. We would suggest that the former is a much better measure of cash tied up in working capital.
Estimating Expected Changes in Noncash Working Capital While we can estimate the noncash working capital change fairly simply for any year using financial statements, this estimate has to be used with caution. Changes in noncash working capital are unstable, with big increases in some years followed by big decreases in subsequent years. To ensure that the projections are not the result of an unusual base year, you should tie the changes in working capital to expected changes in revenues or costs of goods sold at the firm over time. The noncash working capital as a percent of revenues can be used, in conjunction with expected revenue changes each period, to estimate projected changes in noncash working capital over time. You can obtain the noncash working capital as a percent of revenues by looking at the firm's history or at industry standards. Should you break working capital down into more detail? In other words, is there a payoff to estimating individual items, such as accounts receivable, inventory, and accounts payable separately? The answer will depend on both the firm being analyzed and how far into the future working capital is being projected. For firms where inventory and accounts receivable behave in very different ways as revenues grow, it clearly makes sense to break working capital down into detail. The cost, of course, is that it increases the number of inputs needed to value a firm. In addition, the payoff to breaking working capital down into individual items will become smaller as we go further into the future. For most firms, estimating a composite number for noncash working capital is easier to do and often more accurate than breaking it down into more detail.
ILLUSTRATION 10.8: Estimating Noncash Working Capital Needs: The Gap As a specialty retailer, the Gap has substantial inventory and working capital needs. At the end of the 2000 financial year (which concluded in January 2001), the Gap reported $1,904 million in inventory and $335 million in other noncash current assets. At the same time, the accounts payable amounted to $1,067 million and other non-interest-bearing current liabilities were $702 million. The noncash working capital for the Gap in January 2001 can be estimated as follows:
The following table reports on the noncash working capital at the end of the previous year and the total revenues in each year:
The noncash working capital increased by $307 million from the preceding year to this one. When forecasting the noncash working capital needs for the Gap, there are five choices: 1. One is to use the change in noncash working capital from the year ($307 million) and to grow that change at the same rate as earnings are expected to grow in the future. This is probably the least desirable option because changes in noncash working capital from year to year are extremely volatile, and last year's change may in fact be an outlier. 2. The second is to base our changes on noncash working capital as a percent of revenues in the most recent year and expected revenue growth in future years. In the case of the Gap, that would indicate that noncash working capital changes in future years will be 3.44% of revenue changes in that year. This is a much better option than the first one, but the noncash working capital as a percent of revenues can also change from one year to the next. 3. The third is to base our changes on the marginal noncash working capital as a percent of revenues in the most recent year, computed by dividing the change in noncash working capital in the most recent year and the change in revenues in the most recent year, by expected
194
revenue growth in future years. In the case of the Gap, this would lead to noncash working capital changes being 15.06% of revenues in future periods. This approach is best used for firms whose business is changing and where growth is occurring in areas different from the past. For instance, a brick-and-mortar retailer that is growing mostly online may have a very different marginal working capital requirement than the total. 4. The fourth is to base our changes on the noncash working capital as a percent of revenues over a historical period. For instance, noncash working capital as a percent of revenues between 1997 and 2000 averaged out to 4.5% of revenues. The advantage of this approach is that it smooths out year-to-year shifts, but it may not be appropriate if there is a trend (upward or downward) in working capital. 5. The final approach is to ignore the working capital history of the firm and to base the projections on the industry average for noncash working capital as a percent of revenues. This approach is most appropriate when a firm's history reveals a working capital that is volatile and unpredictable. It is also the best way of estimating noncash working capital for very small firms that may see economies of scale as they grow. While these conditions do not apply for the Gap, we can still estimate noncash working capital requirements using the average noncash working capital as a percent of revenues for specialty retailers which is 7.54%. To illustrate how much of a change each of these assumptions can have on working capital requirements, the following table forecasts expected changes in noncash working capital (WC) using each of them. In making these estimates, we have assumed a 10% growth rate in revenues and earnings for the Gap for the next five years.
The noncash working capital investment varies widely across the five approaches that have been described here.
Negative Working Capital (or Changes) Can the change in noncash working capital be negative? The answer is clearly yes. Consider, though, the implications of such a change. When noncash working capital decreases, it releases tied-up cash and increases the cash flow of the firm. If a firm has bloated inventory or gives out credit too easily, managing one or both components more efficiently can reduce working capital and be a source of positive cash flows into the immediate future—three, four, or even five years. The question, however, becomes whether it can be a source of cash flows for longer than that. At some point in time, there will be no more inefficiencies left in the system, and any further decreases in working capital can have negative consequences for revenue growth and profits. Therefore, it appears that for firms with positive working capital, decreases in working capital are feasible only for short periods. In fact, once working capital is being managed efficiently, the working capital changes from year to year should be estimated using working capital as a percent of revenues. For example, consider a firm that has noncash working capital that represents 10 percent of revenues and that you believe that better management of working capital could reduce this to 6 percent of revenues. You could allow working capital to decline each year for the next four years from 10 percent to 6 percent, and, once this adjustment is made, begin estimating the working capital requirement each year as 6 percent of additional revenues. The following table provides estimates of the change in noncash working capital on this firm, assuming that current revenues are $1 billion and that revenues are expected to grow 10 percent a year for the next 15 years.
Can working capital itself be negative? Again, the answer is yes. Firms whose non-debt current liabilities exceed noncash current assets have negative noncash working capital. This is a thornier issue than negative changes in working capital. A firm that has a negative working capital is, in a sense, using supplier credit as a source of capital, especially if the negative working capital becomes larger as the firm becomes larger. A number of firms, with Amazon being the most prominent example, have used this strategy to grow. While this may seem like a costefficient strategy, there are potential downsides. The first is that supplier credit is generally not really free. To the extent that delaying paying supplier bills may lead to the loss of cash discounts and other price breaks, firms are paying for the privilege. Thus a firm that decides to adopt this strategy will have to compare the costs of this capital to more traditional forms of borrowing. The second downside is that a negative noncash working capital has generally been viewed by both accountants and ratings agencies as a source of default risk. To the extent that a firm's rating drops and interest rates paid by the firm increase, there may be costs created for other capital by using supplier credit as a source. As a practical question, you still have an estimation problem on your hands when forecasting working capital requirements for a firm that has negative noncash working capital. As in the previous scenario, with negative changes in noncash working capital, there is no reason why firms cannot continue to use supplier credit as a source of capital in the 195
short term. In the long term, however, we should not assume that noncash working capital will become more and more negative over time. At some point in the future we have to assume either that the change in noncash working capital is zero or that pressure will build for increases in working capital.
wcdata.xls: This dataset on the Web summarizes noncash working capital needs by industry group in the United States for the most recent quarter.
CONCLUSION When valuing a firm, the cash flows that are discounted should be after taxes and reinvestment needs but before debt payments. This chapter considered some of the challenges in coming up with this number for firms. The chapter began with the corrected and updated version of income described in Chapter 9. To state this income in after-tax terms, you need a tax rate. Firms generally state their effective tax rates in their financial statements, but these effective tax rates can be different from marginal tax rates. Although the effective tax rate can be used to arrive at the after-tax operating income in the early years, the tax rate used should converge on the marginal tax rate in future periods. For firms that are losing money and not paying taxes, the net operating losses that they are accumulating will protect some of their future income from taxation. The reinvestment that firms make in their own operations is then considered in two parts. The first part is the net capital expenditure of the firm, which is the difference between capital expenditures (a cash outflow) and depreciation (effectively a cash inflow). In this net capital expenditure, we include the capitalized operating expenses (such as R&D) and acquisitions. The second part relates to investments in noncash working capital, mainly inventory and accounts receivable. Increases in noncash working capital represent cash outflows to the firm, while decreases represent cash inflows. Noncash working capital at most firms tends to be volatile and may need to be smoothed out when forecasting future cash flows.
QUESTIONS AND SHORT PROBLEMS In the problems following, use an equity risk premium of 5.5 percent if none is specified. 1. You are valuing GenFlex, a small manufacturing firm, which reported paying taxes of $12.5 million on taxable income of $50 million and reinvesting $15 million in the most recent year. The firm has no debt outstanding, the cost of capital is 11%, and the marginal tax rate for the firm is 35%. Assuming that the firm's earnings and reinvestment are expected to grow 10% a year for three years and 5% a year forever after that, estimate the value of this firm: a. Using the effective tax rate to estimate after-tax operating income. b. Using the marginal tax rate to estimate after-tax operating income. c. Using the effective tax rate for the next three years and the marginal tax rate in year 4. 2. You are trying to estimate the free cash flow to the firm for RevTech, a technology firm. The firm reported $80 million in earnings before interest and taxes, capital expenditures of $30 million, and depreciation of $20 million in the most recent year. There are two additional complications: The firm had R&D expenses of $50 million in the most recent year. You believe that a three-year amortizable life is appropriate for this firm and the R&D expenses for the past three years have amounted to $20 million, $30 million, and $40 million respectively. The firm also made two acquisitions during the year—a cash-based acquisition for $45 million and a stockbased acquisition for $35 million. If the firm has no working capital requirements and a tax rate of 40%, estimate the free cash flow to the firm in the most recent year. 3. Lewis Clark, a firm in the travel business, reported earnings before interest and taxes of $60 million last year, but you have uncovered the following additional items of interest: The firm had operating lease expenses of $50 million last year and has a commitment to make equivalent payments for the next eight years. The firm reported capital expenditures of $30 million and depreciation of $50 million last year. However, the firm also made two acquisitions, one funded with cash for $50 million and another funded with a stock swap for $30 million. The amortization of these acquisitions is already included in the current year's 196
depreciation. The total working capital increased from $180 million at the start of the year to $200 million at the end of the year. However, the firm's cash balance was a significant portion of this working capital and increased from $80 million at the start of the year to $120 million at the end. (The cash is invested in T-bills.) The tax rate is 40%, and the firm's pretax cost of debt is 6%. Estimate the free cash flows to the firm last year. 4. The following is the balance sheet for Ford Motor Company as of December 31, 1994 (in millions).
The firm had revenues of $154,951 million in 1994 and cost of goods sold of $103,817 million. a. Estimate the net working capital. b. Estimate the noncash working capital. c. Estimate noncash working capital as a percent of revenues. 5. Continuing problem 4, assume that you expect Ford's revenues to grow 10% a year for the next five years. a. Estimate the expected changes in noncash working capital each year, assuming that noncash working capital as a percent of revenues remains at 1994 levels. b. Estimate the expected changes in noncash working capital each year, assuming that noncash working capital as a percent of revenues will converge on the industry average of 4.3% of revenues. 6. Newell Stores is a retail firm that reported $1 billion in revenues, $80 million in after-tax operating income, and noncash working capital of –$50 million last year. a. Assuming that working capital as a percent of revenues remains unchanged next year and that there are no net capital expenditures, estimate the free cash flow to the firm if revenues are expected to grow 10%. b. If you are projecting free cash flows to the firm for the next 10 years, would you make the same assumptions about working capital? Why or why not? A negative effective tax rate usually arises because a firm is reporting an income in its tax books (on which it pays taxes) and a loss in its reporting books. 1
Since the effective tax rate is based on the taxes paid (which comes from the tax statement) and the reported income, the effective tax rate will be lower than the marginal tax rate for firms that change accounting methods to inflate reported earnings. 2
It is only the tax-deductible amortization that really matters. To the extent that amortization is not tax deductible, you would look at the EBI T before the amortization and not consider it while estimating net capital expenditures. 3
197
CHAPTER 11 Estimating Growth The value of a firm is the present value of expected future cash flows generated by the firm. The most critical input in valuation, especially for high-growth firms, is the growth rate to use to forecast future revenues and earnings. This chapter considers how best to estimate these growth rates for firms, including those with low revenues and negative earnings. There are three basic ways of estimating growth for any firm. One is to look at the growth in a firm's past earnings —its historical growth rate. While this can be a useful input when valuing stable firms, there are both dangers and limitations in using this growth rate for high-growth firms. The historical growth rate can often not be estimated, and even if it can, it cannot be relied on as an estimate of expected future growth. The second is to trust the analysts who follow the firm to come up with the right estimate of growth for the firm, and to use that growth rate in valuation. Although many firms are widely followed by analysts, the quality of growth estimates, especially over longer periods, is poor. Relying on these growth estimates in a valuation can lead to erroneous and inconsistent estimates of value. The third is to estimate the growth from a firm's fundamentals. A firm's growth ultimately is determined by how much is reinvested into new assets and the quality of these investments, with investments widely defined to include acquisitions, building distribution channels, or even expanding marketing capabilities. By estimating these inputs, you are, in a sense, estimating a firm's fundamental growth rate.
THE IMPORTANCE OF GROWTH A firm can be valuable because it owns assets that generate cash flows now or because it is expected to acquire such assets in the future. The first group of assets is categorized as assets in place and the second as growth assets. Figure 11.1 presents a financial balance sheet for a firm. Note that an accounting balance sheet can be very different from a financial balance sheet, since accounting for growth assets tends to be both conservative and inconsistent. Figure 11.1 Financial View of a Firm
For high-growth firms, accounting balance sheets do a poor job of summarizing the values of the assets of the firm because they mostly ignore the largest component of value, which is future growth. The problems are exacerbated for firms that invest in R&D, because the book value will not include the most important asset at these firms—the research asset.
HISTORICAL GROWTH When estimating the expected growth for a firm, we generally begin by looking at the firm's history. How rapidly have the firm's operations, as measured by revenues or earnings, grown in the recent past? While past growth is not always a good indicator of future growth, it does convey information that can be valuable while making estimates for the future. This section begins by looking at measurement issues that arise when estimating past growth, and then considers how past growth can be used in projections.
Estimating Historical Growth Given a firm's earnings history, estimating historical growth rates may seem like a simple exercise but there are several measurement problems that may arise. In particular, you can ge t very different values for historical growth, depending on how the average is estimated and whether you allow for compounding in values over time. Estimating growth rates can also be complicated by the presence of negative earnings in the past or in the current period.
Arithmetic versus Geometric Averages 198
The average growth rate can vary depending on whether it is an arithmetic average or a geometric average. The arithmetic average is the simple average of past growth rates, while the geometric mean takes into account the compounding that occurs from period to period:
where gt = Growth rate in year t
where Earningst = earnings in year t The two estimates can be very different, especially for firms with volatile earnings. The geometric average is a much more accurate measure of true growth in past earnings, especially when year-to-year growth has been erratic. In fact, the point about arithmetic and geometric growth rates also applies to revenues, though the difference between the two growth rates tend to be smaller for revenues than for earnings. For firms with volatile earnings and revenues, the caveats about using arithmetic growth carry even more weight.
ILLUSTRATION 11.1: Differences between Arithmetic and Geometric Averages: Motorola (1994–1999) The following table reports the revenues, EBITDA, EBIT, and net income for Motorola for each year from 1994 to 1999. The arithmetic and geometric average growth rates in each series are reported at the bottom of the table.
The arithmetic average growth rate is higher than the geometric average growth rate for all four items, but the difference is much larger with net income and operating income (EBIT) than it is with revenues and EBITDA. This is because the net and operating income are the most volatile of the numbers, with a standard deviation in year-to-year changes of over 140%. Looking at the net and operating income in 1994 and 1999, it is also quite clear that the geometric averages are much better indicators of true growth. Motorola's operating income grew only marginally during the period, and this is reflected in its geometric average growth rate, which is 4.31%, but not in its arithmetic average growth rate, which indicates much faster growth. Motorola's net income dropped by almost 50% during the period. This is reflected in its negative geometric average growth rate but its arithmetic average growth rate is 36.91%.
Linear and Log-Linear Regression Models The arithmetic mean weights percentage changes in earnings in each period equally and ignores compounding effects in earnings. The geometric mean considers compounding but focuses on the first and the last earnings observations in the series—it ignores the information in the intermediate observations and any trend in growth rates that may have developed over the period. These problems are at least partially overcome by using ordinary least squares (OLS)1 regressions of earnings per share (EPS) against time. The linear version of this model is:
where EPSt = Earnings per share in period t t = Time period t The slope coefficient on the time variable is a measure of earnings change per time period. The problem, however, with the linear model is that it specifies growth in terms of dollar EPS and is not appropriate for projecting future growth, given compounding. The log-linear version of this model converts the coefficient into a percentage change: 199
where ln(EPSt) = Natural logarithm of earnings per share in period t t = Time period t The coefficient b on the time variable becomes a measure of the percentage change in earnings per unit time.
ILLUSTRATION 11.2: Linear and Log-Linear Models of Growth: General Electric (1991 to 2000) The earnings per share from 1991 until 2000 are provided for General Electric (GE) in the following table with the percentage changes and the natural logs of the earnings per share computed each year:
There are a number of ways in which we can estimate the growth rate in earnings per share at GE between 1991 and 2000. One is to compute the arithmetic and geometric averages:
The second is to run a linear regression of earnings per share against a time variable (where the earliest year is given a value of 1, the next year a value of 2, and so on):
This regression would indicate that the earnings per share increased 9.52 cents a year from 1991 to 2000. We can convert it into a percent growth in earnings per share by dividing this change by the average earnings per share over the period:
Finally, you can regress ln(EPS) against the time variable:
The coefficient on the time variable here can be viewed as a measure of compounded percent growth in earnings per share; GE's earnings per share grew at 13.35% a year based on this regression. The numbers are close using all the approaches because there is so little variability in the growth rate of earnings per share at GE. For companies with more volatile earnings, the differences will be much larger.
Negative Earnings Measures of historical growth are distorted by the presence of negative earnings numbers. The percentage change in earnings on a year-by-year basis is defined as:
If EPSt–1 is negative, this calculation yields a meaningless number. This extends into the calculation of the geometric mean. If the EPS in the initial time period is negative or zero, the geometric mean is not meaningful. Similar problems arise in log-linear regressions, since the EPS has to be greater than zero for the log transformation to exist. There are at least two ways of trying to get meaningful estimates of earnings growth for firms with negative earnings. One is to run the linear regression of EPS against time specified in the previous regression:
200
The growth rate can then be approximated as follows:
This assumes that the average EPS over the time period is positive. Another approach to estimating growth for these firms uses the higher of the two numbers (EPSt or EPSt–1) in the denominator:
Alternatively, you could use the absolute value of EPS in the previous period. Note that these approaches to estimating historical growth do not provide any information on whether these growth rates are useful in predicting future growth. It is not incorrect, and, in fact, it may be appropriate to conclude that the historical growth rate is not meaningful when earnings are negative and to ignore it in predicting future growth.
ILLUSTRATION 11.3: Negative Earnings: Tesla Motors and Aracruz Celulose The problems with estimating earnings growth when earnings are negative can be seen even for firms that have only negative earnings. For instance, Tesla Motors, the electric automobile company, reported operating earnings (EBIT) of –$52 million in 2009 and –$154 million in 2010. Clearly, the firm's earnings deteriorated, but estimating a standard earnings growth rate would lead us to the following growth rate:
Now consider Aracruz, a Brazilian paper and pulp company, susceptible like other firms in the industry to the ebbs and flows of commodity prices. The following table reports the earnings per share at the firm from 1995 to 2000.
Year EPS in Brazilian Reals 1995 0.302 1996 0.041 1997 0.017 1998 –0.067 1999 0.065 2000 0.437 The negative net income (and earnings per share) numbers in 1998 make the estimation of a growth rate in 1999 problematic. For instance, the firm has a loss per share of 0.067 BR in 1998 and a profit per share of 0.065 BR in 1999. The growth rate in earnings per share estimated using the conventional equation would be:
This growth rate, a negative number, makes no sense given the improvement in earnings during the year. There are two fixes to this problem. One is to replace the actual earnings per share in the denominator with the absolute value:
The other is to use the higher of the earnings per share from the two years, yielding:
While the growth rate is now positive, as you would expect it to be, the values for the growth rates themselves are not very useful for making estimates for the future.
Time Series Models to Predict Earnings per Share Time series models use the same historical information as the simpler models described in the previous section. They attempt to extract better predictions from this data, however, through the use of sophisticated statistical techniques.
Box-Jenkins Models Box and Jenkins (1976) developed a procedure for analyzing and forecasting univariate time series data using an autoregressive integrated moving average (ARIMA) model. Autoregressive integrated moving average models model a value in a time series as a linear combination of past values and past errors (shocks). Since historical data is used, these models are appropriate as long as the data does not show a time trend or a dependence on outside events or variables. ARIMA models are usually denoted by the notation:
where p = Degree of the autoregressive part d = Degree of differencing 201
q = Degree of the moving average process The mathematical model can then be written as follows:
where wt = Original data series or difference of degree d of the original data φ1, φ2 ... φp = Autoregressive parameters θ0 = Constant term θ1, θ2,... θq = Moving average parameters εt = Independent disturbances, random error ARIMA models can also adjust for seasonality in the data, in which case the model is denoted by the notation:
where s = Seasonal parameter of length n
Time Series Models in Earnings Most time series models used in forecasting earnings are built around quarterly earnings per share. In a survey paper, Bathke and Lorek (1984) point out that three time-series models have been shown to be useful in forecasting quarterly earnings per share. All three models are seasonal autoregressive integrated moving average (SARIMA) models, since quarterly earnings per share have a strong seasonal component. The first model, developed by Foster (1977), allows for seasonality in earnings and is a follows:
This model was extended by Griffin2 and Watts3 to allow for a moving average parameter:
where θ1 = First-order moving average [MA(1)] parameter Θ = First-order seasonal moving average parameter εt = Disturbance realization at the end of quarter t The third time series model, developed by Brown and Rozeff (1979), is similar in its use of seasonal moving average parameter:
How Good Are Time Series Models at Predicting Earnings? Time series models do better than naive models (using past earnings) in predicting earnings per share in the next quarter. The forecast error (i.e., the difference between the actual earnings per share and forecasted earnings per share) from the time series models is, on average, smaller than the forecast error from naive models (such as simple averages of past growth). The superiority of the models over naive estimates declines with longer-term forecasts, suggesting that the estimated time series parameters are not stationary. Among the time series models themselves, there is no evidence that any one model is dominant, in terms of minimizing forecast error, for every firm in the sample. The gain from using the firm-specific best models, rather than using the same model for every firm is relatively small.
Limitations in Using Time Series Models in Valuation There are several concerns in using time series models for forecasting earnings in valuation. First, time series models require a lot of data, which is why most of them are built around quarterly earnings per share. In most valuations, the focus is on predicting annual earnings per share and not on quarterly earnings. Second, even with quarterly earnings per share, the number of observations is limited for most firms to 10 to 15 years of data (40 to 60 quarters of data), leading to large estimation errors4 in time series model parameters and in the forecasts. Third, the superiority of earnings forecasts from time series models declines as the forecasting period is extended. Given that earnings forecasts in valuation have to be made for several years rather than a few quarters, the value of time series models may be limited. Finally, studies indicate that analyst forecasts dominate even the best time series 202
models in forecasting earnings. In conclusion, time series models are likely to work best for firms that have a long history of earnings and where the parameters of the models have not shifted significantly over time. For the most part, however, the cost of using these models is likely to exceed their benefits, at least in the context of valuation.
Usefulness of Historical Growth Is the growth rate in the past a good indicator of growth in the future? Not necessarily. In this section we consider how good historical growth is as a predictor of future growth for all firms, and why the changing size and volatile businesses of many firms can undercut growth projections.
Higgledy-Piggledy Growth Past growth rates are useful in forecasting future growth, but they have considerable noise associated with them. In a study of the relationship between past growth rates and future growth rates, Little (1960) coined the term “higgledy-piggledy growth” because he found little evidence that firms that grew fast in one period continued to grow fast in the next period. In the process of running a series of correlations between growth rates in consecutive periods of different length, he frequently found negative correlations between growth rates in the two periods, and the average correlation across the two periods was close to zero (0.02). If past growth is not a reliable indicator of future growth at many firms, it becomes even less so at smaller firms. The growth rates at smaller firms tend to be more volatile than growth rates at other firms in the market. The correlation between growth rates in earnings in consecutive time periods (five-year, three-year, and one-year) for firms in the United States, categorized by market value, is reported in Figure 11.2. Figure 11.2 Correlation in Earnings: One-, Three-, and Five-Year Time Periods
While the correlations tend to be higher across the board for one-year growth rates than for three-year or five-year growth rates in earnings, they are also consistently lower for smaller firms than they are for the rest of the market. This would suggest that you should be more cautious about using past growth, especially in earnings, for forecasting future growth at these firms.
Revenue Growth versus Earnings Growth In general, revenue growth tends to be more persistent and predictable than earnings growth. This is because ac counting choices have a far smaller effect on revenues than they do on earnings. Figure 11.3 compares the correlations in revenue and earnings growth over five-year periods at U.S. firms. Revenue growth is consistently more correlated over time than earnings growth. The implication is that historical growth in revenues is a far more useful number when it comes to forecasting than historical growth in earnings. Figure 11.3 Correlation in Earnings and Revenues—Five-Year Time Periods
203
Effects of Firm Size Since the growth rate is stated in percentage terms, the role of the size of the firm has to be weighed in the analysis. It is easier for a firm with $10 million in earnings to generate a 50 percent growth rate than it is for a firm with $500 million in earnings. Since it becomes harder for firms to sustain high growth rates as they become larger, past growth rates for firms that have grown dramatically in size may be difficult to sustain in the future. While this is a problem for all firms, it is a particular problem when analyzing small and growing firms. While the fundamentals at these firms, in terms of management, products, and underlying markets, may not have changed, it will still be difficult to maintain historical growth rates as the firms double or triple in size. The true test for a small firm lies in how well it handles growth. Some firms continue to deliver their products and services efficiently as they grow. In other words, they are able to scale up successfully. Other firms have had much more difficulty replicating their success as they become larger. In analyzing small firms, therefore, it is important that you look at plans to increase growth but it is even more critical that you examine the systems in place to handle this growth.
ILLUSTRATION 11.4: Cisco: Earnings Growth and Size of the Firm—The Glory Days (1989 to 1999) and Follow-up Cisco's evolution from a firm with $28 million in revenues and net income of about $4 million in 1989 to revenues in excess of $12 billion and net income of more than $2 billion in 1999 is reported in the following table:
While this table presents the results of a phenomenally successful decade for Cisco, it does suggest that you should be cautious about assuming that the firm will continue to grow at a similar rate in the future for two reasons. First, the growth rates tapered off as the firm became larger toward the end of the 1990s. Second, if you assume that Cisco will maintain its historic growth of 1990 to 1999 (estimated using the geometric average) for the following five years, the revenue and earnings will be immense. That is to say, if operating income continued to grow at 86.57% from 2000 to 2005, Cisco's operating income would have been $78 billion in 2005. Third, Cisco's growth came primarily from aquiring of small firms with promising technologies and using its capabilities to commercially develop these technologies. In 1999, for instance, Cisco acquired 15 firms and these acquisitions accounted for almost 80% of its reinvestment that year. If you assume that Cisco will continue to grow at historical rates, you are assuming that the number of acquisitions also will grow at the same rate. Thus Cisco would have to acquire almost 80 firms five years later to maintain the growth rate it had between 1990 and 1999. The difficulties of scaling up growth are clear when we look at Cisco between 2000 and 2010. While Cisco's game plan did not change—it continued to acquire companies and push for higher growth—the aggregate revenues and earnings were not responsive to the company's efforts. The compounded annual growth rate in revenues at Cisco declined to 7.78% between 2000 and 2010, and the compounded annual growth rate in operating income at Cisco between 2000 and 2010 was only 7.12%, both steep drop-offs from the growth rate in the prior decade.
204
histgr.xls: This dataset on the Web summarizes historical growth rates in earnings and revenues by industry group for the United States.
HISTORICAL GROWTH AT HIGH-GROWTH AND YOUNGER FIRMS The presence of negative earnings, volatile growth rates over time, and the rapid changes that high-growth firms go through over time make historical growth rates unreliable indicators of future growth for these firms. Notwithstanding this, you can still find ways to incorporate information from historical growth into estimates of future growth, if you follow these general guidelines: Focus on revenue growth, rather than earnings growth, to get a measure of both the pace of growth and the momentum that can be carried forward into future years. Revenue growth is less volatile than earnings growth and is much less likely to be swayed by accounting adjustments and choices. Rather than looking at average growth over the past few years, look at growth each year. This can provide information on how the growth is changing as the firm becomes larger, and can help when making projections for the future. Use historical growth rates as the basis for projections only in the near future (next year or two), since technologies can change rapidly and undercut future estimates. Consider historical growth in the overall market and in other firms that are serving it. This information can be useful in deciding what the growth rates of the firm that you are valuing will converge on over time.
ANALYST ESTIMATES OF GROWTH Equity research analysts provide not only recommendations on the firms they follow but also estimates of earnings and earnings growth for the future. How useful are these estimates of expected growth from analysts and how, if at all, can they be used in valuing firms? This section considers the process that analysts follow to estimate expected growth and follows up by examining why such growth rates may not be appropriate when valuing some firms.
Who Do Analysts Follow? The number of analysts tracking firms varies widely across firms. At one extreme are firms like Apple, GE, and Microsoft that are followed by dozens of analysts. At the other extreme, there are hundreds of firms that are not followed by any analysts. Why are some firms more heavily followed than others? These seem to be some of the determinants:
Market capitalization. The larger the market capitalization of a firm, the more likely it is to be followed by analysts. Institutional holding. The greater the percent of a firm's stock that is held by institutions, the more likely it is to be followed by analysts. The open question, though, is whether analysts follow institutions or whether institutions follow analysts. Given that institutional investors are the biggest clients of equity research analysts, the causality probably runs both ways. Trading volume. Analysts are more likely to follow liquid stocks. Here again, though, it is worth noting that the presence of analysts and buy (or sell) recommendations on a stock may play a role in increasing trading volume.
Information in Analyst Forecasts There is a simple reason to believe that analyst forecasts of growth should be better than using historical growth rates. Analysts, in addition to using historical data, can avail themselves of five other types of information that may be useful in predicting future growth: 1. Firm-specific information that has been made public since the last earnings report. Analysts can use information that has come out about the firm since the last earnings report, to make predictions about future growth. This information can sometimes lead to significant reevaluation of the firm's expected cash flows. 2. Macroeconomic information that may impact future growth. The expected growth rates of all firms are affected by economic news on GNP growth, interest rates, and inflation. Analysts can update their projections of future growth as new information comes out about the overall economy and about changes in fiscal and monetary policy. Information, for instance, that shows the economy growing at a faster rate than forecast will result in analysts increasing their estimates of expected growth for cyclical firms. 3. Information revealed by competitors on future prospects. Analysts can also condition their growth estimates for a firm on information revealed by competitors on pricing policy and future growth. For instance, a negative earnings report by one telecommunications firm can lead to a reassessment of earnings for other telecommunications firms. 4. Private information about the firm. Analysts sometimes have access to private information about the firms they follow that may be relevant in forecasting future growth. This avoids answering the delicate question of when private information becomes illegal inside information. There is no doubt, however, that good private information can lead to significantly better estimates of future growth. In an attempt to restrict this type of information leakage, the SEC issued new regulations preventing firms from selectively revealing information to a few analysts 205
or investors. Outside the United States, however, firms routinely convey private information to analysts following them. 5. Public information other than earnings. Models for forecasting earnings that depend entirely on past earnings data may ignore other publicly available information that is useful in forecasting future earnings. It has been shown, for instance, that other financial variables such as earnings retention, profit margins, and asset turnover are useful in predicting future growth. Analysts can incorporate information from these variables into their forecasts.
Quality of Earnings Forecasts If firms are followed by a large number of analysts5 and these analysts are indeed better informed than the rest of the market, the forecasts of growth that emerge from analysts should be better than estimates based on either historical growth or other publicly available information. But is this presumption justified? Are analyst forecasts of growth superior to other forecasts? The general consensus from studies that have looked at short-term forecasts (one quarter ahead to four quarters ahead) of earnings is that analysts provide better forecasts of earnings than models that depend purely on historical data. The mean relative absolute error, which measures the absolute difference between the actual earnings and the forecast for the next quarter, in percentage terms, is smaller for analyst forecasts than it is for forecasts based on historical data. Two other studies shed further light on the value of analysts' forecasts. Crichfield, Dyckman, and Lakonishok (1978) examined the relat ive accuracy of forecasts in the “Earnings Forecaster,” a publication from Standard & Poor's that summarizes forecasts of earnings from more than 50 investment firms. They measured the squared forecast errors by month of the year and computed the ratio of analyst forecast error to the forecast error from time series models of earnings. They found that the time series models actually outperform analyst forecasts from April until August, but underperform them from September through January. They hypothesized that this is because there is more firm-specific information available to analysts during the latter part of the year. The other study, by O'Brien (1988), compared consensus analyst forecasts from the Institutions Brokers Estimate System (I/B/ E/S) with time series forecasts from one quarter ahead to four quarters ahead. The analyst forecasts outperformed the time series model for one-quarter-ahead and two-quarters-ahead forecasts, did as well as the time series model for three-quarters-ahead forecasts, and did worse than the time series model for four-quarters-ahead forecasts. Thus, the advantage gained by analysts from firm-specific information seems to deteriorate as the time horizon for forecasting is extended. In valuation, the focus is more on long-term growth rates in earnings than on next quarter's earnings. There is little evidence to suggest that analysts provide superior forecasts of earnings when the forecasts are over three or five years. An early study by Cragg and Malkiel (1968) compared long-term forecasts by five investment management firms in 1962 and 1963 with actual growth over the following three years to conclude that analysts were poor longterm forecasters. This view is contested by Vander Weide and Carleton (1988), who found that the consensus prediction of five-year growth in the I/B/E/S is superior to historically oriented growth measures in predicting future growth. There is an intuitive basis for arguing that analyst predictions of growth rates must be better than time series or other historical data–based models simply because they use more information. The evidence indicates, however, that this superiority in forecasting is surprisingly small for long-term forecasts and that past growth rates play a significant role in determining analyst forecasts. There is one final consideration. Analysts generally forecast earnings per share, and most services report these estimates. When valuing a firm, you need forecasts of operating income and the growth in earnings per share will usually not be equal to the growth in operating income. In general, the growth rate in operating income should be lower than the growth rate in earnings per share. Thus, even if you decide to use analyst forecasts, you will have to adjust them if you are trying to forecast growth rates in operating income or revenues.
How Do You Use Analyst Forecasts in Estimating Future Growth? The information in the growth rates estimated by other analysts can and should be incorporated into the estimation of expected future growth. There are four factors that determine the weight assigned to analyst forecasts in predicting future growth: 1. Amount of recent firm-specific information. Analyst forecasts have an advantage over historical data–based models because they incorporate more recent information about the firm and its future prospects. This advantage is likely to be greater for firms where there have been significant changes in management or business conditions in the recent past, for example, a restructuring or a shift in government policy relating to the firm's underlying business. 2. Number of analysts following the stock. Generally speaking, the larger the number of analysts following a stock, the more informative is their consensus forecast, and the greater should be the weight assigned to it in analysis. The informational gain from having more analysts is diminished somewhat by the well-established fact that most analysts do not act independently, resulting in a high correlation across analysts' revisions of expected earnings. 206
3. Extent of disagreement between analysts. While consensus earnings growth rates are useful in valuation, the extent of disagreement between analysts measured by the standard deviation in growth predictions is also a useful measure of the reliability of the consensus forecasts. Givoly and Lakonsihok (1984) found that the dispersion of earnings is correlated with other measures of risk such as beta and is a good predictor of expected returns. 4. Quality of analysts following the stock. This is the hardest of the variables to quantify. One measure of quality is the size of the forecast error made by analysts following a stock, relative to models that use only historical data— the smaller this relative error, the larger the weight that should be attached to analyst forecasts. Another measure is the effect on stock prices of analyst revisions—the more informative the forecasts, the greater the effect on stock prices. There are some who argue that the focus on consensus forecasts misses the point that some analysts are better than others in predicting earnings, and that their forecasts should be isolated from the rest and weighted more. Analyst forecasts may be useful in coming up with a predicted growth rate for a firm, but there is a danger to blindly following consensus forecasts. Analysts often make significant errors in forecasting earnings, partly because they depend on the same data sources (which might have been erroneous or misleading) and partly because they sometimes overlook significant shifts in the fundamental characteristics of the firm. The secret to successful valuation often lies in discovering inconsistencies between analysts' forecasts of growth and a firm's fundamentals. The next section examines this relationship in more detail.
FUNDAMENTAL DETERMINANTS OF GROWTH With both historical and analyst estimates, growth is an exogenous variable that affects value but is divorced from the operating details of the firm. The soundest way of incorporating growth into value is to make it endogenous i.e., tie in more closely to the actions that a business takes to create and sustain that growth. This section begins by considering the relationship between fundamentals and growth in equity income, and then moves on to look at the determinants of growth in operating income.
Growth in Equity Earnings When estimating cash flows to equity, we usually begin with estimates of net income, if we are valuing equity in the aggregate, or earnings per share, if we are valuing equity per share. This section begins by presenting the fun damentals that determine expected growth in earnings per share and then move on to consider a more expanded version of the model that looks at growth in net income.
Growth in Earnings per Share The simplest relationship determining growth is one based on the retention ratio (percentage of earnings retained in the firm) and the return on equity on its projects. Firms that have higher retention ratios and earn higher returns on equity should have much higher growth rates in earnings per share than firms that do not share these characteristics. To establish this, note that:
where gt = Growth rate in net income NIt = Net income in year t Also note that the ROE in period t can be written as NI in period t divided by the Book value of equity in period t – 1. Given the definition of return on equity, the net income in year t – 1 can be written as:
where ROEt–1 = Return on equity in year t – 1 The net income in year t can be written as:
Assuming that the return on equity is unchanged (i.e., ROEt = ROEt–1 = ROE):
where b is the retention ratio. Note that the firm is not being allowed to raise equity by issuing new shares. Consequently, the growth rate in net income and the growth rate in earnings per share are the same in this 207
formulation.
ILLUSTRATION 11.5: Growth in Earnings per Share This illustration considers the expected growth rate in earnings based on the retention ratio and return on equity for three firms—Consolidated Edison, a regulated utility that provides power to New York City and its environs; Procter & Gamble, a leading brand-name consumer product firm; and Intel, the technology giant—in 2010. The following table summarizes the returns on equity, retention ratios, and expected growth rates in earnings for the three firms in 2010:
Intel has the highest expected growth rate in earnings per share, assuming that it can maintain its current return on equity and retention ratio. Procter & Gamble also can be expected to post a healthy growth rate, notwithstanding the fact that it pays out more than 50% of its earnings as dividends because of its high return on equity. Con Ed, on the other hand, has a very low expected growth rate because its return on equity and retention ratio are anemic.
Growth in Net Income If we relax the assumption that the only source of equity is retained earnings, the growth in net income can be different from the growth in earnings per share. Intuitively, note that a firm can grow net income significantly by issuing new equity to fund new projects, while earnings per share stagnates. To derive the relationship between net income growth and fundamentals, we need a measure of investment that goes beyond retained earnings. One way to obtain such a measure is to estimate how much equity the firm reinvests back into its businesses in the form of net capital expenditures and investments in working capital.
Dividing this number by the net income gives us a much broader measure of the equity reinvestment rate:
Unlike the retention ratio, this number can be well in excess of 100 percent with the excess being funded with new equity. The expected growth in net income can then be written as:
ILLUSTRATION 11.6: Growth in Net Income To estimate growth in operating income based on fundamentals, we look at three firms—Coca-Cola, Nestlé, and Sony. The following table estimates the components of equity reinvestment and uses it to estimate the reinvestment rate for each of the firms. We also present the return on equity and the expected growth rate in net income at each of these firms in 2010:
The pluses and minuses of this approach are visible in the table. The approach much more accurately captures the true reinvestment in the firm by focusing not on what was retained but on what was reinvested. The limitation of the approach is that the ingredients that go into the reinvestment—capital expenditures, working capital change, and net debt issued—are all volatile numbers. Note that Sony had more depreciation than capital expenditures in 2010 and a decrease in working capital, and paid off debt during the year. The net reinvestment rate is negative. If it continues on this path, it will have negative growth. In fact, it would probably be much more realistic to look at the average reinvestment rate over three or five years, rather than just the current year. We will return to examine this question in more depth when we look at growth in operating income.
Determinants of Return on Equity Both earnings per share and net income growth are affected by the return on equity of a firm. The return on equity is affected by how much debt the firm chooses to use to fund its projects. In the broadest terms, increasing debt will lead to a higher return on equity if the after-tax return on capital exceeds the after-tax interest rate paid on debt. This is captured in the following formulation of return on equity:
where ROC = EBIT(1 - t)/(BV of debt + BV of equity-Cash) D/E = BV of debt/BV of equity i = Interest expense on debt/BV of debt t = Tax rate on ordinary income 208
In keeping with the fact that return on equity is based on book value, all of the inputs are also stated in terms of book value. The derivation is simple and is provided in a footnote.6 Using this expanded version of ROE, the growth rate can be written as:
The advantage of this formulation is that is allows use to model changes in leverage and evaluate the effects on growth.
ILLUSTRATION 11.7: Breaking Down Return on Equity To consider the components of return on equity, the following table looks at Consolidated Edison, Procter & Gamble, and Intel, three firms whose returns on equity were shown in Illustration 11.5:
Comparing these numbers to those reported in Illustration 11.5, you will note that the return on equity is close to our earlier estimates for Con Ed and P&G. The return on equity computed here is lower than the earlier estimate for Intel because it posted significant non operating profits in its net income. We have chosen to consider only operating income in the return on capital computation. To the extent that firms routinely report nonoperating income (or losses), the return on equity computed using the standard approach (net income divided by book equity) will be different from the return on equity computed heren. While this is not a serious concern for any of the three firms examined, we should be concerned if a high ROE is caused by a high D/E ratio, a low effective tax rate, or non operating profits. That ROE may not be sustainable. If the firm loses its tax breaks and the sources of nonoperating income dry up, the firm could very easily find itself with a return on capital that is lower than its book interest rate. If this occurs, leverage could bring down the return on equity of the firm.
AVERAGE AND MARGINAL RETURNS The return on equity is conventionally measured by dividing the net income in the most recent year by the book value of equity at the end of the previous year. Consequently, the return on equity measures the quality of both older projects that have been on the books for a substantial period and new projects from more recent periods. Since older investments represent a significant portion of the earnings, the average returns may not shift substantially for larger firms that are facing a decline in returns on new investments, because either of market saturation or competition. In other words, poor returns on new projects will have a lagged effect on the overall returns for the firm. In valuation, it is the returns that firms are making on their newer investments that convey the most information about a quality of a firm's projects. To measure these returns, we could compute a marginal return on equity by dividing the change in net income in the most recent year by the change in book value of equity in the prior year: Marginal return on equity = Δ Net incomet/Δ Book value of equityt–1 For example, Disney reported net income of $3.963 million on book value of equity of $35,425 million in 2010, resulting in an aggregate return on equity of 11.87 percent:
The marginal return on equity is computed as follows:
While we are not suggesting that Disney generated 37.32 percent on its new investments in 2010, it does show the momentum is upward in Disney's return on equity. Thus, a forward-looking estimate greater than 11.87 percent would be merited.
The Effects of Changing Return on Equity So far, this section has operated on the assumption that the overall return on equity remains unchanged over time. If we relax this assumption, we introduce a new component to growth—the effect of changing return on equity on existing investments over time. Consider, for instance, a firm that has a book value of equity of $100 million and a return on equity of 10 percent. If this firm improves its return on equity to 11 percent, it will post an earnings growth rate of 10 percent even if it does not reinvest any money. This additional growth can be written as a function of the change in the return on equity:
where ROEt is the return on equity in period t. This will be in addition to the fundamental growth rate computed as the product of the return on equity and the retention ratio. While increasing return on equity will generate a spurt in the growth rate in the period of the improvement, a decline in the return on equity will create a more than proportional drop in the growth rate in the period of the decline. 209
It is worth differentiating at this point between returns on equity on new investments and returns on equity on existing investments. The additional growth that we are estimating here comes not from new investments but by changing the return on existing investments. For lack of a better term, you could consider it “efficiency-generated growth.”
ILLUSTRATION 11.8: Effects of Changing Return on Equity: Con Ed In Illustration 11.5 we looked at Con Ed's expected growth rate based on its return on equity of 9.79% and its retention ratio of 36%. Assume that the firm will be able to improve its overall return on equity (on both new and existing investments) to 11% next year and that the retention ratio remains at 36%. The expected growth rate in earnings per share next year can then be written as:
After next year, the growth rate will subside to a more sustainable 3.96% (.11 × .36). How would the answer be different if the improvement in return on equity were only on new investments but not on existing assets? The expected growth rate in earnings per share can then be written as:
Thus, there is no additional growth created in this case. What if the improvement had been only on existing assets and not on new investments? Then, the expected growth rate in earnings per share can be written as:
Growth in Operating Income Just as equity income growth is determined by the equity reinvested back into the business and the return made on that equity investment, you can relate growth in operating income to total reinvestment made into the firm and the return earned on capital invested. We will consider three separate scenarios, and examine how to estimate growth in each, in this section. The first is when a firm is earning a high return on capital that it expects to sustain over time. The second is when a firm is earning a positive return on capital that is expected to increase over time. The third is the most general scenario, where a firm expects operating margins to change over time, sometimes from negative values to positive levels.
Stable Return on Capital Scenario When a firm has a stable return on capital, its expected growth in operating income is a product of the reinvestment rate (i.e., the proportion of the after-tax operating income that is invested in net capital expenditures and noncash working capital), and the quality of these reinvestments, measured as the return on the capital invested.
Both measures—the reinvestment rate and return on capital—should be forward looking, and the return on capital should represent the expected return on capital on future investments. In the rest of this section, we consider how best to estimate the reinvestment rate and the return on capital.
Reinvestment Rate The reinvestment rate measures how much a firm is plowing back to generate future growth. The reinvestment rate is often measured using the most recent financial statements for the firm. Although this is a good place to start, it is not necessarily the best estimate of the future reinvestment rate. A firm's reinvestment rate can ebb and flow, especially in firms that invest in relatively few large projects or acquisitions. For these firms, looking at an average reinvestment rate over time may be a better measure of the future. In addition, as firms grow and mature, their reinvestment needs (and rates) tend to decrease. For firms that have expanded significantly over the last few years, the historical reinvestment rate is likely to be higher than the expected future reinvestment rate. For these firms, industry averages for reinvestment rates may provide a better indication of the future than using numbers from the past. Finally, it is important that we continue treating R&D expenses and operating lease expenses consistently. The R&D expenses, in particular, need to be categorized as part of capital expenditures for purposes of measuring the reinvestment rate. 210
Return on Capital The return on capital is often based on the firm's return on capital on existing investments, where the book value of capital is assumed to measure the capital invested in these investments. Implicitly, we assume that the current accounting return on capital is a good measure of the true returns earned on existing investments, and that this return is a good proxy for returns that will be made on future investments. This assumption, of course, is open to question for the following reasons: The book value of capital might not be a good measure of the capital invested in existing investments, since it reflects the historical cost of these assets and accounting decisions on depreciation. When the book value understates the capital invested, the return on capital will be overstated; when book value overstates the capital invested, the return on capital will be understated. This problem is exacerbated if the book value of capital is not adjusted to reflect the value of the research asset or the capital value of operating leases. The operating income, like the book value of capital, is an accounting measure of the earnings made by a firm during a period. All the problems in using unadjusted operating income described in Chapter 9 continue to apply. Even if the operating income and book value of capital are measured correctly, the return on capital on existing investments may not be equal to the marginal return on capital that the firm expects to make on new investments, especially as you go further into the future. Given these concerns, we should consider not only a firm's current return on capital, but any trends in this return as well as the industry average return on capital. If the current return on capital for a firm is significantly higher than the industry average, the forecasted return on capital should be set lower than the current return to reflect the erosion that is likely to occur as competition responds. Finally, any firm that earns a return on capital greater than its cost of capital is earning an excess return. The excess returns are the result of a firm's competitive advantages or barriers to entry into the industry. High excess returns locked in for very long periods imply that this firm has a permanent competitive advantage.
ILLUSTRATION 11.9: Measuring the Reinvestment Rate, Return on Capital, and Expected Growth Rate: Tata Motors in 2010 In May 2010, we looked at Tata Motors, an Indian automobile company, which has been aggressive in its pursuit of growth through both internal investments and acquisitions over much of the past decade. Based upon its financial statements of 2009, we estimated a reinvestment rate of 116.83% and a return on capital of 11.81%:
Note that the effective tax rate (21%) was used to compute the after-tax operating income for both the reinvestment rate and the return on capital. The capital invested was obtained by summing up the book value of debt and equity at the end of the 2008 fiscal year (the beginning of the 2009 fiscal year) and netting out the cash and marketable securities at that point in time. If Tata Motors can maintain this return on capital and reinvestment rate going forward, its expected growth rate would be:
As we will see in the next illustration, maintaining this reinvestment going forward may be very difficult to do.
ILLUSTRATION 11.10: Current and Historical Averages: Reinvestment Rate and Return on Capital for Tata Motors Tata Motors has had a volatile history in terms of both reinvestment and returns on capital. Although the 2009 numbers were computed in the preceding illustration, those values have been in flux over the past five years. We summarize the numbers (in millions of rupees) for 2005 to 2009, with the aggregate in the last column:
211
The reinvestment rate has swung between –5.61% and 228.46% over the period, but the aggregate reinvestment rate over the period was 83.53%. We did a similar computation with the return on capital between 2005 and 2009.
The average return on capital between 2005 and 2009 was 14.98%. Using these averages for the reinvestment rate and return on capital generates a growth rate of 12.51%:
This does seem like a more sustainable value for the future.
fundgrEB.xls: This dataset on the Web summarizes reinvestment rates and return on capital by industry group in the United States for the most recent quarter.
NEGATIVE REINVESTMENT RATES: CAUSES AND CONSEQUENCES The reinvestment rate for a firm can be negative if its depreciation exceeds its capital expenditures or if the working capital declines substantially during the course of the year. For most firms, this negative reinvestment rate will be a temporary phenomenon reflecting lumpy capital expenditures or volatile working capital. For these firms, the current year's reinvestment rate (which is negative) can be replaced with an average reinvestment rate over the past few years or an industry average reinvestment rate. For other firms, the negative reinvestment rate may reflect our failure to incorporate acquisitions into capital expenditures (if the firm grows through acquisitions) or to capitalize R&D (or like expenses). For some firms, though, the negative reinvestment rate may be deliberate and how we deal with it will depend on why the firm is embarking on this path: Firms that have overinvested in capital equipment or working capital in the past may be able to live off past investment for a number of years, reinvesting little and generating higher cash flows for that period. If this is the case, we should use the negative reinvestment rate in forecasts and estimate growth based on improvements in return on capital. Once the firm has reached the point where it is efficiently using its resources, though, we should change the reinvestment rate to reflect expected growth. The more extreme scenario is a firm that has decided to liquidate itself over time, by not replacing assets as they become run-down and by drawing down working capital. In this case, the expected growth should be estimated using the negative reinvestment rate. Not surprisingly, this will lead to a negative expected growth rate and declining earnings over time.
Positive and Changing Return on Capital Scenario The analysis in the preceding section is based on the assumption that the return on capital remains stable over time. If the return on capital changes over time, the expected growth rate for the firm will have a second component, which will increase the growth rate if the return on capital increases and decrease the growth rate if the return on capital decreases.
For example, a firm that sees its return on capital improve from 10 to 11 percent while maintaining a reinvestment rate of 40 percent will have an expected growth rate of:
In effect, the improvement in the return on capital increases the earnings on existing assets and this improvement translates into an additional growth of 10 percent for the firm.
Marginal and Average Returns on Capital So far, we have looked at the return on capital as the measure that determines return. In reality, however, there are 212
two measures of returns on capital. One is the return earned by firm collectively on all of its investments, which we define as the average return on capital. The other is the return earned by a firm on just the new investments it makes in a year, which is the marginal return on capital. Changes in the marginal return on capital do not create a second-order effect, and the expected growth is a product of the marginal return on capital and the reinvestment rate. Changes in the average return on capital, however, will result in the additional impact on growth chronicled earlier.
Candidates for Changing Average Return on Capital What types of firms are likely to see their return on capital change over time? One category includes firms with poor returns on capital that improve their operating efficiency and margins, and consequently their return on capital. In these firms, the expected growth rate will be much higher than the product of the reinvestment rate and the return on capital. In fact, since the return on capital on these firms is usually low before the turnaround, small changes in the return on capital translate into big changes in the growth rate. Thus, an increase in the return on capital on existing assets from 1 percent to 2 percent doubles the earnings (resulting in a growth rate of 100 percent). Another category includes firms that have very high returns on capital on their existing investments but are likely to see these returns slip as competition enters the business, not only on new investments but also on existing investments.
ILLUSTRATION 11.11: Estimating Expected Growth with Changing Return on Capital: Titan Cement and Motorola In 2000, Titan Cement, a Greek cement company, reported operating income of 55,467 million drachmas on capital invested of 135,376 million drachmas. Using its effective tax rate of 24.5%, we estimate a return on capital for the firm of 30.94%:
Assume that the firm will see its return on capital drop on both its existing assets and its new investments to 29% next year and that its reinvestment rate will stay at 35%. The expected growth rate next year can be estimated as follows:
In contrast, consider Motorola in early 2000. The firm had a reinvestment rate of 52.99% and a return on capital of 12.18% in 1999. Assume that Motorola's return on capital will increase toward the industry average of 22.27% as the firm sheds the residue of its ill-fated Iridium investment and returns to its roots. Specifically, assume that Motorola's return on capital will increase from 12.18% to 17.22% over the following five years.7 For simplicity, also assume that the change occurs linearly over the next five years. The expected growth rate in operating income each year for the next five years can then be estimated as follows:8
The improvement in return on capital over the next five years will result in a higher growth rate in operating earnings at Motorola over that period. Note that this calculation assumes that the return on capital on new investments next year will be 17.22%.
chgrowth.xls: This spreadsheet allows you to estimate the expected growth rate in operating income for a firm where the return on capital is expected to change over time.
Negative Return on Capital Scenario The third and most difficult scenario for estimating growth is when a firm is losing money and has a negative return on capital. Since the firm is losing money, the reinvestment rate is also likely to be negative. To estimate growth in these firms, we have to move up the income statement and first project growth in revenues. Next, we use the firm's expected operating margin in future years to estimate the operating income in those years. If the expected margin in future years is positive, the expected operating income will also turn positive, allowing us to apply traditional valuation approaches in valuing these firms. We also estimate how much the firm has to reinvest to generate revenue growth growth, by linking revenues to the capital invested in the firm.
Growth in Revenues Many high-growth firms, while reporting losses, also show large increases in revenues from period to period. The first step in forecasting cash flows is forecasting revenues in future years, usually by forecasting a growth rate in revenues each period. In making these estimates, there are five points to keep in mind. 213
1. The rate of growth in revenues will decrease as the firm's revenues increase. Thus, a tenfold increase in revenues is entirely feasible for a firm with revenues of $2 million but unlikely for a firm with revenues of $2 billion. 2. Compounded growth rates in revenues over time can seem low, but appearances are deceptive. A compounded annual growth rate in revenues of 20 percent over ten years will increase revenues about six fold but an increase of 40 percent over 10 years will result in an almost 30-fold increase in revenues over the period. 3. While growth rates in revenues may be the mechanism that you use to forecast future revenues, you do have to keep track of the dollar revenues to ensure that they are reasonable, given the size of the overall market that the firm operates in. If the projected revenues for a firm 10 years out would give it a 90 or 100 percent share (or greater) of the overall market in a competitive marketplace, you clearly should reassess the revenue growth rate. 4. Assumptions about revenue growth and operating margins have to be internally consistent. Firms can post higher growth rates in revenues by adopting more aggressive pricing strategies but the higher revenue growth will then be accompanied by lower margins. 5. In coming up with an estimate of revenue growth, you have to make a number of subjective judgments about the nature of competition, the capacity of the firm that you are valuing to handle the revenue growth and the marketing capabilities of the firm.
ILLUSTRATION 11.12: Estimating Revenues at Tesla Motors and Linkedin This illustration considers two young, high-growth companies: Tesla Motors, the electric automaker, and Linkedin, a social media firm.
Estimates of growth for the firms in the initial years are based on the growth in revenues over the past year, but those growth rates start declining as the revenues scale up and approach the growth rate of the economy near year 10. As a check, we also examined how much the revenues at each of these firms would be in 10 years relative to more mature companies in the sector now. We compared revenues at Tesla Motors in 10 years to those of more established automobile companies such as Ford, Volvo, Toyota, and Fiat. With $5 billion-plus in revenues, Tesla Motors will remain a very small firm in a large market. It is difficult to find a company directly comparable to Linkedin, but Yahoo! revenues in 2010 were about $6 billion. We are assuming that Linkedin will have revenues that are close (about $5.5 billion) in 10 years.
Operating Margin Forecasts Before considering how to estimate the operating margins, let us begin with an assessment of where many highgrowth firms, early in the life cycle, stand when the valuation begins. They usually have low revenues and negative operating margins. If revenue growth converts low revenues into high revenues and operating margins stay negative, these firms not only will be worth nothing but are unlikely to survive. For firms to be valuable, the higher revenues eventually have to deliver positive earnings. In a valuation model, this translates into positive operating margins in the future. A key input in valuing a high-growth firm then is the operating margin you would expect it to have as it matures. In estimating this margin, you should begin by looking at the business that the firm is in. While many new firms claim to be pioneers in their businesses and some believe that they have no competitors, it is more likely that they are the first to find a new way of delivering a product or service that was previously delivered through other channels. Thus, Amazon.com might have been one of the first online retailers, but retailing was already an established business with hundreds of players. In fact, one can consider online retailers as logical successors to catalog retailers such as L.L. Bean and Lillian Vernon. Similarly, Yahoo! might have been one of the first Internet portals, but it was following the lead of newspapers that have used content and features to attract readers and used their readership to attract advertising. Using the average operating margin of competitors in the business may strike some as conservative. After all, they would point out, Amazon can hold less inventory and does not have the burden of carrying the operating leases that a brick and mortar retailer does (on its stores) and should, therefore, be more efficient about generating its revenues. This may be true, but it is unlikely that the operating margins for 214
online retailers can be persistently higher than their brick-and-mortar counterparts. If they were, you would expect to see a migration of traditional retailers to online retailing and increased competition among online retailers on price and products, driving the margin down. While the margin for the business in which a firm operates provides a target value, there are still two other estimation issues that you need to confront. Given that the operating margins in the early stages of the life cycle are negative, you first have to consider how the margin will improve from current levels to the target values. Generally, the improvements in margins will be greatest in the earlier years (at least in percentage terms) and then taper off as the firm approaches maturity. The second issue is one that is linked to revenue growth. Firms may be able to post higher revenue growth with lower margins but the trade-off has to be considered. While firms generally want both higher revenue growth and higher margin, the margin and revenue growth assumptions have to be consistent.
ILLUSTRATION 11.13: Estimating Operating Margins To estimate the operating margins for Tesla Motors, we begin by estimating the operating margins of established firms in the automobile sector. In 2010, the average pretax operating margin for firms in this sector was 10%. For Linkedin, we will use the average pretax operating margin of firms like Yahoo!, Google and Baidu, which is 25%. We will assume that both Tesla Motors and Linkedin will move toward their target margins, with greater marginal improvements9 in the earlier years and smaller ones in the later years. The following table summarizes the expected operating margins over time for both firms:
Tesla Motors Linkedin Current
–69.87%
8.23%
1
–43.25%
11.62%
2
–25.50%
13.31%
3
–13.67%
14.15%
4
–5.78%
14.58%
5
–0.52%
14.79%
6
2.99%
14.89%
7
5.33%
14.95%
8
6.88%
14.97%
9
7.92%
14.99%
10
8.61%
14.99%
Terminal year 10.00%
15.00%
Note that while margins improve for both companies, we are assuming that it will happen faster at Linkedin, a company that is already profitable, than at Tesla Motors, with its more substantial operating challenges. Since we estimated revenue growth in the preceding section and the margins in this one, we can now estimate the pretax operating income at each of the firms over the next 10 years:
As the margins move toward target levels and revenues grow, the operating income at each of the firms also increases.
MARKET SIZE, MARKET SHARE, AND REVENUE GROWTH Estimating revenue growth rates for a young firm in a new business may seem like an exercise in futility. While it is difficult to do, there are ways in which you can make the process tractable. One way is to work backward by first considering the share of the overall market that you expect your firm to have once it matures, and then determining the growth rate you would need to arrive at this market share. For instance, assume that you are analyzing an online toy retailer with $100 million in revenues currently. Assume also that the entire toy retail market had revenues of $70 billion last year. Assuming a 3 percent growth rate in overall toy market over the next 10 years and a market share of 5 percent for your firm in year 10, you would arrive at expected revenues of $4.703 billion for the firm in 10 years, and a compounded revenue growth rate of 46.98%.
215
Another approach is to forecast the expected growth rate in revenues over the next three to five years based on past growth rates. Once you estimate revenues in year 3 or 5, you can then forecast a growth rate based on the rate at which companies with similar revenues grow currently. For instance, assume that the online toy retailer had revenue growth of 200 percent last year (revenues went from $33 million to $100 million). You could forecast growth rates of 120 percent, 100 percent, 80 percent, and 60 percent for the next four years, leading to revenues of $1.267 billion in four years. You could then look at the average growth rate posted by retail firms with revenues between $1 billion and $1.5 billion last year and use that as the growth rate commencing in year 5.
Sales-to-Capital Ratio High revenue growth is a desirable objective, especially when accompanied by positive operating margins in future years. Firms do, however, have to invest to generate both revenue growth and positive operating margins in future years. This investment can take traditional forms (plant and equipment) but it should also include acquisitions of other firms, partnerships, investments in distribution and marketing capabilities, and research and development. To link revenue growth with reinvestment needs, we look at the revenue generated by each dollar of capital that we invest. This ratio, called the sales-to-capital ratio, allows us to estimate how much additional investment the firm has to make to generate the projected revenue growth. This investment can be in internal projects, acquisitions, or working capital. To estimate the reinvestment needs in any year then, you divide the revenue growth that you have projected (in dollar terms) by the sales-to-capital ratio. Thus, if you expect revenues to grow by $1 billion and you use a sales-to-capital ratio of 2.5, you would estimate a reinvestment need for this firm of $400 million ($1 billion/ 2.5). Lower sales-to-capital ratios increase reinvestment needs (and reduce cash flows) whereas higher sales-tocapital ratios decrease reinvestment needs (and increase cash flows). To estimate the sales-to-capital ratio, you should look at both a firm's past and the business it operates in. To measure this ratio for a firm, you divide changes in revenue each year by the reinvestment made that year. You should also look at the average ratio of sales to book capital invested in the business in which the firm operates. Linking operating margins to reinvestment needs is much more difficult to do, since a firm's capacity to earn operating income and sustain high returns comes from the competitive advantages that it acquires, partly through internal investment and partly through acquisitions. Firms that adopt a two-track strategy in investing, where one track focuses on generating higher revenues and the other on building up competitive strengths, should have higher operating margins and values than firms that concentrate on only revenue growth.
Imputed Return on Capital One of the dangers that you face when using a sales-to-capital ratio to generate reinvestment needs is that you might underestimate or overestimate your reinvestment needs. You can keep tabs on whether this is happening and correct it when it does by also estimating the after-tax return on capital of the firm each year through the analysis. To estimate the return on capital in a future year, you divide the estimated after-tax operating income in that year by the total capital invested in that firm in that year. The former number comes from your estimates of revenue growth and operating margins, while the latter can be estimated by aggregating the reinvestment made by the firm all the way through the future year. For instance, a firm that has $500 million in capital invested today and is assumed to reinvest $300 million next year and $400 million the year after will have capital invested of $1.2 billion at the end of the second year. For firms losing money today, the return on capital will be a negative number when the estimation begins but improve as margins improve. If you reinvest too little, the return on capital in the later years will be too high, while if you don't reinvest enough, it will be too low. Too low or high relative to what, you ask? There are two comparisons that are worth making. The first is to the average return on capital for mature firms in the business in which your firm operates—mature automobile companies in the case of Tesla Motors. The second is to the firm's own cost of capital. A projected return on capital of 40 percent for a firm with a cost of capital of 10 percent in a sector where returns on capital hover around 15 percent is an indicator that the firm is investing too little for the projected revenue growth and operating margins. Decreasing the sales-to-capital ratio until the return on capital converges on 15 percent would be prudent.
ILLUSTRATION 11.14: Estimated Sales-to-Capital Ratios and Implied Return on Capital To estimate how much Tesla Motors and Linkedin have to invest to generate the expected revenue growth, we estimate the current sales-tocapital ratio for each firm, the marginal sales to capital ratio in the last year, and the average sales-to-capital ratio for the businesses that each operates in:
Tesla Motors Linkedin Firm's sales to capital
0.26
1.93
Marginal sales to capital: most recent year 0.31
2.15
216
Industry average sales to capital
1.69
2.20
Sales-to-capital ratio used in valuation
2.00
2.20
We use the industry average of 2.20 for the sales to capital ratio for Linkedin, a little higher than its current sales-to-capital ratio and close to the marginal ratio in the most recent year. For Tesla, we use 2.00, a little higher than the industry average of 1.69, and assume that the current numbers are a reflection of its infrastructure investments, its start-up status, and its technology roots. Based on these estimates of the sales-to-capital ratio for each firm, we can now estimate how much each firm will have to reinvest each year for the next 10 years and the resulting return on capital:
The returns on capital at both firms converge to sustainable levels, at least relative to industry averages, by the terminal year. This suggests that our estimates of sales-to-capital ratios are reasonable.
margins.xls: This dataset on the Web summarizes operating and net margins, by industry, for the United States.
QUALITATIVE ASPECTS OF GROWTH The emphasis on quantitative elements—return on capital and reinvestment rates for profitable firms, and margins, revenue growth, and sales-to-capital ratios for unprofitable firms—may strike some as skewed. After all, growth is determined by a number of subjective factors—the quality of management, the strength of a firm's marketing, its capacity to form partnerships with other firms, and the management's strategic vision, among many others. Where, you might ask, is there room in the growth equations that have been presented in this chapter for these factors? The answer is that qualitative factors matter, but that they all ultimately have to show up in one or more of the quantitative inputs that determine growth. Consider the following: The quality of management plays a significant ro le in the returns on capital that you assume firms can earn on their new investments and in how long they can sustain these returns. Thus, the fact that a firm has a well-regarded management team may be one reason why you allow a firm's return on capital to remain well above the cost of capital. The marketing strengths of a firm and its choice of marketing strategy are reflected in the operating margins and turnover ratios that you assume for firms. Thus, it takes faith in a Coca-Cola's capacity to market its products effectively to assume a high turnover ratio and a high target margin. In fact, you can consider various marketing strategies, which trade off lower margins for higher turnover ratios, and consider the implications for value. The brand name of a firm's products and the strength of its distribution system also affect these estimates. Defining reinvestment broadly to include acquisitions, research and development, and investments in marketing and distribution allows you to consider different ways in which firms can grow. For some firms, reinvestment and growth come from acquisitions, while for other firms it may take the form of more traditional investments in plant and equipment. The effectiveness of these reinvestment strategies is 217
captured in the return on capital that you assume for the future, with more effective firms having higher returns on capital. The strength of the competition that firms face is in the background but it does determine how high excess returns (return on capital less cost of capital) will be, and how quickly they fade toward zero. Thus, qualitative factors are quantified and the growth implications are considered. If you cannot, you should remain skeptical about whether these factors truly affect value. Why is it necessary to impose this quantitative structure on growth estimate? One of the biggest dangers in valuing technology firms is that story telling can be used to justify growth rates that are neither reasonable nor sustainable. Thus, you might be told that Tesla Motors will grow 100% a year because the “green” movement is strong or that Coca-Cola will grow 20 percent a year because it has a great brand name. While there is truth in these stories, a consideration of how these qualitative views translate into the quantitative elements of growth is an essential step toward consistent valuations. Can different investors consider the same qualitative factors and come to different conclusions about the implications for returns on capital, margins, and reinvestment rates, and consequently, about growth? Absolutely. In fact, you would expect differences in opinion about the future and different estimates of value. The payoff to knowing a firm and the sector it operates in better than other investors is that your estimates of growth and value will be better than theirs. Unfortunately, this does not guarantee that your investment returns will be better than theirs.
CONCLUSION Growth is the key input in every valuation, and there are three sources for growth rates. One is the past, though both estimating and using historical growth rates can be difficult for most firms with their volatile and sometimes negative earnings. The second source is analyst estimates of growth. Though analysts may be privy to information that is not available to the rest of the market, this information does not result in growth rates that are superior to historical growth estimates. Furthermore, the analyst's emphasis on earnings per share growth can be a problem when forecasting operating income. The third and soundest way of estimating growth is to base it on a firm's fundamentals. The relationship of growth to fundamentals will depend on what growth rate we are estimating. To estimate growth in earnings per share, we looked at return on equity and retention ratios. To estimate growth in net income, we replaced the retention ratio with the equity reinvestment rate. To evaluate growth in operating income, we used return on capital and reinvestment rate. While the details vary from approach to approach, there are some common themes that emerge from these approaches. The first is that growth and reinvestment are linked, and estimates of one have to be linked with estimates of the other. Firms that want to grow at high rates over long periods have to reinvest to create that growth. The second is that the quality of growth can vary widely across firms, and the best measure of the quality of growth is the returns earned on investments. Firms that earn higher returns on equity and capital not only will generate higher growth, but that growth will add more to their value.
QUESTIONS AND SHORT PROBLEMS In the problems following, use an equity risk premium of 5.5 percent if none is specified. 1. Walgreen Company reported the following earnings per share from 1989 to 1994. Year EPS 1989 $1.28 1990 $1.42 1991 $1.58 1992 $1.78 1993 $1.98 1994 $2.30
a. Estimate the arithmetic average and geometric average growth rate in earnings per share between 1989 and 1994. Why are they different? Which is more reliable? b. Estimate the growth rate using a linear growth model. c. Estimate the growth rate using a log-linear growth model. 2. BIC Corporation reported a return on equity of 20% and paid out 37% of its earnings as dividends in the most recent year. 218
a. Assuming that these fundamentals do not change, estimate the expected growth rate in earnings per share. b. Now assume that you expect the return on equity to increase to 25% on both new and existing investments next year. Estimate the expected growth rate in earnings per share. 3. You are trying to estimate the expected growth in net income at Metallica Corporation, a manufacturing firm that reported $150 million in net income in the just-completed financial year; the book value of equity at the beginning of the year was $1 billion. The firm had capital expenditures of $160 million, depreciation of $100 million, and an increase in working capital of $40 million during the year. The debt outstanding increased by $40 million during the year. Estimate the equity reinvestment rate and expected growth in net income. 4. You are trying to estimate a growth rate for HipHop Inc., a record producer and distributor. The firm earned $100 million in after-tax operating income on capital invested of $800 million last year. In addition, the firm reported net capital expenditures of $25 million and an increase in noncash working capital of $15 million. a. Assuming that the firm's return on capital and reinvestment rate remain unchanged, estimate the expected growth in operating income next year. b. How would your answer to (a) change if you were told that the firm's return on capital next year will increase by 2.5%? (Next year's return on capital = This year's return on capital + 2.5%.) 5. InVideo Inc. is an online retailer of videos and DVDs. The firm reported an operating loss of $10 million on revenues of $100 million in the most recent financial year. You expect revenue growth to be 100% next year, 75% in year 2, 50% in year 3, and 30% in years 4 and 5. You also expect the pretax operating margin to improve to 8% of revenues by year 5. Estimate the expected revenues and operating income (or loss) each year for the next five years. 6. SoftTech Inc. is a small manufacturer of entertainment software that reported revenues of $25 million in the most recent financial year. You expect the firm to grow significantly over time and capture 8% of the overall entertainment software market in 10 years. If the total revenues from entertainment software in the most recent year amounted to $2 billion and you expect an annual growth rate of 6% in these revenues for the next 10 years, estimate the compounded annual revenue growth rate at SoftTech for the next 10 years. An ordinary least squares (OLS) regression estimates regression coefficients by minimizing the squared differences of predicted values from actual values. 1
Griffin, P.A., “The Time-Series Behavior of Quarterly Earnings: Preliminary Evidence,” Journal of Accounting Research 15 (Spring 1977): 71–83. 2
3
Watts, R.L., “The Time-Series Behavior of Quarterly Earnings,” Working Paper, University of Newcastle, 1975.
Time series models generally can be run as long as there are at lea st 30 observations, but the estimation error declines as the number of observations increases. 4
Sell-side analysts work for brokerage houses and investment banks, and their research is offered to clients of these firms as a service. In contrast, buy-side analysts work for institutional investors, and their research is generally proprietary. 5
6
7
Note that 17.22% is halfway between the current return on capital and the industry average (22.27 percent).
You are allowing for a compounded growth rate over time. Th us, if earnings are expected to grow 25 percent over three years, you estimate the expected growth rate each year to be: Expected growth rate each year = (1.25)1/3 – 1. 8
The margin each year is computed as follows for Linkedin: (Margin this year + Target margin)/2. For Tesla Motors, the margin each year is computed as: (Margin this year – Target margin)/3. 9
219
CHAPTER 12 Closure in Valuation: Estimating Terminal Value In the previous chap ter, we examined the determinants of expected growth. Firms that reinvest substantial portions of their earnings and earn high returns on these investments should be able to grow at high rates. But for how long? And what happens after that? This chapter looks at two ways of bringing closure to a valuation: a going concern approach, where we assume that the firm continues to deliver cash flows in perpetuity and a liquidation approach, where we assume that the business is shut down and the assets are sold at some point in time. Consider the going concern approach first. As a firm grows, it becomes more difficult for it to maintain high growth and it eventually will grow at a rate less than or equal to the growth rate of the economy in which it operates. This growth rate, labeled stable growth, can be sustained in perpetuity, allowing us to estimate the value of all cash flows beyond that point as a terminal value for a going concern. The key question that we confront is the estimation of when and how this transition to stable growth will occur for the firm that we are valuing. Will the growth rate drop abruptly at a point in time to a stable growth rate or will it occur more gradually over time? To answer these questions, we will look at a firm's size (relative to the market that it serves), its current growth rate, and its competitive advantages. We also consider an alternate route, which is that firms do not last forever and that they will be liquidated at some point in the future. We will consider how best to estimate liquidation value and when it makes more sense to use this approach rather than the going concern approach.
CLOSURE IN VALUATION Since you cannot estimate cash flows forever, you generally impose closure in discounted cash flow valuation by stopping your estimation of cash flows sometime in the future and then computing a terminal value that reflects the value of the firm at that point.
You can find the terminal value in one of three ways. One is to assume a liquidation of the firm's assets in the terminal year and estimate what others would pay for the assets that the firm has accumulated at that point. The other two approaches value the firm as a going concern at the time of the terminal value estimation. One applies a multiple to earnings, revenues, or book value to estimate the value in the terminal year. The other assumes that the cash flows of the firm will grow at a constant rate forever—a stable growth rate. With stable growt h, the terminal value can be estimated using a perpetual growth model.
Liquidation Value In some valuations, we can assume that the firm will cease operations at a point in time in the future and sell the assets it has accumulated to the highest bidders. The estimate that emerges is called a liquidation value. There are two ways in which the liquidation value can be estimated. One is to base it on the book value of the assets, adjusted for any inflation during the period. Thus, if the book value of assets 10 years from now is expected to be $2 billion, the average age of the assets at that point is five years and the expected inflation rate is 3 percent, the expected liquidation value can be estimated as:
The limitation of this approach is that it is based on accounting book value and does not reflect the earning power of the assets. The alternative approach is to estimate the value based on the earning power of the assets. To make this estimate, we would first have to estimate the expected cash flows from the assets and then discount these cash flows back to the present, using an appropriate discount rate. In the preceding example, for instance, if we assumed that the assets in question could be expected to generate $400 million in after-tax cash flows for 15 years (after the terminal year) and the cost of capital was 10 percent, our estimate of the expected liquidation value would be:
When valuing equity, there is one additional step that needs to be taken. The estimated value of debt outstanding in the terminal year has to be subtracted from the liquidation value to arrive at the liquidation proceeds for equity 220
investors.
Multiple Approach In this approach, the value of a firm in a future year is estimated by applying a multiple to the firm's earnings or revenues in that year. For instance, a firm with expected revenues of $6 billion, 10 years from now will have an estimated terminal value in that year of $12 billion, if a value-to-sales multiple of 2 is used. If valuing equity, we use equity multiples such as price-earnings ratios to arrive at the terminal value. Although this approach has the virtue of simplicity, the multiple determines the final value and where it is obtained can be critical. If, as is common, the multiple is estimated by looking at how comparable firms in the business today are priced by the market, the valuation becomes a relative valuation, rather than a discounted cash flow valuation. If the multiple is estimated using fundamentals, it converges on the stable growth model that is described in the next section. All in all, using multiples to estimate terminal value, when those multiples are estimated from comparable firms, results in a dangerous mix of relative and discounted cash flow valuation. While there are advantages to relative valuation, and we consider these in a later chapter, a discounted cash flow valuation should provide you with an estimate of intrinsic value, not relative value. Consequently, the only consistent way of estimating terminal value in a discounted cash flow model is to use either a liquidation value or a stable growth model.
Stable Growth Model In the liquidation value approach, you are assuming that your firm has a finite life and that it will be liquidated at the end of that life. Firms, however, can reinvest some of their cash flows back into new assets and extend their lives. If you assume that cash flows, beyond the terminal year, will grow at a constant rate forever, the terminal value can be estimated as follows:
The cash flow and the discount rate used will depend on whether you are valuing the firm or valuing equity. If you are valuing equity, the terminal value of equity can be written as:
The cash flow to equity can be defined strictly as dividends (in the dividend discount model) or as free cash flow to equity. If valuing a firm, the terminal value can be written as:
where the cost of capital and the growth rate in the model are sustainable forever. In this section, we will begin by considering how high a stable growth rate can be, how to best estimate when your firm will be a stable growth firm, and what inputs need to be adjusted as a firm approaches stable growth.
Constraints on Stable Growth Of all the inputs into a discounted cash flow valuation model, none creates as much angst as estimating the stable growth rate. Part of the reason for it is that small changes in the stable growth rate can change the terminal value significantly, and the effect gets larger as the growth rate approaches the discount rate used in the estimation. The fact that a stable growth rate is constant forever, however, puts strong constraints on how high it can be. Since no firm can grow forever at a rate higher than the growth rate of the economy in which it operates, the constant growth rate cannot be greater than the overall growth rate of the economy. In making a judgment on what the limits on a stable growth rate are, we have to consider the following three questions: 1. Is the company constrained to operate as a domestic company, or does it operate (or have the capacity to operate) multinationally? If a firm is a purely domestic company, either because of internal constraints (such as those imposed by management) or external constraints (such as those imposed by a government), the growth rate in the domestic economy will be the limiting value. If the company is a multinational or has aspirations to be one, the growth rate in the global economy (or at least those parts of the globe that the firm operates in) will be the limiting value. 2. Is the valuation being done in nominal or real terms? If the valuation is a nominal valuation, the stable growth rate should also be a nominal growth rate (i.e., include an expected inflation component). If the valuation is a real valuation, the stable growth rate will be constrained to be lower. Using a U.S. company in 2011 as an example, the stable growth rate can be as high as 2.0 percent if the valuation is done in nominal U.S. dollars but only 1 percent if the valuation is done in real terms. 221
3. What currency is being used to estimate cash flows and discount rates in the valuation? The limits on stable growth will vary depending on what currency is used in the valuation. If a high-inflation currency is used to estimate cash flows and discount rates, the stable growth rate will be much higher, since the expected inflation rate is added on to real growth. If a low-inflation currency is used to estimate cash flows, the stable growth rate will be much lower. For instance, the stable growth rate that would be used to value Cemex, the Mexican cement company, will be much higher if the valuation is done in Mexican pesos than in U.S. dollars. Although the stable growth rate cannot exceed the growth rate of the economy in which a firm operates, it can be lower. There is nothing that prevents us from assuming that mature firms will become a smaller part of the economy and it may, in fact, be the more reasonable assumption to make. Note that the growth rate of an economy reflects the contributions of both young, higher-growth firms and mature, stable-growth firms. If the former grow at a rate much higher than the growth rate of the economy, the latter have to grow at a rate that is lower. Setting the stable growth rate to be less than or equal to the growth rate of the economy not only is the consistent thing to do but it also ensures that the growth rate will be less than the discount rate. This is because there is a link between the riskless rate that goes into the discount rate and the growth rate of the economy. Note that the riskless rate can be written as:
In the long term, the real riskless rate will converge on the real growth rate of the economy, and the nominal riskless rate will approach the nominal growth rate of the economy. In fact, a simple rule of thumb on the stable growth rate is that it generally should not exceed the riskless rate used in the valuation.
CAN THE STABLE GROWTH RATE BE NEGATIVE? The previous section noted that the stable growth rate has to be less than or equal to the growth rate of the economy. But can it be negative? There is no reason why not since the terminal value can still be estimated. For instance, a firm with $100 million in after-tax cash flows growing at –5% a year forever and a cost of capital of 10 percent has a value of:
Intuitively, though, what does a negative growth rate imply? It essentially allows a firm to partially liquidate itself each year until it just about disappears. Thus, it is an intermediate choice between complete liquidation and the going concern that gets larger each year forever. This may be the right choice to make when valuing firms in industries that are being phased out because of technological advances (such as the manufacturers of landline phones, with the advent of the cellphones) or where an external and critical customer is scaling back purchases for the long term (as was the case with defense contractors after the end of the cold war).
Key Assumptions about Stable Growth In every discounted cash flow valuation, there are three critical assumptions you need to make on stable growth. The first relates to when the firm that you are valuing will become a stable growth firm, if it is not one already. The second relates to what the characteristics of the firm will be in stable growth, in terms of return on investments and costs of equity and capital. The final assumption relates to how the firm that you are valuing will make the transition from high growth to stable growth.
Length of the High Growth Period The question of how long a firm will be able to sustain high growth is perhaps one of the more difficult questions to answer in a valuation, but two points are worth making. One is that it is not a question of whether but when firms hit the stable growth wall. All firms ultimately become stable growth firms, in the best case, because high growth makes a firm larger, and the firm's size will eventually become a barrier to further high growth. In the worst-case scenario, firms do not survive and will be liquidated. The second is that high growth in valuation, or at least high growth that creates value,1 comes from firms earning excess returns on their marginal investments. In other words, increased value comes from firms having a return on capital that is higher than the cost of capital (or a return on equity that exceeds the cost of equity). Thus, when you assume that a firm will experience high growth for the next 5 or 10 years, you are also implicitly assuming that it will earn excess returns (over and above the required return) during that period. In a competitive market, these excess returns will eventually draw in new competitors, and the excess returns will disappear. You should look at three factors when considering how long a firm will be able to maintain high growth. 1. Size of the firm. Smaller firms are much more likely to earn excess returns and maintain these excess returns than otherwise similar larger firms. This is because they have mor e room to grow and a larger potential market. Small firms in large markets should have the potential for high growth (at least in revenues) over long periods. When looking at the size of the firm, you should look not only at its current market share, but also at the potential growth in the total market for its products or services. A firm may have a large market share of its current market, but it may be able to grow in spite of this because the entire market is growing rapidly. 222
2. Existing growth rate and excess returns. Momentum does matter, when it comes to projecting growth. Firms that have been reporting rapidly growing revenues are more likely to see revenues grow rapidly at least in the near future. Firms that are earning high returns on capital and high excess returns in the current period are likely to sustain these excess returns for the next few years. 3. Magnitude and sustainability of competitive advantages. This is perhaps the most critical determinant of the length of the high growth period. If there are significant barriers to entry and sustainable competitive advantages, firms can maintain high growth for longer periods. If, on the other hand, there are no or minor barriers to entry, or if the firm's existing competitive advantages are fading, you should be far more conservative about allowing for long growth periods. The quality of existing management also influences growth. Some top managers have the capacity to make the strategic choices that increase competitive advantages and create new ones.2
COMPETITIVE ADVANTAGE PERIOD (CAP) The confluence of high growth and excess returns that is the source of value has led to the coining of the term competitive advantage period (CAP) to capture the joint effect. This term, popularized by Michael Mauboussin at Credit Suisse First Boston, measures the period during which a firm can be expected to earn excess returns. The value of such a firm can then be written as the sum of the capital invested today and the present value of the excess returns that the firm will earn over its life. Since there are no excess returns after the competitive advantage period, there is no additional value added. In an inventive variant, analysts sometimes try to estimate how long the competitive advantage period will have to be to sustai n a current market value, assuming that the current return on capital and cost of capital remain unchanged. The resulting market-implied competitive advantage period (MICAP) can then be either compared across firms in a sector or evaluated on a qualitative basis.
ILLUSTRATION 12.1: Length of High Growth Period To illustrate the process of estimating the length of the high growth period, we will consider a number of companies and make subjective judgments about how long each one will be able to maintain high growth. CONSOLIDATED EDISON
Background: The firm has a near-monopoly in generating and selling power in the environs of New York City. In return for the monopoly, though, the firm is restricted in both its investment and its pricing policy. A regulatory commission determines how much Con Ed can raise prices, and it makes this decision based on the returns made by Con Ed on its investments; if the firm is making high returns on its investments, it is unlikely to be allowed to increase prices. Finally, the demand for power in New York is stable, as the population levels off. Implication: The firm is already a stable growth firm. There is little potential for either high growth or excess returns. PROCTER & GAMBLE
Background: Procter & Gamble comes in with some obvious strengths. Its valuable brand names have allowed it to earn high excess returns (as manifested in its high return on equity of 20.09% in 2010) and sustain high growth rates in earnings over the past few decades. The firm faces two challenges. One is that it has a significant market share in a mature market in the United States, and its brand names are less recognized and therefore less likely to command premiums abroad. The other is the increasing assault on brand names in general by generic manufacturers. Implication: Brand name can sustain excess returns and growth higher than the stable growth rate for a short period—we will assume five years. Beyond that, we will assume that the firm will be in stable growth albeit with some residual excess returns. If the firm is able to extend its brand names overseas, its potential for high growth will be higher. AMGEN
Background: Amgen has a stable of drugs, on which it has patent protection, that generate cash flows currently, and several drugs in its R&D pipeline. While it is the largest independent biotechnology firm in the world, the market for biotechnology products is expanding significantly and will continue to do so. Finally, Amgen has had a track record of delivering solid earnings growth. Implication: The patents that Amgen has will protect it from competition, and the long lead time to drug approval will ensure that new products will take a while getting to the market. We will allow for 10 years of growth and excess returns. There is clearly a strong subjective component to making a judgment on how long high growth will last. Much of what was said about the interrelationships between qualitative variables and growth toward the end of Chapter 11 has relevance for this discussion as well.
Characteristics of Stable Growth Firm As firms move from high growth to stable growth, you need to give them the characteristics of stable growth firms. A firm in stable growth is different from that same firm in high growth on a number of dimensions. In general, you would expect stable growth firms to have average risk, use more debt, have lower (or no) excess returns, and reinvest less than high growth firms. In this section, we will consider how best to adjust each of these variables.
Equity Risk When looking at the cost of equity, high growth firms tend to be more exposed to market risk (and have higher betas) than stable growth firms. Part of the reason for this is that they tend to be niche players supplying discretionary products, and part of the reason is high operating leverage. Thus, young technology or social media firms will have high betas. As these firms mature, you would expect them to have less exposure to market risk and 223
betas that are closer to 1—the average for the market. One option is to set the beta in stable growth to 1 for all firms, arguing that firms in stable growth should all be average risk. Another is to allow for small differences to persist even in stable growth, with firms in more volatile businesses having higher betas than firms in stable businesses. We would recommend that, as a rule of thumb, stable period betas not exceed 1.2.3 But what about firms that have betas well below 1, such as commodity companies? If you are assuming that these firms will stay in their existing businesses, there is no harm in assuming that the beta remains at existing levels. However, if your estimates of growth in perpetuity will require them to branch out into other businesses, you should adjust the beta upward toward 1; invoking another rule of thumb, stable period betas should not be lower than 8.4
betas.xls: This dataset on the Web summarizes the average lever ed and unlevered betas, by industry group, for firms in the United States.
Project Returns High growth firms tend to have high returns on capital (and equity) and earn excess returns. In stable growth, it becomes much more difficult to sustain excess returns. There are some who believe that the only assumption consistent with stable growth is to assume no excess returns; the return on capital is set equal to the cost of cap ital. While, in principle, excess returns in perpetuity may not seem reasonable, it is difficult in practice to assume that firms will suddenly lose the capacity to earn excess returns at a point in time (say 5 years or 10 years). To provide a simple example, consider Proctor and Gamble, a company that we estimated a high growth period of 5 years for in illustration 12.1. While the growth rate for P&G may drop to a stable level by year 6, the strong brand name and other competitive advantages are likely to persist for much longer (say 30 to 40 years). Rather than estimate cash flows for 30 to 40 years, we would stop estimating cash flows in year 5 but still allow the company to continue earning more than its cost of capital in perpetuity. Since entire industries often earn excess returns over long periods, assuming a firm's returns on equity and capital will move toward industry averages will yield more reasonable estimates of value.
eva.xls: This dataset on the Web summarizes the returns on capital (equity), costs of capital (equity), and excess returns, by industry group, for firms in the United States.
Debt Ratios and Costs of Debt High growth firms tend to use less debt than stable growth firms. As firms mature, their debt capacity increases. When valuing firms, this will change the debt ratio that we use to compute the cost of capital. When valuing equity, changing the debt ratio will change both the cost of equity and the expected cash flows. The question of whether the debt ratio for a firm should be moved toward a more sustainable level in stable growth cannot be answered without looking at the incumbent managers' views on debt, and how much power stockholders have in these firms. If managers are willing to change their financing policy, and stockholders retain some power, it is reasonable to assume that the debt ratio will move to a higher level in stable growth; if not, it is safer to leave the debt ratio at existing levels. As earnings and cash flows increase, the perceived default risk in the firm will also change. A firm that is currently losing $10 million on revenues of $100 million may be rated B, but its rating should be much better if your forecasts of $10 billion in revenues and $1 billion in operating income come to fruition. In fact, internal consistency requires that you reestimate the rating and the cost of debt for a firm as you change its revenues and operating income. As a general rule, stable growth firms should have at least investment grade ratings (Baa or higher). On the practical question of what debt ratio and cost of debt to use in stable growth, you should look at the financial leverage of larger and more mature firms in the industry. One solution is to use the industry average debt ratio and cost of debt as the debt ratio and cost of debt for the firm in stable growth.
wacc.xls: This dataset on the Web summarizes the debt ratios and costs of debt, by industry group, for firms in the United States.
Reinvestment and Retention Ratios 224
Stable growth firms tend to reinvest less than high growth firms, and it is critical that we capture the effects of lower growth on reinvestment and that we ensure that the firm reinvests enough to sustain its stable growth rate in the terminal phase. The actual adjustment will vary depending on whether we are discounting dividends, free cash flows to equity, or free cash flows to the firm. In the dividend discount model, note that the expected growth rate in earnings per share can be written as a function of the retention ratio and the return on equity.
Algebraic manipulation can allow us to state the retention ratio as a function of the expected growth rate and return on equity:
If we assume, for instance, a stable growth rate of 3 percent (based on the growth rate of the economy) for Procter & Gamble (P&G) and a return on equity of 12 percent (based on industry averages), we would be able to compute the retention ratio of the firm in stable growth:
Procter & Gamble will have to retain 25 percent of its earnings to generate its expected growth of 3 percent; it can pay out the remaining 75 percent. In a free cash flow to equity model, where we are focusing on net income growth, the expected growth rate is a function of the equity reinvestment rate and the return on equity:
The equity reinvestment rate can then be computed as follows:
If, for instance, we assume that Coca-Cola will have a stable growth rate of 3 percent and have a return on equity in stable growth of 15 percent, we can estimate an equity reinvestment rate of 20%; the remaining 80% can be paid out as cash flows to equity investors:
Finally, looking at free cash flows to the firm, we estimated the expected growth in operating income as a function of the return on capital (ROC) in stable growth and the reinvestment rate:
Again, algebraic manipulation yields the following measure of the reinvestment rate in stable growth:
where ROCn is the return on capital that the firm can sustain in stable growth. This reinvestment rate can then be used to generate the free cash flow to the firm in the first year of stable growth. Linking the reinvestment rate and retention ratio to the stable growth rate also makes the valuation less sensitive to assumptions about the stable growth rate. Whereas increasing the stable growth rate, holding all else constant, can dramatically increase value, changing the reinvestment rate as the growth rate changes will create an offsetting effect. The gains from increasing the growth rate will be partially or completely offset by the loss in cash flows because of the higher reinvestment rate. Whether value increases or decreases as the stable growth increases will entirely depend on what you assume about excess returns. If the return on capital is higher than the cost of capital in the stable growth period, increasing the stable growth rate will increase value. If the return on capital is equal to the stable growth rate, increasing the stable growth rate will have no effect on value. This can be proved quite easily:
Substituting in the stable growth rate as a function of the reinvestment rate, from the equation, you get:
225
Setting the return on capital equal to the cost of capital, you arrive at:
Simplifying, the terminal value can be stated as:
Put simply, when there are no excess returns, your terminal value is unaffected by your assumptions about expected growth. You could establish the same proposition with equity income and cash flows, and show that a return on equity equal to the cost of equity in stable growth nullifies the positive effect of growth.
divfund.xls: This dataset on the Web summarizes retention ratios, by industry group, for firms in the United States.
capex.xls: This dataset on the Web summarizes the reinvestment rates, by industry group, for firms in the United States.
ILLUSTRATION 12.2: Stable Growth Rates and Excess Returns Alloy Mills is a textile firm that is currently reporting after-tax operating income of $100 million. The firm has a return on capital currently of 20% and reinvests 50% of its earnings back into the firm, giving it an expected growth rate of 10% for the next five years:
After year 5 the growth rate is expected to drop to 5% and the return on capital is expected to stay at 20%. The terminal value can be estimated as follows:
The value of the firm today would then be:
If we did change the return on capital in stable growth to 10% while keeping the growth rate at 5%, the effect on value would be dramatic:
Now consider the effect of lowering the growth rate to 4% while keeping the return on capital at 10% in stable growth:
Note that the terminal value decreases by $16 million but the cash flow in year 5 also increases by $16 million because the reinvestment rate at the end of year 5 drops to 40%. The value of the firm remains unchanged at $1,300 million. In fact, changing the stable growth rate to 0% has no effect on value:
226
ILLUSTRATION 12.3: Stable Growth Inputs To illustrate how the inputs to valuation change as we go from high growth to stable growth, we will consider three firms—Procter & Gamble, with the dividend discount model; Coca-Cola, with a free cash flow to equity model; and Amgen with a free cash flow to firm model. Consider Procter & Gamble first in the context of the dividend discount model. Although we do the valuation in the next chapter, note that there are three key inputs to the dividend discount model—the payout ratio (which determines dividends), the expected return on equity (which determines the expected growth rate), and the beta (which affects the cost of equity). In Illustration 12.1, we argued that Procter & Gamble would have a five-year high-growth period. The following table summarizes the inputs into the dividend discount model for the valuation of Procter & Gamble. High Growth Stable Growth Payout ratio
50.00%
75.00%
Return on equity
20.00%
12.00%
Expected growth rate 10%
3.00%
Beta
1.00
0.90
Note that the payout ratio, return on equity, and beta for the high growth period are based on the current year's values. The expected growth rate of 10% for the next five years is the product of the return on equity and retention ratio. In stable growth, we adjust the beta to 1, though the adjustment has little effect on value since the beta is already close to one. We assume that the stable growth rate will be 3%, just slightly below the nominal growth rate in the global economy (and the risk-free rate of 3.5% at the time). We also assume that the return on equity will drop to 12%, reflecting our assumption that returns on equity will decline for the entire industry as competition from generics eats into profit margins. The retention ratio decreases to 25%, as both growth and return on equity drop. To analyze Coca-Cola in a free cash flow to equity model, the following table summarizes our inputs for high growth and stable growth:
High Growth Stable Growth Return on equity
30.00%
15.00%
Equity reinvestment rate 25.00%
20.00%
Expected growth rate
7.50%
3.00%
Beta
0.80
0.80
In high growth, the high return on equity allows the firm to generate an expected growth rate of 7.50% a year. In stable growth, we reduce the return on equity for Coca-Cola to the industry average for beverage companies and estimate the expected equity reinvestment rate based on a stable growth rate of 3%. The beta for the firm is left unchanged at its existing level, since Coca-Cola's management has been fairly disciplined in staying focused on the core businesses. Finally, let us consider a valuation of Amgen beginning in early 2010. The following table reports on the return on capital, reinvestment rate, and debt ratio for the firm in high growth and stable growth periods.
High Growth Stable Growth Return on capital
17.41%
10.00%
Reinvestment rate
33.23%
30.00%
Expected growth rate 5.78%
3.00%
Beta
1.10
1.65
Note that the reinvestment rate and return on capital for the firm reflect the decision we made to capitalize R&D and operating leases. The operating income is adjusted for R&D and the book value of equity is augmented by the capitalized value of R&D (see Chapter 9). The firm has a high return on capital currently, and we assume that this return will decrease in stable growth to 10% as the firm becomes larger and patents expire. Since the stable growth rate drops to 3%, the resulting reinvestment rate at Amgen will decrease to 30%. We also assume that the beta for Amgen will converge toward the market average. For all of the firms, it is worth noting that we are assuming that excess returns continue in perpetuity by setting the return on capital above the cost of capital. While this is potentially troublesome, the competitive advantages that these firms have built up historically or will build up over the high growth phase will not disappear in an instant. The excess returns will fade over time, but moving them to or toward the cost of capital in stable growth seems like a reasonable compromise.
Transition to Stable Growth Once you have decided that a firm will be in stable growth at a point in time in the future, you have to consider how the firm will change as it approaches stable growth. There are three distinct scenarios. In the first, the firm will maintain its high growth rate for a period of time and then become a stable growth firm abruptly; this is a twostage model. In the second, the firm will maintain its high growth rate for a period and then have a transition period when its characteristics change gradually toward stable growth levels; this is a three-stage model. In the third, the firm's characteristics change each year from the initial period to the stable growth period; this can be considered an 227
n-stage model. Which of these three scenarios gets chosen depends on the firm being valued. Since the firm goes from high growth to stable growth in one year in the two-stage model, this model is more appropriate for firms with moderate growth rates, where the shift will not be too dramatic. For firms with very high growth rates in operating income, a transition phase allows for a gradual adjustment not just of growth rates but also of risk characteristics, returns on capital and reinvestment rates toward stable growth levels. For very young firms or for firms with negative operating margins, allowing for changes in each year (in an n-stage model) is prudent.
ILLUSTRATION 12.4: Choosing a Growth Pattern Consider the three firms analyzed in Illustration 12.3. We assumed a growth rate of 10% and a high-growth period of five years for P&G, a growth rate of 7.5% and a high growth period of 10years for Coca-Cola, and a growth rate of 5.78% and a high-growth period of 10 years for Amgen. For Procter & Gamble, we will use a two-stage model—growth of 10% for five years and 3% thereafter. For both Coca-Cola and Amgen, we will allow for a transition phase between years 6 and 10 in which the inputs will change gradually from high growth to stable growth levels. Figure 12.1 reports on how the equity reinvestment rate and expected growth change at Coca-Cola from years 6 through 10, as well as the change in the expected growth rate and reinvestment rate at Amgen over the same period.
Figure 12.1 Transition Period Estimates: Coca-Cola and Amgen
EXTRAORDINARY GROWTH PERIODS WITHOUT A HIGH GROWTH RATE OR A NEGATIVE GROWTH RATE Can you have extraordinary growth periods for firms that have expected growth rates that are less than or equal to the growth rate of the economy? The answer is yes, for some firms. This is because stable growth requires not just that the growth r ate be less than the growth rate of the economy, but that the other inputs into the valuation are also appropriate for a stable growth firm. Consider, for instance, a firm whose operating income is growing at 2 percent a year but whose current return on capital is 20 percent and whose beta is 1.5. You would still need a transition period in which the return on capital d eclined to more sustainable levels (say 12 percent) and the beta moved toward 1. By the same token, you can have an extraordinary growth period, where the growth rate is less than the stable growth rate and then moves up to the stable growth rate. For instance, you could have a firm that is expected to see its earnings decline 5 percent a year for the next five years (which would be the extraordinary growth period) and grow 2 percent thereafter.
THE SURVIVAL ISSUE Implicit in the use of a terminal value in discounted cash flow valuation is the assumption that the value of a firm comes from it being a going concern with a perpetual life. For many risky firms, there is the very real possibility that they might not be in existence in 5 or 10 years, with volatile earnings and shifting technology. Should the valuation reflect this chance of failure, and, if so, how can the likelihood that a firm will not survive be built into a valuation?
Life Cycle and Firm Survival There is a link between where a firm is in the life cycle and survival. Young firms with negative earnings and cash flows can run into serious cash flow problems and end up being acquired by firms with more resources at bargain basem*nt prices. Why are young firms more exposed to this problem? The negative cash flows from operations, when combined with significan t reinvestment needs, can result in a rapid depletion of cash reserves. When financial markets are accessible and additional equity (or debt) can be raised at will, raising more funds to meet these funding needs is not a problem. However, when stock prices drop and access to markets becomes more limited, these firms can be in trouble. A widely used measure of the potential for a cash flow problem for firms with negative earnings is the cash burn 228
ratio, which is estimated as the cash balance of the firm divided by its earnings before interest, taxes, depreciation, and amortization (EBITDA).
where EBITDA is a negative number and the absolute value of EBITDA is used to estimate this ratio. Thus a firm with a cash balance of $1 billion and EBITDA of –$1.5 billion will burn through its cash balance in eight months.
Likelihood of Failure and Valuation One view of survival is that the expected cash flows that you use in a valuation reflect cash flows under a wide range of scenarios from very good to abysmal and the probabilities of the scenarios occurring. Thus, the expected value already has built into it the likelihood that the firm will not survive. Any market risk associated with survival or failure is assumed to be incorporated into the cost of capital. Firms with a high likelihood of failure will therefore have higher discount rates and lower expected cash flows. Another view of survival is that discounted cash flow valuations tend to have an optimistic bias and that the likelihood that the firm will not survive is not considered adequately in the value. With this view, the discounted cash flow value that emerges from the analysis in the prior section overstates the value of operating assets and has to be adjusted to reflect the likelihood that the firm will not survive to deliver its terminal value or even the positive cash flows that you have forecast in future years.
Should You or Should You Not Discount Value for Survival? For firms that have substantial assets in place and relatively small probabilities of distress, the first view is the more appropriate one. Attaching an extra discount for nonsurvival is double counting risk. For younger and smaller firms, it is a tougher call and depends on whether expected cash flows incorporate the probability that these firms may not make it past the first few years. If they do, the valuation already reflects the likelihood that the firms will not survive past the first few years. If they do not, you do have to discount the value for the likelihood that the firm will not survive the near future. One way to estimate this discount is to estimate a probability of failure, and adjust the operating asset value for this probability:
For a firm with a discounted cash flow value of $1 billion on its assets, a distress sale value of $500 million and a 20 percent probability of distress, the adjusted value would be $900 million:
There are two points worth noting here. It is not the failure to survive per se that causes the loss of value but the fact that the distressed sale value is at a discount on the fair value. The second is that this approach requires estimating the probability of failure. This probability is difficult to estimate because it will depend upon both the magnitude of the cash reserves of the firm (relative to its cash needs) and the state of the market. In buoyant equity markets, even firms with little or no cash can survive because they can access markets for more funds. Under more negative market conditions, even firms with significant cash balances may find themselves under threat.
ESTIMATING THE PROBABILITY OF DISTRESS There are two ways in which we can estimate the probability that a firm will not survive. One is to draw on the past, look at firms that have failed, compare them to firms that did not, and look for variables that seem to set them apart. For instance, firms with high debt ratios and negative cash flows from operations may be more likely to fail than firms without these characteristics. In fact, you can use statistical techniques such as probits to estimate the probability that a firm will fail. To run a probit, you would begin, for instance, with all listed firms in 1990 and their financial characteristics, identify the firms that failed during the 1991–1999 time period and then estimate the probability of failure as a function of variables that were observable in 1990. The output, which resembles regression output, will then let you estimate the probability of default for any firm today. The other way of estimating the probability of default is to use the bond rating for the firm, if it is available. For instance, assume that Tesla Motors has a B rating. An empirical examination of B-rated bonds over the past decade reveals that the likelihood of default with this rating is 36.80 percent.5 While this approach is simpler, it is limiting insofar as it can be used only for rated firms, and it assumes that the standards used by ratings agencies have not changed significantly over time.
CLOSING THOUGHTS ON TERMINAL VALUE The role played by the terminal value in discounted cash flow valuations has often been the source of much of the criticism of the discounted cash flow approach. Critics of the approach argue that too great a proportion of the discounted cash flow value comes from the terminal value and that it is easy to manipulate the terminal value to yield any number you want. They are wrong on both counts. 229
It is true that a large portion of the value of any stock or equity in a business comes from the terminal value, but it would be surprising if it were not so. When you buy a stock or invest in the equity in a business, consider how you get your returns. Assuming that your investment is a good investment, the bulk of the returns come not while you hold the equity (from dividends or other cash flows) but when you sell it (from price appreciation). The terminal value is designed to capture the latter. Consequently, the greater the growth potential in a business, the higher the proportion of the value that comes from the terminal value. Is it easy to manipulate the terminal value? We concede that terminal value is manipulated often and easily, but it is because analysts either use multiples to get these values or because they violate one or both of two basic propositions in stable growth models. One is that the growth rate cannot exceed the growth rate of the economy. The other is that firms have to reinvest enough in stable growth to generate the growth rate. In fact, as we showed earlier in the chapter, it is not the stable growth rate that drives value as much as what we assume about excess returns in perpetuity. When excess returns are zero, changes in the stable growth rate have no impact on value.
CONCLUSION The value of a firm is the present value of its expected cash flows over its life. Since firms have infinite lives, you apply closure to a valuation by estimating cash flows for a period and then estimating a value for the firm at the end of the period—a terminal value. Many analysts estimate the terminal value using a multiple of earnings or revenues in the final estimation year. If you assume that firms have infinite lives, an approach that is more consistent with discounted cash flow valuation is to assume that the cash flows of the firm will grow at a constant rate forever beyond a point in time. When the firm that you are valuing will approach this growth rate, which you label a stable growth rate, is a key part of any discounted cash flow valuation. Small firms that are growing fast and have significant competitive advantages should be able to grow at high rates for much longer periods than larger and more mature firms, without these competitive advantages. If you do not want to assume an infinite life for a firm, you can estimate a liqui dation value based on what others will pay for the assets that the firm has accumulated during the high-growth phase.
QUESTIONS AND SHORT PROBLEMS In the problems following, use an equity risk premium of 5.5 percent if none is specified. 1. Ulysses Inc. is a shipping company with $100 million in earnings before interest and taxes that is expected to have earnings growth of 10% for the next five years. At the end of the fifth year, you estimate the terminal value using a multiple of 8 times operating income (which is the average for the sector). a. Estimate the terminal value of the firm. b. If the cost of capital for Ulysses is 10%, the tax rate is 40%, and you expect the stable growth rate to be 5%, what is the return on capital that you are assuming in perpetuity if you use a multiple of 8 times operating in come? 2. Genoa Pasta manufactures Italian food products and currently earns $80 million in earnings before interest and taxes. You expect the firm's earnings to grow 20 percent a year for the next six years and 5% thereafter. The firm's current after-tax return on capital is 28%, but you expect it to be halved after the sixth year. If the cost of capital for the firm is expected to be 10% in perpetuity, estimate the terminal value for the firm. (The tax rate for the firm is 40%.) 3. Lamps Galore Inc. manufactures table lamps and earns an after-tax return on capital of 15% on its current capital invested (which is $100 million). You expect the firm to reinvest 80% of its after-tax operating income back into the business for the next four years and 30% thereafter (the stable growth period). The cost of capital for the firm is 9%. a. Estimate the terminal value for the firm (at the end of the fourth year). b. If you expect the after-tax return on capital to drop to 9% after the fourth year, what would your estimate of terminal value be? 4. Bevan Real Estate Inc. is a real estate holding company with four properties. You estimate that the income from these properties, which is currently $50 million after taxes, will grow 8% a year for the next 10 years and 3% thereafter. The current market value of the properties is $500 million, and you expect this value to appreciate at 3% a year for the next 10 years. a. Estimate the terminal value of the properties, based on the current market value and the expected appreciation rate in property values. b. Assuming that your projections of income growth are right, what is the terminal value as a multiple of after-tax 230
operating income in the tenth year? c. If you assume that no reinvestment is needed after the tenth year, estimate the cost of capital that you are implicitly assuming with your estimate of the terminal value. 5. Latin Beats Corporation is a firm that specializes in Spanish music and videos. In the current year, the firm reported $20 million in after-tax operating income, $15 million in capital expenditures, and $5 million in depreciation. The firm expects all three items to grow at 10% for the next five years. Beyond the fifth year, the firm expects to be in stable growth and grow at 4% a year in perpetuity. You assume that earnings, capital expenditures, and depreciation will grow at 4% in perpetuity and that your cost of capital is 12%. (There is no working capital.) a. Estimate the terminal value of the firm. b. What reinvestment rate and return and capital are you implicitly assuming in perpetuity when you do this? c. What would your terminal value have been if you had assumed that capital expenditures offset depreciation in stable growth? d. What return on capital are you implicitly assuming in perpetuity when you set capital expenditures equal to depreciation? 6. Crabbe Steel owns a number of steel plants in Pennsylvania. The firm reported after-tax operating income of $40 million in the most recent year on capital invested of $400 million. The firm expects operating income to grow 7% a year for the next three years, and 3% thereafter. a. If the firm's cost of capital is 10% and you expect the firm's current return on capital to continue in perpetuity, estimate the value at the end of the third year. b. If you expect operating income to stay fixed after year 3 (what you earn in year 3 is what you will earn every year thereafter), estimate the terminal value. c. If you expect operating income to drop 5% a year in perpetuity after year 3, estimate the terminal value. 7. How would your answers to the preceding problem change if you were told that the cost of capital for the firm is 8%? 1
Growth without excess returns will make a firm larger but not add value.
While Jack Welch (GE) and Robert Goisueta (Coca-Cola) represent traditional examples of CEOs who made a difference. Steve Jobs at Apple set a new standard for the difference making CEO. 2
Two-thirds of U.S. firms have betas that fall between 0.8 and 1.2. That becomes the range for stable period betas. 3
If you are valuing a commodity company and assuming any growth rate that exceeds inflation, you are assuming that your firm will branch out into other businesses and you have to adjust the beta accordingly. 4
Profes sor Altman at NYU's Stern School of Business estimates these probabilities as part of an annual series that he updates. The latest version is available from the Stern School of Business working paper series. 5
231
CHAPTER 13 Dividend Discount Models In the strictest sense, the only cash flow you receive when you buy shares in a publicly traded firm is a dividend. The simplest model for valu ing equity is the dividend discount model (DDM)—the value of a stock is the present value of expected dividends on it. While many analysts have turned away from the dividend discount model and view it as outmoded, much of the intuition that drives discounted cash flow valuation stems from the dividend discount model. In fact, there are companies where the dividend discount model remains a useful tool for estimating value. This chapter explores the general model as well as specific versions of it tailored for different assumptions about future growth. It also examines issues in using the dividend discount model and the results of studies that have looked at its efficacy.
THE GENERAL MODEL When an investor buys stock, he or she generally expects to get two types of cash flows—dividends during the period the stock is held and an expected price at the end of the holding period. Since this expected price is itself determined by future dividends, the value of a stock is the present value of dividends through infinity:
where DPSt = Expected dividends per share ke = Cost of equity The rationale for the model lies in the present value rule—the value of any asset is the present value of expected future cash flows, discounted at a rate appropriate to the riskiness o f the cash flows being discounted. There are two basic inputs to the model—expected dividends and the cost on equity. To obtain the expected dividends, we make assumptions about expected future growth rates in earnings and payout ratios. The required rate of return on a stock is determined by its riskiness, measured differently in different models—the market beta in the capital asset pricing model (CAPM) and the factor betas in the arbitrage and multifactor models. The model is flexible enough to allow for time-varying discount rates, where the time variation is because of expected changes in interest rates or risk across time.
VERSIONS OF THE MODEL Since projections of dollar dividends cannot be made through infinity, several versions of the dividend discount model have been developed based on different assumptions about future growth. We will begin with the simplest—a model designed to value stock in a stable growth firm that pays out what it can afford to in dividends—and then look at how the model can be adapted to value companies in high growth that may be paying little or no dividends.
The Gordon Growth Model The Gordon growth model can be used to value a firm that is in “steady state” with dividends growing at a rate that can be sustained forever.
The Model The Gordon growth model relates the value of a stock to its expected dividends i n the next time period, the cost of equity, and the expected growth rate in dividends.
What Is a Stable Growth Rate? While the Gordon growth model provides a simple approach to valuing equity, its use is limited to firms that are growing at a stable growth rate. There are two insights worth keeping in mind when estimating a stable growth rate. First, since the growth rate in the firm’s dividends is expected to last forever, the firm’s other measures of performance (including earnings) can also be expected to grow at the same rate. To see why, consider the consequences in the long term of a firm whose earnings grow 2 percent a year forever, while its dividends grow at 3 232
percent. Over time, the dividends will exceed earnings. If a firm’s earnings grow at a faster rate than dividends in the long term, the payout ratio, in the long term, will converge toward zero, which is also not a steady state. Thus, though the model’s requirement is for the expected growth rate in dividends, analysts should be able to substitute in the expected growth rate in earnings and get precisely the same result, if the firm is truly in steady state. The second issue relates to what growth rate is reasonable as a stable growth rate. As noted in Chapter 12, this growth rate has to be less than or equal to the growth rate of the economy in which the firm operates. No firm, no matter how well run, can be assumed to grow forever at a rate that exceeds the growth rate of the economy (or as a proxy, the risk-free rate). In addition, the caveats made in Chapter 12 about stable growth apply: The return on equity (ROE) that we assume in perpetuity should reflect not what the company may have made last year nor what it is expected to make next year, but, rather, a longer-term estimate. The estimate of ROE matters because the payout ratio in stable growth has to be consistent: The cost of equity has to be consistent with the firm being mature; if a beta is being used, it should be close to 1.
Limitations of the Model As most analysts discover quickly, the Gordon growth model is extremely sensitive to assumptions about the growth rate, as long as other inputs to the model (payout ratio, cost of equity) are kept constant. Consider a stock with an expected dividend per share next period of $2.50, a cost of equity of 15 percent, and an expected growth rate of 5 percent forever. The value of this stock is:
Note, however, the sensitivity of this value to estimates of the growth rate in Figure 13.1. As the growth rate approaches the cost of equity, the value per share approaches infinity. If the growth rate exceeds the cost of equity, the value per share becomes negative. Figure 13.1 Value per Share and Expected Growth Rate
DOES A STABLE GROWTH RATE HAVE TO BE CONSTANT OVER TIME? The assumption that the growth rate in dividends has to be constant over time may seem a difficult assumption to meet, especially given the volatility of earnings. If a firm has an average growth rate that is close to a stable growth rate, the model can be used with little real effect on value. Thus a cyclical firm that can be expected to have year-to-year swings in growth rates, but has an average growth rate that is 2 percent, can be valued using the Gordon growth model, without a significant loss of generality. There are two reasons for this result. First, since dividends are smoothed even when earnings are volatile, they are less likely to be affected by year-to-year change s in earnings growth. Second, the mathematical effects on present value of using year-specific growth rates rather than a constant growth rate are small.
There are, of course, two common sense fixes to this problem. The first is to work with the constraint that a stable growth rate cannot exceed the risk-free rate; in the preceding example, this would limit the growth rate to a number well below 15 percent. The second is to recognize that growth is not free; when the growth rate is increased, the payout ratio should be decreased. This creates a trade-off on growth, with the net effect of increasing growth being positive, neutral, or even negative.
Firms Model Works Best For In summary, the Gordon growth model is best suited for firms growing at a rate equal to or lower than the nominal 233
growth in the economy with well-established dividend payout policies that they intend to continue into the future. The dividend payout and cost of equity of the firm have to be consistent with the assumption of stability, since stable firms generally pay substantial dividends and have betas close to 1.1 In particular, this model will underestimate the value of the stock in firms that consistently pay out less than they can afford to and accumulate cash in the process.
ILLUSTRATION 13.1: Valuing a Regulated Monopoly: Consolidated Edison in May 2011 Consolidated Edison (Con Ed) is the electric utility that supplies power to residences and businesses in New York City. It is a quasi-monopoly whose prices and profits are regulated by the state of New York. We will be valuing Con Ed using a stable growth dividend discount model because it fits the criteria for the model: The firm operates in a region where the population and power usage have leveled off over the past f ew decades. The regulatory authorities will restrict price increases to be about the inflation rate. The firm has had a stable mix of debt and equity funding its operations for decades. Con Ed has a clientele of dividend-loving investors, and attempts to pay out as much as it can in dividends. During the period 2006– 2010, the firm returned about 95% of its free cash flows to equity (FCFE) as dividends. To value the company using the stable growth dividend discount model, we start with the earnings per share of $3.47 that the firm reported for 2010 and the dividends per share of $2.22 it paid out for the year. Using the average beta of 0.80 for power utilities and an equity risk premium of 5% for mature markets allows us to estimate a cost of equity of 7.5% (the risk-free rate was 3.5%)
Capping the growth rate at the risk-free rate of 3.5%, we generated a value per share of $57.46.
We check to see whether the expected growth rate was consistent with fundamentals for Con Ed. Retention ratio = 1 - ($2.22/$3.47) = 36% Return on equity = 9.79% Expected growth rate = .36 × .0979 = .0352 The fundamental growth rate is very close to our estimate of growth of 3.5%. The stock was trading at $53.47 a share in May 2011, making it slightly under valued.
IMPLIED GROWTH RATE The value for Con Ed is different from the market price, and this is likely to be the case with almost any company that you value. There are three possible explanations for this deviation. One is that you are right and the market is wrong. While this may be the correct explanation, you should probably make sure that the other two explanations do not hold—that the market is right and you are wrong or that the difference is too small to draw any conclusions. To examine the magnitude of the difference between the market price and your estimate of value, you can hold the other variables constant and change the growth rate in your valuation until the value converges on the price. Figure 13.2 estimates value as a function of the expected growth rate (assuming a beta of 0.80 and current dividends per share of $2.22). Solving for the expected growth rate that provides the current price, we get:
Figure 13.2 Con Ed: Value versus Growth Rate
234
The growth rate in earnings and dividends would have to be 3.21 percent a year to justify the stock price of $53.47. This growth rate is called an implied growth rate. Since we estimate growth from fundamentals, this allows us to estimate an implied return on equity:
ILLUSTRATION 13.2: Valuing a Mature Commodity Company Using Stable DDM: Total SA Total SA is a mature integrated, multinational oil company based in France. With its vo latile income, driven largely by swings in oil prices, it may seem to fit poorly into the mold of a stable company, but looking past the commodity price variance, it is a mature company, and its high dividend payout ratio reflects this stability. To value the company, we first average the net income and dividends from 2007 to 2010:
Using the average income and dividends over the four-year period, we estimate a payout ratio of 49.28%, a little low for a stable growth company but one that we will use for this valuation. The cost of equity was estimated using a beta of 0.90 (reflecting the average beta of integrated oil companies, a euro risk-free rate of 3.25%, and an equity risk premium of 5.50%; we augment the mature market premium of 5% with an additional 0.50% to capture the risk of the many risky markets where Total gets its oil:
Finally, we assume a stable growth rate of 2% (set just above the expected inflation rate) and valued the equity:
Total SA had a market capitalization of 97,286 million euros at the time of the valuation, making it over-valued by about 12.2%. It is worth noting that this is a conservative estimate of value. With a growth rate of 2%, Total SA should be capable of paying more in dividends, and its free cash flows to equity over this four-year period are higher than what it paid out.
DDMst.xls: This spreadsheet allows you to value a stable growth firm, with stable firm characteristics (beta and return on equity) and dividends that roughly match cash flows.
Two-Stage Dividend Discount Model The two-stage growth model allows for two stages of growth—an initial phase where the growth rate is not a stable growth rate and a subsequent steady state where the growth rate is stable and is expected to remain so for the long term. While, in most cases, the growth rate during the initial phase is higher than the stable growth rate, the model can be adapted to value companies that are expected to post low or even negative growth rates for a few years and then revert back to stable growth.
The Model 235
The model is based on two stages of growth, an extraordinary growth phase that lasts n years, and a stable growth phase that lasts forever after that:
where DPSt = Expected dividends per share in year t ke = Cost of equity (hg: high growth period; st: stable growth period) Pn = Price at the end of year n g = Extraordinary growth rate for the first n years gn = Growth rate forever after year n In the case where the extraordinary growth rate (g) and payout ratio are unchanged for the first n years, this formula can be simplified as follows:
where the inputs are as defined previously.
Calculating the Terminal Price The same constraint that applies to the growth rate for the Gordon growth model (i.e., that the growth rate in the firm is comparable to the nominal growth rate in the economy) applies for the terminal growth rate (gn) in this model as well. In addition, the payout ratio has to be consistent with the estimated growth rate. If the growth rate is expected to drop significantly after the initial growth phase, the payout ratio should be higher in the stable phase than in the growth phase. A stable firm can pay out more of its earnings in dividends than a growing firm. One way of estimating this new payout ratio is to use the fundamental growth model described in Chapter 12:
Algebraic manipulation yields the following stable period payout ratio:
Thus a firm with a 5 percent growth rate and a return on equity of 15 percent will have a stable period payout ratio of 66.67 percent. The other characteristics of the firm in the stable period should be consistent with the assumption of stability. For instance, it is reasonable to assume that a high growth firm has a beta of 2.0, but unreasonable to assume that this beta will remain unchanged when the firm becomes stable. In fact, the rule of thumb that we developed in the previous chapter—that stable period betas be between 0.8 and 1.2—is worth repeating here. Similarly, the return on equity, which can be high during the initial growth phase, should come down to levels commensurate with a stable firm in the stable growth phase. What is a reasonable stable period return on equity? The industry average return on equity and the firm’s own stable period cost of equity provide useful information to make this judgment.
Limitations of the Model There are three problems with the two-stage dividend discount model; the first two would apply to any two-stage model, and the third is specific to the dividend discount model. 1. The first practical problem is in defining the length of the extraordinary growth period. Since the growth rate is expected to decline to a stable level after this period, the value of an investment will increase as this period is 236
made longer. While we did develop criteria that might be useful in making this judgment in Chapter 12, it is difficult in practice to convert these qualitative considerations into a specific time period. 2. The second problem with this model lies in the assumption that the growth rate is high during the initial period and is transformed overnight to a lower stable rate at the end of the period. While these sudden transformations in growth can happen, it is much more realistic to assume that the shift from high growth to stable growth happens gradually over time. 3. The focus on dividends in this model can lead to skewed estimates of value for firms that are not paying out what they can afford to in dividends. In particular, we will underestimate the value of firms that accumulate cash and pay out too little in dividends.
Firms Model Works Best For Since the two-stage dividend discount model is based on two clearly delineated growth stages—high growth and stable growth—it is best suited for firms that are in high growth and expect to maintain that growth rate for a specific time period, after which the sources of the high growth are expected to disappear. One scenario, for instance, where this may apply is when a company has patent rights to a very profitable product for the next few years, and is expected to enjoy supernormal growth during this period. Once the patent expires, it is expected to settle back into stable growth. Another scenario where it may be reasonable to make this assumption about growth is when a firm is in an industry that is enjoying supernormal growth because there are significant barriers to entry (either legal or as a consequence of infrastructure requirements), which can be expected to keep new entrants out for several years. The assumption that the growth rate drops precipitously from its level in the initial phase to a stable rate also implies that this model is more appropriate for firms with modest growth rates in the initial phase. For instance, it is more reasonable to assume that a firm growing at 7 percent in the high growth period will see its growth rate drop to 2 percent afterward than it is for a firm growing at 40 percent in the high-growth period. Finally, the model works best for firms that maintain a policy of paying out residual cash flows (i.e., cash flows left over after debt payments and reinvestment needs have been met) as dividends.
ILLUSTRATION 13.3: Valuing a Firm with a Two-Stage Dividend Discount Model: Procter & Gamble in May 2011 Procter & Gamble (P&G) is one the leading global consumer product companies, owning some of the most valuable brands in the world, including Gillette razors, Pampers diapers, Tide detergent, Crest toothpaste, and Vicks cough medicine. P&G’s long history of paying dividends makes it a good candidate for the dividend discount model, and while it is a large company, its brand names and global expansion provide it with a platform to deliver high growth at least for the next few years. Consequently, we will use the two-stage dividend discount model to value the company. To set the stage, P&G reported $12,736 million in earnings for 2010 and paid out 49.74% of these earnings as dividends; on a per share basis, earnings were $3.82 and dividends were $1.92 in 2010. We will use a beta of 0.90, reflecting the beta of large consumer product companies in 2010, a risk-free rate of 3.50%, and a mature market equity risk premium of 5% to estimate the cost of equity:
To estimate the expected growth rate, we will start with the firm’s current return on equity (20.09%) and payout ratio (49.74%) and assume numbers very close to these for the next five years: Expected ROE for next 5 years = 20% Expected retention ratio for next 5 years = 50% Expected growth rate for next 5 years = 20% × 50% = 10% Applying this growth rate to earnings and dividends for the next 5 years and discounting these dividends back at the cost of equity, we arrive at a value of $10.09/share for the high growth period:
After year 5, we assume that P&G will be in stable growth, growing 3% a year (set just below the risk-free rate). We also assume that the return on equity for the firm will drop to a more sustainable 12% in perpetuity, resulting in an estimated payout ratio of 75% in perpetuity:
Assuming that the beta moves up to 1 in stable growth (resulting in a cost of equity of 8.5%), we estimate the value per share at the end of year 5:
237
Discounting this price to the present at 8% (the cost of equity for the high growth period) and adding the present value of expected dividends during the high growth period yields a value per share of $79.00.
The stock was trading at $68 in May 2011, making it fairly valued.
DDM2st.xls: This spreadsheet allows you to value a growth firm, with an initial period of high growth and stable growth thereafter, using expected dividends.
A TROUBLESHOOTING GUIDE: WHAT IS WRONG WITH THIS VALUATION? (TWOSTAGE DDM) If This Is Your Problem This May Be the Solution If you get an extremely low value from the two-stage DDM, the likely culprits are: If using fundamentals, use a higher ROE. The stable period payout ratio is too low for a stable firm (< 40%). If entering directly, enter a higher payout. The beta in the stable period is too high for a stable firm. The two-stage model is being used when the three-stage model is more appropriate. If you get an extremely high value:
Use a beta closer to 1.
The growth rate in the stable growth period is too high for a stable firm.
Use a growth rate less than the risk-free rate, and make sure that your retention ratio is consistently estimated.
Use a three-stage model.
Modifying the Model to Include Stock Buybacks In recent years, firms in the United States have increasingly turned to stock buybacks as a way of returning cash to stockholders. Figure 13.3 presents the cumulative amounts paid out by firms in the form of dividends and stock buybacks from 1988 to 2010. The trend toward stock buybacks was very strong, especially in the 1990s. Even the banking crisis of 2008 created only a momentary blip in buybacks in 2009, before they returned in force in 2010. Figure 13.3 Stock Buybacks and Dividends: Aggregate for U.S. Firms—1989 to 2010
What are the implications for the dividend discount model? Focusing strictly on dividends paid as the only cash returned to stockholders exposes us to the risk that we might be missing significant cash returned to stockholders in the form of stock buybacks. The simplest way to incorporate stock buybacks into a dividend discount model is to add them onto the dividends and compute an augmented payout ratio: 238
While this adjustment is straightforward, the resulting ratio for any one year can be skewed by the fact that stock buybacks, unlike dividends, are not smoothed out. In other words, a firm may buy back $3 billion in stock in one year, and not buy back stock for th e next three years. Consequently, a much better estimate of the modified payout ratio can be obtained by looking at the average value over a four- or five-year period. In addition, firms may sometimes buy back stock as a way of increasing financial leverage. We could adjust for this by netting out new debt issued from the earlier calculation:
Adjusting the payout ratio to include stock buybacks will have ripple effects on estimated growth and the terminal value. In particular, the modified growth rate in earnings per share can be written as:
Even the return on equity can be affected by stock buybacks. Since the book value of equity is reduced by the market value of equity bought back, a firm that buys back stock can reduce its book equity (and increase its return on equity) dramatically. If we use this return on equity as a measure of the marginal return on equity (on new investments), we will overstate the value of a firm. Adding back stock buybacks in recent years to the book equity and reestimating the return on equity can sometimes yield a more reasonable estimate of the return on equity on investments.
ILLUSTRATION 13.4: Augmented versus Conventional Dividend Payout Ratios: Coca-Cola To illustrate the effect of using augmented dividends versus actual dividends, we will look at Coca-Cola, a company that has bought back stock between 2006 and 2010. In the following table, we estimate the total cash returned to stockholders each year from 2006 to 2010 and contrast the augmented payout ratio with the conventional payout ratio:
The augmented dividend payout is higher than the dividend payout ratio in each year, but stock buybacks are volatile. That is why we would look at the augmented dividend payout ratio in the aggregate over the entire period; that number is 63.60%, higher than the conventional payout ratio of 49.15%. How would this play out in a valuation of Coca-Cola? Using the higher augmented payout ratio will result in higher cash flows to stockholders in the high growth phase, which should increase value. This effect, however, will be partly or even fully offset by a lower fundamental growth rate. In the case of Coca-Cola, where we will assume a return on equity of 25%, the expected growth rate using the higher augmented payout ratio can be computed as follows:
In contrast, using the conventional payout ratio would have yielded an expected growth rate of more than 12.5%.
Valuing an Entire Market Using the Dividend Discount Model All our examples of the dividend discount model so far have involved individual companies, but there is no reason why we cannot apply the same model to value a sector or even the entire market. The market price of the stock would be replaced by the cumulative market value of all of the stocks in the sector or market. The expected dividends would be the cumulated dividends of all these stocks, and could be expanded to include stock buybacks by all firms. The expected growth rate would be the growth rate in cumulated earnings of the index. There would be no need for a beta or betas, since you are looking at the entire market (which should have a beta of 1), and you could add the risk premium (or premiums) to the risk-free rate to estimate a cost of equity. You could use a twostage model, where this growth rate is greater than the growth rate of the economy, but you should be cautious about setting the growth rate too high or the growth period too long, because it will be difficult for cumulated earnings growth of all firms in an economy to run ahead of the growth rate in the economy for extended periods. Consider a simple example. Assume that you have an index trading at 700, and that the average dividend yield of stocks in the index is 5 percent. Earnings and dividends can be expected to grow at 4 percent a year forever, and the riskless rate is 5.4 percent. If you use a market risk premium of 4 percent, the value of the index can be estimated as follows: 239
At its existing level of 700, the market is slightly overpriced.
ILLUSTRATION 13.5: Valuing the S&P 500 Using Dividends and Augmented Dividends On January 1, 2011, the S&P 500 was trading at 1,257.64, and the dividends on the index amounted to 23.12 over the previous year. On the same date, analysts were estimating an expected growth rate of 6.95% in earnings for the index for the following five years. Assuming that dividends grow at the same rate as earnings, we obtain the following:
To estimate the cost of equity, we assume a beta of 1 for the index and use the risk-free rate on January 1, 2011, of 3.29% and an equity risk premium of 5%:
After year 5, earnings and dividends are expected to grow at 3.29%, the same nominal rate as the economy (assumed to be equal to the riskfree rate). The value that we obtained for the index follows:
This suggest that the index was massively over valued on January 1, 2011. Since many of the companies in the index have chosen to return cash in the form of stock buybacks, rather than dividends, a more realistic estimate of value would incorporate these expected buybacks. To do so, we added the buybacks in 2010 to the dividends to arrive at a value of 53.96 for augmented dividends on the index. Applying the same parameters that we used for conventional dividends (growth rate of 6.95% for the next five years and 3.29% beyond year 5), we estimate a new value for the index:
With buybacks incorporated, the index looks slightly undervalued (by about 4%).
The Value of Growth Investors pay a premium when they acquire companies with high growth potential. This premium takes the form of higher price-earnings or price–book value ratios. While no one will contest the proposition that growth is valuable, it is possible to pay too much for growth. In fact, empirical studies that show low price-earnings ratio stocks earning return premiums over high price-earnings ratio stocks in the long term support the notion that investors overpay for growth. This section uses the two-stage dividend discount model to examine the value of growth, and it provides a benchmark that can be used to compare the actual prices paid for growth.
Estimating the Value of Growth The value of the equity in any firm can be written in terms of three components:
240
where DPSt = Expected dividends per share in year t ke = Cost of equity gn = Growth rate forever after year n Value of extraordinary growth = Value of the firm with extraordinary growth in first n years Value of the firm as a stable growth firm2 Value of stable growth = Value of the firm as a stable growth firm - Value of firm with no growth Assets in place = Value of firm with no growth In making these estimates, though, we have to remain consistent. For instance, to value assets in place, you would have to assume that the entire earnings could be paid out in dividends, while the payout ratio used to value stable growth should be a stable period payout ratio.
ILLUSTRATION 13.6: The Value of Growth: P&G in May 2011 In Illustration 13.3, we valued P&G using a two-stage dividend discount model at $68.90. We first value the assets in place using current earnings ($3.82) and assume that all earnings are paid out as dividends. We also use the stable growth cost of equity as the discount rate.
To estimate the value of stable growth, we assume that the expected growth rate will be 3% and that the payout ratio is the stable period payout ratio of 75%:
Note that $68.90 was our estimate of value per share in Illustration 13.3. DETERMINANTS OF THE VALUE OF GROWTH
Growth rate during extraordinary period. The higher the growth rate in the extraordinary period, the higher the estimated value of growth. Conversely, the value of high growth companies can drop precipitously if the expected growth rate is reduced, either because of disappointing earnings news from the firm or as a consequence of external events. Length of the extraordinary growth period. The longer the extraordinary growth period, the greater the value of growth. At an intuitive level, this is fairly simple to illustrate. The value of $15.25 obtained for extraordinary growth in P&G is predicated on the assumption that high growth will last for five years. If this is revised to last 10 years, the value of extraordinary growth will increase. Profitability of projects. The excess returns earned by projects determine the value of the expected growth. At the limit, growth becomes worthless if the firm earns a return on equity that is equal to the cost of equity.
H Model for Valuing Growth The H model is a two-stage model for growth, but unlike the classic two-stage model, the growth rate in the initial growth phase is not constant but declines linearly over time to reach the stable growth rate in steady state. This model was presented in Fuller and Hsia (1984).
The Model The model is based on the assumption that the earnings growth rate starts at a high initial rate (ga) and declines linearly over the extraordinary growth period (which is assumed to last 2H periods) to a stable growth rate (gn). It also assumes that the dividend payout and cost of equity are constant over time, and are not affected by the 241
shifting growth rates. Figure 13.4 graphs the expected growth over time in the H model. Figure 13.4 Expected Growth in the H Model
The value of expected dividends in the H model can be written as follows:
where P0 = Value of the firm now per share DPSt = DPS in year t ke = Cost of equity ga = Growth rate initially gn = Growth r ate at end of 2H years, applies forever after that
Limitations This model avoids the problems associated with the growth rate dropping precipitously from the high growth to the stable growth phase, but it does so at a cost. First, the decline in the growth rate is expected to follow the strict structure laid out in the model—it drops in linear increments each year based on the initial growth rate, the stable growth rate, and the length of the extraordinary growth period. While small deviations from this assumption do not affect the value significantly, large deviations can cause problems. Second, the assumption that the payout ratio is constant through both phases of growth exposes the model to an inconsistency—as growth rates decline, the payout ratio usually increases.
Firms Model Works Best For The allowance for a gradual decrease in growth rates over time may make this a useful model for firms that are growing rapidly right now, but where the growth is expected to decline gradually over time as the firms get larger and the differential advantage they have over their competitors declines. The assumption that the payout ratio is constant, however, makes this an inappropriate model to use for any firm that has low or no dividends currently. Thus, the model, by requiring a combination of high growth and high payout, may be quite limited3 in its applicability.
ILLUSTRATION 13.7: Valuing with the H Model: Vodafone Vodafone, a UK-based telecommunications firm, paid dividends per share of 9.8 pence on earnings per share of 16.1 pence in 2010. The firm’s earnings per share had grown at 6% over the prior five years but the growth rate is expected to decline linearly over the next five years to 3%, while the payout ratio remains unchanged. The beta for the stock is 1, the risk-free rate in British pounds is 4% and the market risk premium is 5%.
The stock can be valued using the H model:
242
The stock was trading at 173.3 pence in May 2011, making it slightly under-valued.
DDMH.xls: This spreadsheet allows you to value a firm, with an initial period when the high growth declines to stable growth, using expected dividends.
Three-Stage Dividend Discount Model The three-stage dividend discount model combines the features of the two-stage model and the H model. It allows for an initial period of high growth, a transitional period where growth declines, and a final stable growth phase. It is the most general of the models because it does not impose any restrictions on the payout ratio.
The Model This model assumes an initial period of stable high growth, a second period of declining growth, and a third period of stable low growth that lasts forever. Figure 13.5 graphs the expected growth over the three time periods. Figure 13.5 Expected Growth in the Three-Stage Dividend Discount Model
The value of the stock is then the present value of expected dividends during the high-growth and the transitional periods, and of the terminal price at the start of the final stable growth phase.
where EPSt = Earnings per share in year t DPSt = Dividends per share in year t g a = Growth rate in high-growth phase (lasts n1 periods) g n = Growth rate in stable phase Πa = Payout ratio in high-growth phase Πn = Payout ratio in stable growth phase ke = Cost of equity in high growth (hg), transition (t), and stable growth (st)
Assumptions 243
This model removes many of the constraints imposed by other versions of the dividend discount model. In return, however, it requires a much larger number of inputs—year-specific payout ratios, growth rates, and betas. For firms where there is substantial noise in the estimation process, the errors in these inputs can overwhelm any benefits that accrue from the additional flexibility in the model.
Firms Model Works Best For This model’s flexibility makes it a useful model for any firm that in addition to changing growth over time is expected to change on other dimensions as well—in particular, payout policies and risk. It is best suited for firms that are growing at an extraordinary rate now and are expected to maintain this rate for an initial period, after which the differential advantage of the firm is expected to deplete leading to gradual declines in the growth rate to a stable growth rate. Practically speaking, this may be the more appropriate model to use for a firm whose earnings are growing at very high rates,4 are expected to continue growing at those rates for an initial period, but are expected to start declining gradually toward a stable rate as the firm become larger and loses its competitive advantages.
ILLUSTRATION 13.8: Valuing Coca-Cola Using a Three-Stage Dividend Discount Model To value Coca-Cola in May 2011, we used a three-stage dividend discount model, partly because we expect the firm to maintain a growth rate higher than the economy for the next few years and partly because it has a history of paying substantial dividends. It has had a low debt ratio and has shown no indications that it plans to alter its approach to financing. In 2010, the company reported earnings per share of $3.56 and paid out $1.88 per share in dividends. To estimate the expected growth rate, we assumed that the firm would be able to generate 25% as a return on equity on futur e investments, lower than its current return on equity but close to its marginal return on equity over the last few years. We also assumed that the firm would reinvest about 36.40% of its earnings as dividends; while this is lower than the existing retention ratio of 47.19%, it is consistent with the retention ratio that we estimated in Illustration 13.4, using the augmented dividends. Expected ROE for next 5 years = 25% Expected Retention Ratio for next 5 years = 36.40% Expected growth rate in EPS and DPS for next 5 years = 25% × 36.4% = 9.10% During this high growth phase, we will assume that the cost of equity for Coca-Cola will be 8.45%, estimated based upon a beta of 0.9, the U.S. treasury bond rate in 2011 of 3.5% and an equity risk premium of 5.5% (with the premium augmented to reflect Coca-Cola’s exposure in emerging markets).
The expected dividends over the next 5 years are shown in the following table, with the present values computed using the cost of equity:
After year 5, we allow for a transition period of 5 more years to stable growth after year 10. In the stable growth phase, we assume the following changes: Expected growth rate of 3% forever, set just below the risk-free rate. A return on equity of 15%; while this is lower than the current ROE, it is an impressive return for a mature firm and reflects our belief that Coca-Cola’s brand name will endure. A payout ratio of 80%, based on the return on equity and expected growth rate:
A cost of equity of 9.00%, based upon the assumption that the beta will increase to 1 in stable growth. The transition period (years 6–10) allow us to change each of the inputs (payout ratio, cost of equity and growth rate) from high growth levels to stable growth levels in linear increments. The resulting dividends and present values are summarized here:
244
Note that the changing cost of equity requires us to estimate a cumulated cost of equity. Thus, the cumulated cost of equity for year 7 is:
The value per share at the end of year 10 can now be obtained:
Discounting the terminal value back at the cumulated cost of equity for year 10 and adding to the present value of dividends, we get a value per share of $67.15.
Coca-Cola was trading at $68.22 in May 2011, making it fairly valued.
DDM3st.xls: This spreadsheet allows you to value a firm with a period of high growth followed by a transition period where growth declines to a stable growth rate.
A TROUBLESHOOTING GUIDE: WHAT IS WRONG WITH THIS MODEL? (THREESTAGE DDM) If This Is Your Problem If you are getting too low a value from this model:
This May Be the Solution
The stable period payout ratio is too low for a stable firm (< 40%).
If using fundamentals, use a higher ROE. If entering directly, enter a higher payout.
The beta in the stable period is too high for a stable firm. If you get an extremely high value:
Use a beta closer to 1.
The growth rate in the stable growth period is too high for stable firm. Use a growth rate less than the riskfree rate. The period of growth (high + transition) is too high.
Use shorter high growth and transition periods.
ISSUES IN USING THE DIVIDEND DISCOUNT MODEL The dividend discount model’s primary attraction is its simplicity and its intuitive logic. There are many analysts, however, who view its results with suspicion because of limitations that they perceive it to possess. The model, they claim, is not really useful in valuation except for a limited number of stable, high-dividend-paying stocks. This section examines some of the areas where the dividend discount model is perceived to fall short.
Valuing Non-Dividend-Paying or Low-Dividend-Paying Stocks The conventional wisdom is that the dividend discount model cannot be used to value a stock that pays low or no dividends. It is wrong. If the dividend payout ratio is adjusted to reflect changes in the expected growth rate, a value can be obtained even for non-dividend-paying firms. Thus, a high-growth firm, paying no dividends currently, can still be valued based on dividends that it is expected to pay out when the gr owth rate declines. If the payout ratio is not adjusted to reflect changes in the growth rate, however, the dividend discount model will underestimate the value of non-dividend-paying or low-dividend-paying stocks.
Is the Model Too Conservative in Estimating Value? A standard critique of the dividend discount model is that it provides too conservative an estimate of value. This criticism is predicated on the notion that the value is determined by more than the present value of expected dividends. For instance, it is argued that the dividend discount model does not reflect the value of “unutilized assets.” There is no reason, however, that these unutilized assets cannot be valued separately and added on to the value from the dividend discount model. Some of the assets that are supposedly ignored by the dividend discount model, such as the value of brand names, can be dealt with fairly simply within the context of the model. A more legitimate criticism of the model is that it does not incorporate other ways of returning cash to stockholders (such as stock buybacks). If you use the augmented version of the dividend discount model, this criticism can also be countered. 245
Contrarian Nature of the Model The dividend discount model is also considered by many to be a contrarian model. As the market rises, fewer and fewer stocks, they argue, will be found to be undervalued using the dividend discount model. This is not necessarily true. If the market increase is due to an improvement in economic fundamentals, such as higher expected growth in the economy and/or lower interest rates, there is no reason, a priori, to believe that the values from the dividend discount model will not increase by an equivalent amount. If the market increase is not due to fundamentals, the dividend discount model values will not follow suit, but that is more a sign of strength than weakness. The model is signaling that the market is overvalued relative to dividends and cash flows, and the cautious investor will pay heed.
TESTS OF THE DIVIDEND DISCOUNT MODEL The ultimate test of a model lies in how well it works at identifying undervalued and overvalued stocks. The dividend discount model has been tested and the results indicate that it does, in the long term, provide for excess returns. It is unclear, however, whether this is because the model is good at finding undervalued stocks or because it proxies for well-known empirical irregularities in returns relating to price-earnings ratios and dividend yields.
Simple Test of the Dividend Discount Model A simple study of the dividend discount model was conducted by Sorensen and Williamson, where they valued 150 stocks from the S&P 400 in December 1980 using the dividend discount model. They used the difference between the market price at that time and the model value to form five portfolios based on the degree of under or over valuation. They made fairly broad assumptions in using the dividend discount model: The average of the earnings per share between 1976 and 1980 was used as the current earnings per share. The cost of equity was estimated using the CAPM. The extraordinary growth period was assumed to be five years for all stocks, and the I/B/E/S consensus forecast of earnings growth was used as the growth rate for this period. The stable growth rate, after the extraordinary growth period, was assumed to be 8 percent for all stocks. The payout ratio was assumed to be 45 percent for all stocks. The returns on these five portfolios were estimated for the following two years (January 1981 to January 1983) and excess returns were estimated relative to the S&P 500 index using the betas estimated at the first stage and the CAPM. Figure 13.6 illustrates the excess returns earned by the portfolio that was undervalued by the dividend discount model relative to both the market and the overvalued portfolio. Figure 13.6 Performance of the Dividend Discount Model, 1981–1983
The undervalued portfolio had a positive excess return of 16 percent per annum between 1981 and 1983, while the overvalued portfolio had a negative excess return of 15 percent per annum during the same time period. Other studies that focus only on the dividend discount model come to similar conclusions. In the long term, undervalued and overvalued stocks from the dividend discount model outperform and underperform, respectively, the market index on a risk-adjusted basis.
Caveats on the Use of the Dividend Discount Model The dividend discount model provides impressive results in the long term. There are, however, three considerations in generalizing the findings from these studies:
The Dividend Discount Model Does Not Beat the Market Every Year 246
The dividend discount model outperforms the market over five-year time periods, but there have been individual years where the model has significantly underperformed the market. Haugen reports on the results of a fund that used the dividend discount model to analyze 250 large capitalization firms and to classify them into five quintiles from the first quarter of 1979 to the last quarter of 1991. The betas of these quintiles were roughly equal. The valuation was done by six analysts who estimated an extraordinary growth rate for the initial high-growth phase, the length of the high-growth phase, and a transitional phase for each of the firms. The returns on the five portfolios, as well as the returns on all 250 stocks and the S&P 500 from 1979 to 1991, are reported in Table 13.1. The undervalued portfolio earned significantly higher returns than the overvalued portfolio and the S&P 500 for the 1979–1991 period, but it underperformed the market in 6 of the 13 years and the overvalued portfolio in 4 of the 13 years. Table 13.1 Returns on Quintiles: Dividend Discount Model
Is the Model Just Proxying for Low PE Ratios and Dividend Yields? The dividend discount model weights expected earnings and dividends in near periods more than earnings and dividends in far periods, and is biased toward finding low price-earnings ratio stocks with high dividend yields to be undervalued and high price-earnings ratio stocks with low or no dividend yields to be overvalued. As noted in Chapter 6, studies of market efficiency indicate that low-PE-ratio stocks have outperformed (in terms of excess returns) high-PE-ratio stocks over extended time periods. Similar conclusions have been drawn about high-dividendyield stocks relative to low-dividend-yield stocks. Thus, the valuation findings of the model are consistent with empirical irregularities observed in the market. It is unclear how much the model adds in value to investment strategies that use PE ratios or dividend yields to screen stocks. Jacobs and Levy (1988b) indicate that the marginal gain is relatively small. Attribute
Average Excess Return per Quarter: 1982–1987
Dividend discount model 0.06% per quarter Low P/E ratio
0.92% per quarter
Book/price ratio
0.01% per quarter
Cash flow/price
0.18% per quarter
Sales/price ratio
0.96% per quarter
Dividend yield
–0.51% per quarter
This suggests that using low PE ratios to pick stocks adds 0.92 percent to your quarterly returns, whereas using the dividend discount model adds only a further 0.06 percent to quarterly returns. If, in fact, the gain from using the dividend discount model is that small, screening stocks on the basis of observables (such as PE ratio or cash flow measures) may provide a much larger benefit in terms of excess returns.
Tax Disadvantages from High-Dividend Stocks Portfolios created with the dividend discount model are generally characterized by high dividend yield, which can create a tax disadvantage if dividends are taxed at a rate greater than capital gains or if there is a substantial tax timing liability associated with dividends.5 Since the excess returns uncovered in the studies presented earlier are pretax to the investor, the introduction of personal taxes may significantly reduce or even eliminate these excess returns. In summary, the dividend discount model’s impressive results in studies looking at past data have to be considered with caution. For a tax-exempt investment with a long time horizon, the dividend discount model is a good tool (though it may not be the only one) to pick stocks. For a taxable investor, the benefits are murkier, since the tax 247
consequences of the strategy have to be considered. For investors with shorter time horizons, the dividend discount model may not deliver on its promised excess returns because of the year-to-year volatility in its performance.
CONCLUSION When you buy stock in a publicly traded firm, the only cash flow you receive directly from this investment in expected dividends. The dividend discount model builds on this simple proposition and argues that the value of a stock then has to be the present value of expected dividends over time. Dividend discount models can range from simple growing perpetuity models such as the Gordon growth model, where a stock’s value is a function of its expected dividends next year, the cost of equity, and the stable growth rate, to complex three-stage models, where payout ratios and growth rates change over time. While the model is often criticized as being of limited value, it has proven to be surprisingly adaptable and useful in a wide range of circ*mstances. It may be a conservative model that finds fewer and fewer undervalued firms as market prices rise relative to fundamentals (earnings, dividends, etc.); but that can also be viewed as a strength. Tests of the model also seem to indicate its usefulness in gauging value, though much of its effectiveness may be derived from its finding low-PE-ratio, high-dividend-yield stocks to be undervalued .
QUESTIONS AND SHORT PROBLEMS In the problems following, use an equity risk premium of 5.5 percent if none is specified. 1. Respond true or false to the following statements relating to the dividend discount model: a. The dividend discount model cannot be used to value a high-growth company that pays no dividends. True ____ False ____ b. The dividend discount model will undervalue stocks, because it is too conservative. True ____ False ____ c. The dividend discount model will find more undervalued stocks when the overall stock market is depressed. True ____ Fa lse ____ d. Stocks that are undervalued using the dividend discount model have generally made significant positive excess returns over long time periods (five years or more). True ____ False ____ e. Stocks that pay high dividends and have low price-earnings ratios are more likely to come out as undervalued using the dividend discount model. True ____ False ____ 2. Ameritech Corporation paid dividends per share of $3.56 in 1992, and dividends are expected to grow 5.5% a year forever. The stock has a beta of 0.90, and the Treasury bond rate is 6.25%. (Risk premium is 5.5%.) a. What is the value per share, using the Gordon growth model? b. The stock was trading for $80 per share. What would the growth rate in dividends have to be to justify this price? 3. Church & Dwight, a large producer of sodium bicarbonate, reported earnings per share of $1.50 in 1993 and paid dividends per share of $0.42. In 1993, the firm also reported the following:
The firm faced a corporate tax rate of 38.5%. (The market value debt-to-equity ratio is 5%. The Treasury bond rate is 7%.) The firm expected to maintain these financial fundamentals from 1994 to 1998, after which it was expected to become a stable firm, with an earnings growth rate of 6%. The firm’s financial characteristics were expected to approach industry averages after 1998. The industry averages were as follows:
248
Church & Dwight had a beta of 0.85 in 1993, and the unlevered beta was not expected to change over time. a. What is the expected growth rate in earnings, based on fundamentals, for the high-growth period (1994 to 1998)? b. What is the expected payout ratio after 1998? c. What is the expected beta after 1998? d. What is the expected price at the end of 1998? e. What is the value of the stock, using the two-stage dividend discount model? f. How much of this value can be attributed to extraordinary growth? To stable growth? 4. Oneida Inc, the world’s largest producer of stainless steel and silverplated flatware, reported earnings per share of $0.80 in 1993, and paid dividends per share of $0.48 in that year. The firm was expected to report earnings growth of 25% in 1994, after which the growth rate was expected to decline linearly over the following six years to 7% in 1999. The stock was expected to have a beta of 0.85. (The Treasury bond rate is 6.25%, and the risk premium is 5.5%.) a. Estimate the value of stable growth, using the H model. b. Estimate the value of extraordinary growth, using the H model. c. What are the assumptions about dividend payout in the H model? 5. Medtronic Inc., the world’s largest manufacturer of implantable biomedical devices, reported earnings per share in 1993 of $3.95, and paid dividends per share of $0.68. Its earnings were expected to grow 16% from 1994 to 1998, but the growth rate was expected to decline each year after that to a stable growth rate of 6% in 2003. The payout ratio was expected to remain unchanged from 1994 to 1998, after which it would increase each year to reach 60% in steady state. The stock was expected to have a beta of 1.25 from 1994 to 1998, after which the beta would decline each year to reach 1.00 by the time the firm becomes stable. (The Treasury bond rate is 6.25%, and the risk premium is 5.5%.) a. Assuming that the growth rate declines linearly (and the payout ratio increases linearly) from 1999 to 2003, estimate the dividends per share each year from 1994 to 2003. b. Estimate the expected price at the end of 2003. c. Estimate the value per share, using the three-stage dividend discount model. 6. Yuletide Inc. is a manufacturer of Christmas ornaments. The firm earned $100 million last year and paid out 20% of its earnings as dividends. The firm also has bought back $180 million of stock over the past four years, in varying amounts each year. The firm is in stable growth, expects to grow 5% a year in perpetuity, and has a cost of equity of 12%. a. Assuming that the dividend payout ratio will not change over time, estimate the value of equity. b. How would your answer change if your dividend payout ratio is modified to include stock buybacks? 1
The average payout ratio for large stable firms in the United States is about 60 percent.
The payout ratio used to calculate the value of the firm as a stable growth firm can be either the current payout ratio, if it is reasonable, or the new payout ratio calculated using the fundamental growth formula. 2
Proponents of the model would argue that using a steady-state payout ratio for firms that pay little or no dividends is likely to cause only small errors in the valuation. 3
The definition of a “very high” growth rate is largely subjective. As a rule of thumb, growth rates over 25 percent would qualify as v ery high when the stable growth rate is 6 to 8 percent. 4
Investors do not have a choice of when they receive dividends, whereas they have a choice on the timing of capital gains. 5
249
CHAPTER 14 Free Cash Flow to Equity Discount Models The dividend discount model is based on the premise that the only cash flows received by stockholders are dividends. Even if we use the modified version of the model and treat stock buybacks as dividends, we may misvalue firms that consistently fail to return what they can afford to their stockholders. This chapter uses a more expansive definition of cash flows to equity as the cash flows left over after meeting all financial obligations, including debt payments, and after covering capital expenditure and working capital needs. It discusses the reasons for differences between dividends and free cash flows to equity (FCFE), and presents the discou nted free cash flow to equity model for valuation.
MEASURING WHAT FIRMS CAN RETURN TO THEIR STOCKHOLDERS G iven what firms are returning to their stockholders in the form of dividends or stock buybacks, how do we decide whether they are returning too much or too little? We propose a simple measure how much cash is available to be paid out to stockholders after meeting reinvestment needs and compare this amount to the amount actually returned to stockholders.
Free Cash Flows to Equity To estimate how much cash a firm can afford to return to its stockholders, we begin with the net income—the accounting measure of the stockholders' earnings during the period—and convert it to a cash flow by subtracting out a firm's reinvestment needs. First, any capital expenditures, defined broadly to include acquisitions, are subtracted from the net income, since they represent cash outflows. Depreciation and amortization, on the other hand, are added back in because they are accounting but not cash expenses. The difference between capital expenditures and depreciation (net capital expenditures) is usually a function of the growth characteristics of the firm. High-growth firms tend to have high net capital expenditures relative to earnings, whereas low-growth firms may have low, and sometimes even negative, net capital expenditures. Second, increases in working capital drain a firm's cash flows, while decreases in working capital increase the cash flows available to equity investors. Firms that are growing fast, in industries with high working capital requirements (retailing, for instance), typically have large increases in working capital. Since we are interested in the cash flow effects, we consider only changes in noncash working capital in this analysis. Finally, equity investors also have to consider the effect of changes in the levels of debt on their cash flows. Repaying the principal on existing debt represents a cash outflow, but the debt repayment may be fully or partially financed by the issue of new debt, which is a cash inflow. Again, netting the repayment of old debt against the new debt issues provides a measure of the cash flow effects of changes in debt. Allowing for the cash flow effects of net capital expenditures, changes in working capital, and net changes in debt on equity investors, we can define the cash flows left over after these changes as the free cash flow to equity (FCFE):
This is the cash flow available to be paid out as dividends. Deconstructing this equation, the reinvestment by equity investors into the firm can be written as:
This calculation can be simplified if we assume that the net capital expenditures and working capital changes are financed using a fixed mix1 of debt and equity. If δ is the proportion of the net capital expenditures and working capital changes that is raised from debt financing, the effect on cash flows to equity of these items can be represented as follows:
250
Accordingly, the cash flow available for equity investors after meeting capital expenditure and working capital needs is:
Note that the net debt payment item is eliminated, because debt repayments are financed with new debt issues to keep the debt ratio fixed. It is appropriate to assume that a specified proportion of net capital expenditures and working capital needs will be financed with debt if the target or optimal debt ratio of the firm is used to forecast the free cash flow to equity that will be available in future periods. Alternatively, in examining past periods, we can use the firm's a verage debt ratio over the period to arrive at approximate free cash flows to equity.
WHAT ABOUT PREFERRED DIVIDENDS? In both the long and short formulations of free cash flows to equity described in this section, we assume that there are no preferred dividends paid. Since the equity that we value is only common equity, you would need to modify the formulas slightly for the existence of preferred stock and dividends. In particular, you would subtract the preferred dividends to arrive at the free cash flow to equity:
The debt ratio (δ) would then have to include the expected financing from new preferred stock issues.
ILLUSTRATION 14.1: Estimating Free Cash Flows to Equity—Disney In this illustration, we will compute the free cash flows to equity generated by Disney, the U.S. entertainment company, from 2001 to 2010, using the full calculation described in the last section:
To use the shortcut, first estimate the net debt used in aggregate over the entire period as a percentage of reinvestment (net cap ex and change in working capital):
Applying this net debt ratio to reinvestment yields the shorter version of FCFE:
251
While the aggregate FCFE over the period remains the same, the shortcut version yields smoother FCFE over the period.
Comparing Dividends to Free Cash Flows to Equity The conventional measure of dividend policy—the dividend payout ratio—gives us the value of dividends as a proportion of earnings. The modified approach measures the total cash returned to stockholders as a proportion of the free cash flow to equity:
The ratio of cash to stockholders to FCFE shows how much of the cash available to be paid out to stockholders is actually returned to them in the form of dividends and stock buybacks. If this ratio, over time, is equal or close to 1, the firm is paying out all that it can to its stockholders. If it is significantly less than 1, the firm is paying out less than it can afford to and is using the difference to increase its cash balance or to invest in marketable securities. If it is significantly over 1, the firm is paying out more than it can afford and is either drawing on an existing cash balance or issuing new securities (stocks or bonds). We can observe the tendency of firms to pay out less to stockholders than they have available in free cash flows to equity by examining cash returned to stockholders paid as a percentage of free cash flow to equity. In 2010, the global median of dividends as a percent of FCFE was about 60 percent, with most companies paying out less in dividends than they had available in FCFE. Figure 14.1 provides a global comparison of dividends to FCFE. While there is a significant segment of companies where dividends exceed FCFE, for the majority of companies the reverse is true. Figure 14.1 Dividends versus FCFE—Global Comparison
When a firm is paying out less in dividends than it has available in free cash flows, it is generating surplus cash. For those firms, this cash surplus appears as an increase in the cash balance. Firms that pay dividends that exceed FCFE have to finance these dividend payments either out of existing cash balances or by making new stock issues. The implications for valuation are simple. If we use the dividend discount model and do not allow for the buildup of cash that occurs when firms pay out less than they can afford, we will underestimate the value of equity in firms. If we use the model to value firms that pay out more dividends than they have available, we will overvalue the firms. The rest of this chapter is designed to correct for this limitation.
252
dividends.xls: This spreadsheet allows you to estimate the free cash flow to equity and the cash returned to stockholders for a period of up to 10 years.
divfcfe.xls: This dataset on the Web summarizes dividends, cash returned to stockholders, and free cash flows to equity, by sector, in the United States.
Why Firms May Pay Out Less than Is Available Many firms pay out less to stockholders, in the form of dividends and stock buybacks, than they have available in free cash flows to equity. The reasons vary from firm to firm.
Desire for Stability Firms are generally reluctant to change dividends, and dividends are considered “sticky” because the variability in dividends is significantly lower than the variability in earnings or cash flows. The unwillingness to change dividends is accentuated when firms have to reduce dividends, and increases in dividends outnumber cuts in dividends by at least a five-to-one margin in most periods. As a consequence of this reluctance to cut dividends, firms will often refuse to increase dividends even when earnings and FCFE go up, because they are uncertain about their capacity to maintain these higher dividends. This leads to a lag between earnings increases and dividend increases. Similarly, firms frequently keep dividends unchanged in the face of declining earnings and FCFE. Figure 14.2 reports the number of dividend changes (increases, decreases, no change) between 1988 and 2008. Figure 14.2 Dividend Changes by Year: U.S. Companies
The number of firms increasing dividends outnumbers those decreasing dividends seven to one. The number of firms, however, that do not change dividends outnumbers firms that do about four to one. Dividends are also less variable than either FCFE or earnings, but this reduced volatility is a result of keeping dividends significantly below the FCFE.
Future Investment Needs A firm might hold back on paying its entire FCFE as dividends if it expects substantial increases in capital expenditure needs in the future. Since issuing stocks is expensive (with floatation costs and issuance fees), it m ay choose to keep the excess cash to finance these future needs. Thus, to the degree that a firm is unsure about its future financing needs, it will retain some cash to take on unexpected investments or meet unanticipated needs.
Tax Factors If dividends are taxed at a higher tax rate than capital gains, a firm may choose to retain the excess cash and pay out much less in dividends than it has available. This is likely to be accentuated if the stockholders in the firm are in high tax brackets, as is the case with many family-controlled firms. If, however, investors in the firm like dividends or tax laws favor dividends, the firm may pay more out in dividends than it has available in FCFE, often borrowing or issuing new stock to do so. 253
Signaling Prerogatives Firms often use dividends as signals of future prospects, with increases in dividends being viewed as positive signals and decreases as negative signals. The empirical evidence is consistent with this signaling story, since stock prices generally go up on dividend increases and down on dividend decreases. The use of dividends as signals may lead to differences between dividends and FCFE.
Managerial Self-interest The managers of a firm may gain by retaining cash rather than paying it out as a dividend. The desire for empire building may make increasing the size of the firm an objective on its own. Or management may feel the need to build up a cash cushion to tide over periods when earnings may dip; in such periods, the cash cushion may reduce or obscure the earnings drop and may allow managers to remain in control.
FCFE VALUATION MODELS The free cash flow to equity model does not represent a radical departure from the traditional dividend discount model. In fact, one way to describe a free cash flow to equity model is that it represents a model where we discount potential dividends rather than actual dividends. Consequently, the three versions of the FCFE valuation model presented in this section are simple variants on the dividend discount model, with one significant change— free cash flows to equity replace dividends in the models.
Underlying Principle When we replace the dividends with FCFE to value equity, we are doing more than substituting one cash flow for another. We are implicitly assuming that the FCFE will be paid out to stockholders. There are two consequences: 1. There will be no future cash buildup in the firm, since the cash that is available after debt payments and reinvestment needs i s assumed to be paid out to stockholders each period. 2. The expected growth in FCFE will include growth in income from operating assets and not growth in income from increases in marketable securities. This follows directly from the last point. How does discounting free cash flows to equity compare with the augmented dividend discount model, where stock buybacks are added back to dividends and discounted? You can consider stock buybacks to be the return of excess cash accumulated largely as a consequence of not paying out their FCFE as dividends. Thus, FCFE represents a smoothed-out measure of what companies can return to their stockholders over time in the form of dividends and stock buybacks.
Estimating Growth in FCFE Free cash flows to equity, like dividends, are cash flows to equity investors and you could use the same approach that you used to estimate the fundamental growth rate in dividends per share:
The use of the retention ratio in this equation implies that whatever is not paid out as dividends is reinvested back into the firm. There is a strong argument to be made, though, that this is not consistent with the assumption that free cash flows to equity are paid out to stockholders, which underlies FCFE models. It is far more consistent to replace the retention ratio with the equity reinvestment rate, which measures the percent of net income that is invested back into the firm.
When discounting FCFE, it is safest to separate the existing cash balance from the operating assets of the firm, to value the equity in the operating assets and then add on the existing cash balance. Consequently, the return on equity can also have to be modified to reflect the fact that the conventional measure of the return includes interest income from cash and marketable securities in the numerator and the book value of equity also includes the value of the cash and marketable securities. In the FCFE model, there is no excess cash left in the firm and the return on equity should measure the return on noncash investments. You could construct a modified version of the return on equity that measures this:
The product of the equity reinvestment rate and the modified ROE will yield the expected growth rate in FCFE: 254
Constant Growth FCFE Model The constant growth FCFE model is designed to value firms that are growing at a stable growth rate and are hence in steady state.
The Model The value of equity, under the constant growth model, is a function of the expected FCFE in the next period, the stable growth rate, and the required rate of return.
where Value = Value of stock today FCFE1 = Expected FCFE next year ke = Cost of equity of the firm gn = Growth rate in FCFE for the firm forever
Caveats The model is very similar to the Gordon growth model in its underlying assumptions and works under some of the same constraints. The growth rate used in the model has to be reasonable, relative to the nominal growth rate in the economy in which the firm operates. As a general rule, a stable growth rate cannot exceed the growth rate of the economy in which the firm operates. The assumption that a firm is in steady state also implies that it possesses other characteristics shared by stable firms. This would mean, for instance, that capital expenditures are not disproportionately large, relative to depreciation, and the firm is of average risk. (If the capital asset pricing model is used, the beta of the equity should be close to 1.) To estimate the reinvestment for a stable growth firm, you can use one of two approaches: (a) You can use the typical reinvestment rates for firms in the industry to which the firm belongs. A simple way to do this is to use the average capital expenditure to depreciation ratio for the industry (or better still, just stable firms in the industry) to estimate a normalized capital expenditure for the firm. The danger of doing so is that the industry itself may not be steady state and the average will therefore be skewed (high or low). (b) Alternatively, you can use the relationship between growth and fundamentals to estimate the required reinvestment. The expected growth in net income can be written as:
This allows us to estimate the equity reinvestment rate:
To illustrate, a firm with a stable growth rate of 4 percent and a return on equity of 12 percent would need to reinvest about one-third of its net income back into net capital expenditures and working capital needs. Put differently, the free cash flows to equity should be two-thirds of net income.
Best Suited for Firms This model, like the stable growth dividend discount model, is best suited for firms growing at a rate comparable to or lower than the nominal growth in the economy. It is, however, a better model to use than the dividend discount model for stable firms that pay out dividends that are unsustainably high (because they exceed FCFE by a significant amount) or are significantly lower than the FCFE. Note, though, that if the firm is stable and pays out its FCFE as dividends the value obtained from this model will be the same as the one obtained from the Gordon growth model.
ILLUSTRATION 14.2: Stable Growth FCFE Model—Volkswagen Volkswagen is a mature German automobile company. Notwithstanding the cyclical swings in net income that are characteristic of the business, the firm is assumed to be in stable growth, and the following inputs were used to value it in May 2011: 1. The net income, not including the interest income from cash, for the company in 2010 was 5,279 million euros, and we will use this as the base year income. (We did check the level to see if it was an outlier, in either direction. If it had been, we would have use a normalized value.) 2. The expected growth in net income over time is assumed to be 3% and the noncash return on equity that Volkswagen is expected to deliver is 10%. The resulting equity reinvestment rate for the stable growth model is 30%:
255
The firm did report capital expenditures of 11,462 million euros, depreciation of 10,089 million euros, and an increase in noncash working capital of 423 million euros in 2010. The reinvestment rate using those inputs was 20.41%,
We could have used this reinvestment rate in the valuation, but with an expected growth rate in perpetuity of 2.04%:
3. Volkswagen's cost of equity is estimated using a beta of 1.20, reflecting the average beta across European auto companies, a euro risk-free rate of 3.2%, and an equity risk premium of 5%:
With the inputs, we can estimate the overall value of equity:
Note that this is the value of the equity in the non-cash operating assets, since we took out the income from cash from our base FCFE. Adding the cash balance of 18,670 million euros yields an overall value of equity of 80,062 million errors, significantly higher than the market capitalization of 53,560 million euros in May 2011.
FCFEst.xls: This spreadsheet allows you to value the equity in a firm in stable growth, with all of the inputs of a stable growth firm.
LEVERAGE, FCFE, AND EQUITY VALUE Embedded in the FCFE computation seems to be the makings of a free lunch. Increasing the debt ratio increases free cash flow to equity because more of a firm's reinvestment needs will come from borrowing and less is needed from equity investors. The released cash can be paid out as additional dividends or used for stock buybacks. If the free cash flow to equity increases as the leverage increases, does it follow that the value of equity will also increase with leverage? Not necessarily. The discount rate used is the cost of equity, which is estimated based on a beta or betas. As leverage increases, the beta will also increase, pushing up the cost of equity. In fact, in the levered beta equation that we introduced in Chapter 8 the levered beta is:
This, in turn, will have a negative effect on equity value. The net effect on value will then depend on which effect—the increase in cash flows or the increase in betas—dominates.
A TROUBLESHOOTING GUIDE: WHAT IS WRONG WITH THIS VALUATION? (CONSTANT GROWTH FCFE MODEL) If This Is Your Problem This May Be the Solution If you get a low value from this model, it may be because: Capital expenditures are too high relative to depreciation.
Use a smaller cap ex or use the two- stage model.
Working capital as a percent of revenues is too high.
Normalize this ratio, using historical averages.
The beta is high for a stable firm.
Use a beta closer to 1.
If you get too high a value, it is because: Capital expenditures are lower than depreciation.
Estimate an appropriate reinvestment rate = g/ROE.
Working capital ratio as percent of revenue is negative.
Set equal to zero.
The expected growth rate is too high for a stable firm.
Use a growth rate less than or equal to GNP growth (risk-free rate).
Two-Stage FCFE Model The two-stage FCFE model is designed to value a firm that is expected to grow much faster than a stable firm in the initial period and at a stable rate after that.
The Model 256
The value of any stock is the present value of the FCFE per year for the extraordinary growth period plus the present value of the terminal price at the end of the period.
where FCFEt = Free cash flow to equity in year t Pn = Price at the end of the extraordinary growth period ke = Cost of equity in high growth (hg) and stable growth (st) periods The terminal price is generally calculated using the infinite growth rate model:
where gn = Growth rate after the terminal year forever
Calculating the Terminal Price The same caveats that apply to the growth rate for the stable growth rate model, described in the previous section, apply here as well. In addition, the assumptions made to derive the free cash flow to equity after the terminal year have to be consistent with this assumption of stability. For instance, while capital spending may be much greater than depreciation in the initial high-growth phase, the difference should narrow as the firm enters its stable growth phase. We can use the two approaches described for the stable growth model—industry average capital expenditure requirements or the fundamental growth equation (equity reinvestment rate = g/ROE)—to make this estimate. The beta and debt ratio may also need to be adjusted in stable growth to reflect the fact that stable growth firms tend to have average risk (betas closer to 1) and use more debt than high-growth firms.
ILLUSTRATION 14.3: Capital Expenditure, Depreciation, and Growth Rates Assume you have a firm that is expected to have earnings growth of 20% for the next five years and 5% thereafter. The current earnings per share is $2.50. Current capital spending is $2.00, and current depreciation is $1.00. If we assume that capital spending and depreciation grow at the same rate as earnings and there are no working capital requirements or debt:
If we use the perpetual growth rate model, but fail to adjust the imbalance between capital expenditures and depreciation, the free cash flow to equity in the terminal year is:
This free cash flow to equity can then be used to compute the value per share at the end of year 5, but it will understate the true value. There are two ways in which you can adjust for this: 1. Adjust capital expenditures in year 6 to reflect industry average capital expenditure needs: Assume, for instance, that capital expenditures are 150% of depreciation for the industry in which the firm operates. You could compute the capital expenditures in year 6 as follows (EPS in year 6 = $6.21 × 1.05 = $6.53):
2. Estimate the equity reinvestment rate in year 6, based on expected growth and the firm's return on equity. For instance, if we assume that this firm's return on equity will be 15% in stable growth, the equity reinvestment rate would need to be:
257
Firms Model Works Best For This model makes the same assumptions about growth as the two-stage dividend discount model (i.e., that growth will be high and constant in the initial period and drop abruptly to stable growth after that). It is different because of its emphasis on FCFE rather than dividends. Consequently, it provides much better results than the dividend discount model when valuing firms which either have dividends which are unsustainable (because they are higher than FCFE) or which pay less in dividends than they can afford to (i.e., dividends are less than FCFE).
ILLUSTRATION 14.4: Two-Stage FCFE Model: Nestlé in 2001 Nestlé has operations all over the world, with 97% of its revenues coming from markets outside Switzerland, where it is headquartered. The firm, like many large European corporations, has a weak corporate governance system, and stockholders have little power over managers. RATIONALE FOR USING THE MODEL
Why two-stage? Nestlé has a long and impressive history of growth, and while we believe that its growth will be moderate, we assume that it will be able to maintain high growth for 10 years. Why FCFE? Given its weak corporate governance structure and a history of accumulating cash, the dividends paid by Nestlé bear little resemblance to what the firm could have paid out. BACKGROUND INFORMATION Current net income = Sfr 5,763 million
Earnings per share = Sfr 148.33
Current capital spending = Sfr 5,058 million Capital expenditures/share = Sfr 130.18 Current depreciation = Sfr 3,330 million
Depreciation/share = Sfr 85.71
Current revenues = Sfr 81,422 million
Revenue/share = Sfr 2,095.64
Noncash working capital = Sfr 5,818 million Working capital/share = Sfr 149.74 Change in working capital = Sfr 368 million Change in working capital/share = Sfr 9.47 Net debt issues = Sfr 272 million ESTIMATES We will begin by estimating the cost of equity for Nestlé during the high growth period in Swiss francs. We will use the 10-year Swiss government Sfr bond rate of 4% as the risk-free rate. To estimate the risk premium, we used the breakdown of Nestlé's revenues by region:
The risk premiums for each region represent an average of the risk premiums of the countries in the region. Using a bottom-up beta of 0.85 for Nestlé, we estimated a cost of equity of:
We will assume that this cost of equity will remain unchanged in perpetuity. To estimate the expected growth rate in free cash flows to equity, we first computed the free cash flows to equity in the current year:
The equity reinvestment rate can be estimated from this value:
The return on equity in 2000 was estimated using the net income from 2000 and the book value of equity from the end of the previous year:
The expected growth rate in FCFE is a product of the equity reinvestment rate and the return on equity:
We will assume that net capital expenditures and working capital will grow at the same rate as earnings for the high growth period and that the firm will raise 33.92% of its reinvestment needs from debt (which is its current book value debt-to-capital ratio). In stable growth, we assume a growth rate of 4%. We also assume that the cost of equity remains unchanged but that the return on equity
258
drops to 15%. The equity reinvestment rate in stable growth can be estimated as follows:
VALUATION The first component of value is the present value of the expected FCFE during the high-growth period, (see table) assuming earnings, net capital expenditures, and working capital grow at 7.27% and 33.92% of reinvestment needs come from debt (in Sfr):
Note that the change in working capital each year is computed based on the existing working capital of Sfr 149.74 per share, and that the present value is computed using the cost of equity of 8.47%. To estimate the terminal value, we first estimate the free cash flows to equity in year 11:
The value per share can be estimated as the sum of the present value of FCFE during the high growth phase and the present value of the terminal value of equity:
The stock was trading at 3,390 Sfr per share in May 2001 at the time of this valuation, making the stock slightly under valued.
FCFE2st.xls: This spreadsheet allows you to value a firm with a temporary period of high growth in FCFE, followed by stable growth.
REINVESTMENT ASSUMPTIONS, TERMINAL VALUE, AND EQUITY VALUE We have repeatedly emphasized the importance of linking growth assumptions to assumptions about reinvestment, and especially so in stable growth. A very common assumption in many discounted cash flow valuations is that capital expenditures offset depreciation in stable growth. When combined with the assumption of no working capital changes, this translates into zero reinvestment. While this may be a reasonable assumption for a year or two, it is not consistent with the assumption that operating income will grow in perpetuity. How much of a difference can one assumption make? In the Nestlé valuation, we reestimated terminal value of equity per share assuming no reinvestment:
Keeping all of our other assumptions intact, this results in a value of equity per share of 4,144 Sfr per share—an increase in value of approximately 22 percent.
E Model—A Three-Stage FCFE Model The E model is designed to value firms that are expected to go through three stages of growth—an initial phase of high growth rates, a transitional period in which the growth rate declines, and a steady-state period in which growth is stable.
The Model The E model calculates the present value of expected free cash flow to equity over all three stages of growth:
259
where P0 = Value of the stock today FCFEt = FCFE in year t ke = Cost of equity Pn2 = Terminal price at the end of transitional period = FCFEn2+1/(ke - gn) n1 = End of initial high growth period n2 = End of transition period
Caveats in Using Model Since the model assumes that the growth rate goes through three distinct phases—high growth, transitional growth, and stable growth—it is important that assumptions about other variables are consistent with these assumptions about growth.
Capital Spending versus Depreciation It is reasonable to assume that as the firm goes from high growth to stable growth, the relationship between capital spending and depreciation will change. In the high-growth phase, capital spending is likely to be much larger than depreciation. In the transitional phase, the difference is likely to narrow and the difference between capital spending and depreciation will be lower still in stable growth, reflecting the lower expected growth rate. (See Figure 14.3.) Figure 14.3 Three-Stage FCFE Model: Reinvestment Needs
A TROUBLESHOOTING GUIDE: WHAT IS WRONG WITH THIS VALUATION? (TWOSTAGE FCFE MODEL) If This Is Your Problem If you get a extremely low value from the two-stage FCFE, the likely culprits are:
This May Be the Solution
Earnings are depressed due to some reason (economy, etc.). Capital expenditures are significantly higher than depreciation in stable growth phase.
Use normalized earnings. Reduce the difference for stable growth period. (Compute the appropriate reinvestment rate—you might need a higher ROE.)
The beta in the stable period is too high for a stable firm.
Use a beta closer to 1.
Working capital as percent of revenue is too high to sustain.
Use a working capital ratio closer to industry.
260
The use of the two-stage model when the three-stage model is more appropriate. If you get an extremely high value:
Use a three-stage model.
Earnings are inflated above normal levels. Capital expenditures offset or lag are lower than depreciation during high growth period.
Use normalized earnings.
The growth rate in the stable growth period is too high for stable firm.
Use a growth rate closer to GNP growth (risk free rate).
Compute the appropriate reinvestment rate = g/ROE.
Risk As the growth characteristics of a firm change, so do its risk characteristics. In the context of the CAPM, as the growth rate declines the beta of the firm can be expected to change. The tendency of betas to converge toward one in the long term has been confirmed by empirical observation of portfolios of firms with high betas. Over time, as these firms get larger and more diversified, the average betas of these portfolios move toward 1.
Firms Model Works Best For Since the model allows for three stages of growth and for a gradual decline from high to stable growth, it is the appropriate model to use to value firms with very high growth rates currently. The assumptions about growth are similar to the ones made by the three-stage dividend discount model, but the focus is on FCFE instead of dividends, making it more suited to value firms whose dividends are significantly higher or lower than the FCFE.
ILLUSTRATION 14.5: Three-Stage FCFE Model: Tsingtao Breweries (China) in 2001 Tsingtao Breweries produces and distributes beer and other alcoholic beverages in China and around the world under the Tsingtao brand name. The firm has 653.15 million shares listed on the Shanghai and Hong Kong exchanges. RATIONALE FOR USING THE THREE-STAGE FCFE MODEL
Why three-stage? Tsingtao is a small firm serving a huge and growing market—China, in particular, and the rest of Asia in general. The firm's current return on equity is low, and we anticipate that it will improve over the next five years. As it increases, earnings growth will be pushed up. Why FCFE? Corporate governance in China tends to be weak and dividends are unlikely to reflect free cash flow to equity. In addition, the firm consistently funds a portion of its reinvestment needs with new debt issues. BACKGROUND INFORMATION In 2000, Tsingtao Breweries earned 72.36 million CY (Chinese yuan) in net income on a book value of equity of 2,588 million CY, giving it a return on equity of 2.80%. The firm had capital expenditures of 335 million CY and depreciation of 204 million CY during the year, and noncash working capital dropped by 1.2 million CY during the year. The total reinvestment in 2000 was therefore:
The working capital changes over the past four years have been volatile, and we normalize the change using noncash working capital as a percent of revenues in 2000:
The normalized reinvestment in 2000 can then be estimated as follows:
As with working capital, debt issues have been volatile. We estimate the firm's book debt to capital ratio of 40.94% at the end of 2000 and use it to estimate the normalized equity reinvestment in 2000:
As a percent of net income,
ESTIMATION To estimate free cash flows to equity for the high-growth period, we make the assumption that the return on equity, which is 2.80% today, will drift up to 12% by the fifth year. In addition, we will assume that new investments from now on will earn a return on equity of 12%. Finally, we will assume that the equity reinvestment rate will remain at its current level (149.97%) each year for the next five years. The expected growth rate over the next five years can then be estimated as follows:
261
After year 5, we will assume that the expected growth rate declines linearly each year from year 6 through 10 to reach a stable growth rate of 10% in year 10. (Note that the growth rate is in nominal CY; the higher stable growth rate reflects the higher expected inflation in that currency.) As the growth rate declines, the equity reinvestment rate also drops off to a stable period equity reinvestment rate of 50%, estimated using the 10% stable growth rate and an assumed return on equity in stable growth of 20%.
To estimate the cost of equity, we used a risk-free rate of 10% (in nominal CY), a risk premium of 6.28% (4% for mature market risk and 2.28% as the country risk premium for China), and a beta of 0.75 (reflecting the bottom-up beta for breweries):
In stable growth, we assume that the beta will drift up to 0.80 and that the country risk premium will drop to 0.95%:
The cost of equity adjusts in linear increments from 14.71% in year 5 to 13.96% in year 10. VALUATION To value Tsingtao, we will begin by projecting the free cash flows to equity during the high growth and transition phases, using an expected growth rate of 44.91% in net income and an equity reinvestment rate of 149.97% for the first five years. The next five years represent a transition period, where the growth drops in linear increments from 44.91% to 10% and the equity reinvestment rate drops from 149.97% to 50%. The resulting free cash flows to equity are shown in the following table:
To estimate the terminal value of equity, we use the net income in the year 11, reduce it by the equity reinvestment needs in that year, and then assume a perpetual growth rate to get to a value.
To estimate the value of equity today, we sum up the present value of the FCFE over the high growth period and add to it the present value of the terminal value of equity:
The stock was trading at 10.10 yuan per share, which would make it overvalued based on this valuation.
NEGATIVE FCFE, EQUITY DILUTION, AND VALUE PER SHARE Unlike dividends, free cash flows to equity can be negative. This can occur either because net income is negative or because a firm's reinvestment needs are significant; this is the case with Tsingtao in Illustration 14.5. The resulting net capital expenditure and working capital needs are much larger than the net income. In fact, this is likely to occur fairly frequently with high-growth firms. The FCFE model is flexible enough to deal with this issue. The free cash flows to equity will be negative as the firm reinvests substantial amounts to generate high growth. As the growth declines, the reinvestment needs also drop off and free cash flows to equity turn positive. Intuitively, though, consider what a negative free cash flow to equity implies. It indicates that the firm does not generate enough equity cash flows from current operations to meet its equity reinvestment needs. Since the free cash flow to equity is after net debt issues, the firm will have to issue new equity in years when the cash flow is negative. This expected dilution in future years will reduce the value of equity per share
262
today. In the FCFE model, the negative free cash flows to equity in the earlier years will reduce the estimated value of equity today. Thus the dilution effect is captured in the present value, and no additional consideration is needed of new stock issues in future years and the effect on value per share today.
A TROUBLESHOOTING GUIDE: WHAT IS WRONG WITH THIS VALUATION? (THREESTAGE FCFE MODEL) If This Is Your Problem This May Be the Solution If you get an extremely low value from the three-stage FCFE, the likely culprits are: Capital expenditures are significantly higher than depreciation in stable Reduce net cap ex in stable growth. Cap ex grows slower than growth phase. depreciation during transition period. The beta in the stable period is too high for a stable firm.
Use a beta closer to 1.
Working capital as percent of revenue is too high to sustain. If you get an extremely high value:
Use working capital ratio closer to industry average.
Capital expenditures offset depreciation during high-growth period.
Capital expenditures should be set higher.
Capital expenditures are less than depreciation.
Calculate reinvestment rate = g/ROC
Growth period (high growth and transition) is too long.
Use a shorter growth period. Use a growth rate lower than or equal to GNP growth (riskfree rate).
The growth rate in the stable growth period is too high for stable firm.
FCFE3st.xls: This spreadsheet allows you to value a firm with a temporary period of high growth in FCFE, followed by a transition period, followed by stable growth.
FCFE VALUATION VERSUS DIVIDEND DISCOUNT MODEL VALUATION The discounted cash flow model that uses FCFE can be viewed as an alternative to the dividend discount model. Since the two approaches sometimes provide different estimates of value, it is worth examining when they provide similar estimates of value, when they provide different estimates of value, and what the difference tells us about the firm.
When They Are Similar There are two conditions under which the value from using the FCFE in discounted cash flow valuation will be the same as the value obtained from using the dividend discount model. The first is the obvious one, where the dividends are equal to the FCFE. The second condition is more subtle, where the FCFE is greater than dividends, but the excess cash (FCFE minus dividends) is invested in projects with net present value of zero. (For instance, investing in financial assets that are fairly priced should yield a net present value of zero.)
When They Are Different There are several cases where the two models will provide different estimates of value. First, when the FCFE is greater than the dividend and the excess cash either earns below-market interest rates or is invested in negative net present value projects, the value from the FCFE model will be greater than the value from the dividend discount model. There is reason to believe that this is not as unusual as it would seem at first glance. There are numerous case studies of firms that, having accumulated large cash balances, by paying out low dividends relative to FCFE, have chosen to use this cash to finance unwise takeovers (where the price paid is greater than the value received from the takeover). Second, the payment of smaller dividends than can be afforded to be paid out by a firm may lead to a lower debt ratio and a higher cost of capital, causing a loss in value. In the cases where dividends are greater than FCFE, the firm will have to issue either new stock or new debt to pay these dividends leading to at least three negative consequences for value. One is the issuance cost on these security issues, which can be substantial for equity issues, creates an unnecessary expenditure that decreases value. Second, if the firm borrows the money to pay the dividends, the firm may become overlevered (relative to the optimal), exposing itself to distress/default and leading to a loss in value. Finally, paying too much in dividends can lead to capital rationing constraints where good projects are rejected, resulting in a loss of value. There is a third possibility and it reflects different assumptions about reinvestment and growth in the two models. If the same growth rate is used in the dividend discount and FCFE models, the FCFE model will give a higher value than the dividend discount model whenever FCFE is higher than dividends and a lower value when dividends exceed FCFE. In reality, the growth rate in FCFE should be different from the growth rate in dividends, because the free cash flow to equity is assumed to be paid out to stockholders. This will affect the reinvestment rate of the firm. In 263
addition, the return on equity used in the FCFE model should reflect the return on equity on noncash investments, whereas the return on equity used in the dividend discount model should be the overall return on equity. Table 14.1 summarizes the differences in assumptions between the two models. Table 14.1 Differences between DDM and FCFE Models Dividend Discount Model Only dividends are paid. Remaining portions of earnings are invested back into the firm, some in operating assets and some in cash and marketable securities. Measures growth in income from both operating Expected growth and cash assets. In terms of fundamentals, it is the product of the retention ratio and the return on equity. Dealing with cash and marketable securities The income from cash and marketable securities is built into earnings and ultimately into dividends. Therefore, cash and marketable securities do not need to be added in. Implicit assumption
FCFE Model The FCFE is paid out to stockholders. The remaining earnings are invested only in operating assets. Measures growth only in income from operating assets. In terms of fundamentals, it is the product of the equity reinvestment rate and the noncash return on equity. You have two choices: 1. Build income from cash and marketable securities into projections of income, and estimate the value of equity. 2. Ignore income from cash and marketable securities, and add their value to equity value in model.
In general, when firms pay out much less in dividends than they have available in FCFE, the expected growth rate and terminal value will be higher in the dividend discount model, but the year-to-year cash flows will be higher in the FCFE model. The net effect on value will vary from company to company.
What Does It Mean When They Are Different? When the value using the FCFE model is different from the value using the dividend discount model, with consistent growth assumptions, there are two questions that need to be addressed: What does the difference between the two models tell us? Which of the two models is the appropriate one to use in evaluating the market price? The less frequent scenario is that the dividend discount model yields a higher value than the FCFE model, largely because dividends exceed FCFE. In this case, it is best to go with the FCFE model because the dividends are not sustainable. The more common occurrence is for the value from the FCFE model to exceed the value from the dividend discount model. The difference between the value from the FCFE model and the value using the dividend discount model can be considered one component of the value of controlling a firm—it measures the value of controlling dividend policy. In a hostile takeover, the bidder can expect to control the firm and change the dividend policy (to reflect FCFE), thus capturing the higher FCFE value. As for which of the two values is the more appropriate one for use in evaluating the market price, the answer lies in the openness of the market for corporate control. If there is a sizable probability that a firm can be taken over or its management changed, the market price will reflect that likelihood, and the appropriate benchmark to use is the value from the FCFE model. As changes in corporate control become more difficult because of a firm's size and/or legal or market restrictions on takeovers, the value from the dividend discount model will provide the appropriate benchmark for comparison.
ILLUSTRATION 14.6: Valuing Coca-Cola with a Three-Stage FCFE Model In Chapter 13, we valued Coca-Cola using a three-stage dividend discount model and estimated a value of $67.15 per share, a tad under the market price of $68.22. Implicitly, we were assuming that Coca-Cola's managers are paying out what they can afford to in dividends and that there will be no cash buildup in the company. To test the proposition, we will now value Coca-Cola using a three-stage FCFE model. We begin by first separating the after-tax interest income of $105.32 million earned by Coca-Cola from its net income of $11,809 million and computing a noncash net income:
To compute the noncash return on equity, we modify the return on equity computation by netting cash ($7.021 million) out of book value of equity ($25,346 million):
As in the three-stage dividend discount model, we assume a much lower noncash return on equity of 30% going forward (which is higher than the 25% return on equity we assumed in the dividend discount model, because of our exclusion of cash). Rather than use the retention ratio to estimate reinvestment, we computed an equity reinvestment rate (ERR), using the net capital expenditures and working capital investment made by Coca-Cola in 2010:
264
Since Coca-Cola has done acquisitions over time, its equity reinvestment rate has been volatile, and the average over the past five years has been approximately 25%. Assuming a noncash ROE of 30% and an equity reinvestment rate of 25% for the next five years yields an expected growth rate in noncash net income of 7.5%:
During this high growth period, we will assume that Coca-Cola will have a beta of 0.90 (the same beta that we used in the dividend discount model). With a risk-free rate of 3.5% and an equity risk premium of 5.5%, we arrive at the same cost of equity of 8.45% that we used in the dividend discount model:
Applying the growth to net income and netting out the equity reinvestment rate yields the free cash flows to equity for the next five years, with present values computed using the cost of equity:
As in the dividend discount model, we will assume that the firm will be in stable growth 10 years forward, with the following assumptions: The growth rate in perpetuity after year 10 will be 3% and the noncash ROE is expected to be 15%. The resulting equity reinvestment rate in stable growth is:
The beta for the company will rise to 1, causing the cost of equity to increase to 9%. The transition period from year 6 to 10 is again used to adjust the growth rate, equity reinvestment rate, and cost of equity from high growth values to stable growth levels. The resulting FCFE are shown in the following table, with the present value computed using the cumulated cost of equity (since the cost of equity changes from period to period).
At the end of year 10, the terminal value of equity is computed using the stable growth inputs:
Discounting the terminal value back at the cumulated cost of equity for year 10 and adding to the present value of FCFE over the next 10 years yields an overall value for equity from operating assets. Note, though, that cash has been set apart in this model and needs to be added back at this stage to get to the value of equity for Coca-Cola.
Dividing by the number of shares outstanding today (2,289.254 million) yields a value per share of $95.54.
The FCFE model suggests that Coca-Cola is significantly undervalued at $68.22 a share.
So, why are they different? The FCFE model does use a lower growth rate than the dividend discount model, but it 265
allows for more cash to be returned to stockholders. In effect, we are incorporating the fact that Coca-Cola does not pay out its FCFE as dividends. The net effect, at least in this case, is an increase in value per share. For companies that pay out more dividends than they have available in FCFE, the value per share might drop with the FCFE model. In either case, we would argue that the FCFE estimate of value per share is a more realistic one than the dividend discount model estimate of value per share.
CONCLUSION The primary difference between the dividend discount models described in the previous chapter and the free cash flow to equity models described in this one lies in the definition of cash flows; the dividend discount model uses a strict definition of cash flow to equity (i.e., the expected dividends on the stock), while the FCFE model uses an expansive definition of cash flow to equity as the residual cash flow after meeting all financial obligations and investment needs. When dividends are different from the FCFE, the values from the two models will be different. In valuing firms for takeovers or in valuing firms where there is a reasonable chance of changing corporate control, the value from the FCFE model provides the better estimate of value.
QUESTIONS AND SHORT PROBLEMS In the problems following, use an equity risk premium of 5.5 percent if none is specified. 1. Respond true or false to the following statements relating to the calculation and use of FCFE: a. The free cash flow to equity will generally be more volatile than dividends. True ____ False ____ b. The free cash flow to equity will always be higher than dividends. True ____ False ____ c. The free cash flow to equity will always be higher than net income. True ____ False ____ d. The fre e cash flow to equity can never be negative. True ____ False ____ 2. Kimberly-Clark, a household product manufacturer, reported earnings per share of $3.20 in 1993 and paid dividends per share of $1.70 in that year. The firm reported depreciation of $315 million in 1993 and capital expenditures of $475 million. (There were 160 million shares outstanding, trading at $51 per share.) This ratio of capital expenditures to depreciation is expected to be maintained in the long term. The working capital needs are negligible. Kimberly-Clark had debt outstanding of $1.6 billion, and intended to maintain its current financing mix (of debt and equity) to finance future investment needs. The firm was in steady state and earnings were expected to grow 7% a year. The stock had a beta of 1.05. (The Treasury bond rate was 6.25%, and the risk premium was 5.5%.) a. Estimate the value per share, using the dividend discount model. b. Estimate the value per share, using the FCFE model. c. How would you explain the difference between the two models, and which one would you use as your benchmark for comparison to the market price? 3. Ecolab Inc. sells chemicals and systems for cleaning, sanitizing, and maintenance. It reported earnings per share of $2.35 in 1993, and expected earnings growth of 15.5% a year from 1994 to 1998 and 6% a year after that. The capital expenditure per share was $2.25, and depreciation was $1.125 per share in 1993. Both were expected to grow at the same rate as earnings from 1994 to 1998. Working capital was expected to remain at 5% of revenues, and revenues, which were $1 billion in 1993, were expected to increase 6% a year from 1994 to 1998, and 4% a year after that. The firm had has a debt ratio [D/(D + E)] of 5%, but planned to finance future investment needs (including working capital investments) using a debt ratio of 20%. The stock was expected to have a beta of 1 for the period of the analysis, and the Treasury bond rate was 6.50%. (There were 63 million shares outstanding, and the market risk premium was 5.5%.) a. Assuming that capital expenditures and depreciation offset each other after 1998, estimate the value per share. Is this a realistic estimate? b. Assuming that capital expenditures continue to be 200% of depreciation even after 1998, estimate the value per share. c. What would the value per share have been, if the firm had continued to finance new investments with its old 266
financing mix (5%)? Is it fair to use the same beta for this analysis? 4. Dionex Corporation, a leader in the development and manufacture of ion chromography systems (used to identify contaminants in electronic devices), reported earnings per share of $2.02 in 1993, and paid no dividends. These earnings were expected to grow 14% a year for five years (1994 to 1998) and 7% a year after that. The firm reported depreciation of $2 million in 1993 and capital spending of $4.20 million, and had 7 million shares outstanding. The working capital was expected to remain at 50% of revenues, which were $106 million in 1993, and were expected to grow 6% a year from 1994 to 1998 and 4% a year after that. The firm was expected to finance 10% of its capital expenditures and working capital needs with debt. Dionex had a beta of 1.20 in 1993, and this beta was expected to drop to 1.10 after 1998. (The Treasury bond rate was 7%, and the market risk premium was 5.5%.) a. Estimate the expected free cash flow to equity from 1994 to 1998, assuming that capital expenditures and depreciation grow at the same rate as earnings. b. Estimate the terminal price per share (at the end of 1998). Stable firms in this industry have capital expenditures that are 150% of revenues, and maintain working capital at 25% of revenues. c. Estimate the value per share today, based on the FCFE model. 5. Biomet Inc., which designs, manufactures, and markets reconstructive and trauma devices, reported earnings per share of $0.56 in 1993, on which it paid no dividends (it had revenues per share in 1993 of $2.91). It had capital expenditures of $0.13 per share in 1993, and depreciation in the same year of $0.08 per share. The working capital was 60% of revenues in 1993 and was expected to remain at that level from 1994 to 1998, while earnings and revenues were expected to grow 17% a year. The earnings growth rate was expected to decline linearly over the following five years to a rate of 5% in 2003. During the high growth and transition periods, capital spending and depreciation were expected to grow at the same rate as earnings, but capital spending would be 120% of depreciation when the firm reaches steady state. Working capital was expected to drop from 60% of revenues during the 1994–1998 period to 30% of revenues after 2003. The firm had no debt currently, but planned to finance 10% of its net capital investment and working capital requirements with debt. The stock was expected to have a beta of 1.45 for the high growth period (1994 to 1998), and the beta was expected to decline to 1.10 by the time the firm goes into steady state (in 2003). The Treasury bond rate is 7%, and the market risk premium is 5.5%. a. Estimate the value per share, using the FCFE model. b. Estimate the value per share, assuming that working capital stays at 60% of revenues forever. c. Estimate the value per share, assuming that the beta remains unchanged at 1.45 forever. 6. Will the following firms be likely to have a higher value from the dividend discount model, a higher value from the FCFE model, or the same value from both models? a. A firm that pays out less in dividends than it has available in FCFE, but invests the balance in Treasury bonds. b. A firm that pays out more in dividends than it has available in FCFE, and then issues stock to cover the difference. c. A firm that pays out, on average, its FCFE as dividends. d. A firm that pays out less in dividends that it has available in FCFE, but uses the cash at regular intervals to acquire other firms with the intent of diversifying. e. A firm that pays out more in dividends than it has available in FCFE, but borrows money to cover the difference. (The firm is overlevered to begin with.) 7. You have been asked to value Oneida Steel, a midsize steel company. The firm reported $80 million in net income, $50 million in capital expenditures, and $20 million in depreciation in the just-completed financial year. The firm reported that its noncash working capital increased by $20 million during the year and that total debt outstanding increased by $10 million during the year. The book value of equity at Oneida Steel at the beginning of the last financial year was $400 million. The cost of equity is 10%. a. Estimate the equity reinvestment rate, return on equity, and expected growth rate for Oneida Steel. (You can assume that the firm will continue to maintain the same debt ratio that it used last year to finance its reinvestment needs.) b. If this growth rate is expected to last five years and then drop to a 4% stable growth rate after that and the return on equity after year 5 is expected to be 12%, estimate the value of equity today, using the projected free 267
cash flows to equity. 8. Luminos Corporation, a manufacturer of lightbulbs, is a firm in stable growth. The firm reported net income of $100 million on a book value of equity of $1 billion. However, the firm also had a cash balance of $200 million on which it earned after-tax interest income of $10 million last year. (This interest income is included in the net income, and the cash is part of the book value of equity.) The cost of equity for the firm is 9%. a. Estimate the noncash return on equity at Luminos Corporation. b. If you expect the cash flows from the operating assets of Luminos to increase 3% a year in perpetuity, estimate the value of equity at Luminos. 1
The mix has to be fixed in book value terms. It can be varying in market value terms.
268
CHAPTER 15 Firm Valuation: Cost of Capital and Adjusted Present Value Approaches The preceding two chapters examined two approaches to valuing the equity in the firm—the dividend discount model and the free cash flow to equity (FCFE) valuation model. This chapter examines approaches to valuation in which the entire firm is valued, by either discounting the cumulated cash flows to all claim holders in the firm by the weighted average cost of capital (the cost of capital approach) or by adding the marginal impact of debt on value to the unlevered firm value—the adjusted present value (APV) approach. In the process of looking at firm valuation, we also look at how leverage may or may not affect firm value. We note that in the presence of default risk, taxes, and agency costs, increasing leverage c an sometimes increase firm value and sometimes decrease it. In fact, we argue that the optimal financing mix for a firm is the one that maximizes firm value.
FREE CASH FLOW TO THE FIRM The free cash flow to the firm (FCFF) is the sum of the cash flows to all claim holders in the firm, including common stockholders, bondholders, and preferred stockholders. There are two ways of measuring the free cash flow to the firm. One is to add up the cash flows to the claim holders, which would include cash flows to equity (defined either as free cash flow to equity or as dividends); cash flows to lenders (which would include principal payments, interest expenses, and new debt issues); and cash flows to preferred stockholders (usually preferred dividends):
Note, however, that we are reversing the process that we used to get to free cash flow to equity, where we subtracted out payments to lenders and preferred stockholders to estimate the cash flow left for stockholders. A simpler way of getting to free cash flow to the firm is to estimate the cash flows prior to any of these claims. Thus we could begin with the e arnings before interest and taxes, net out taxes and reinvestment needs, and arrive at an estimate of the free cash flow to the firm:
Since this cash flow is prior to debt payments, it is often referred to as an unlevered cash flow. Note that this free cash flow to the firm does not incorporate any of the tax benefits due to interest payments. This is by design, because the use of the after-tax cost of debt in the cost of capital already considers this benefit, and including it in the cash flows would double count it.
FCFF and Other Cash Flow Measures The differences between FCFF and FCFE arise primarily from cash flows associated with debt—interest payments, principal repayments, and new debt issues—and other nonequity claims, such as preferred dividends. For firms at their desired debt level, which finance their capital expenditures and working capital needs with this mix of debt and equity and use new debt issues to finance principal repayments, the free cash flow to the firm will exceed the free cash flow to equity. One metric that is widely used in valuation is the earnings before interest, taxes, depreciation, and amortization (EBITDA), a rough measure of cash flows form operations. The free cash flow to the firm is a related concept but it is more complete because it takes into account the potential tax liability from the earnings as well as capital expenditures and working capital requirements. Some analysts also use after-tax operating income as a proxy for free cash flow to the firm, with alternative definitions of operating income. The first, earnings before interest and taxes (EBIT) or operating income, comes directly from a firm's income statements. Adjustments to EBIT yield the net operating profit or loss after taxes (NOPLAT) or the net operating income (NOI). The net operating income is defined to be the income from operations prior to taxes and nonoperating expenses. Each of these measures is used in valuation models, and each can be related to the free cash flow to the firm. Each, however, makes some assumptions about the relationship between depreciation and capital expenditures that are made explicit in Table 15.1. Table 15.1 Free Cash Flows to the Firm: Comparison to Other Measures 269
Cash Flow Used
Definition
FCFF
Free cash flow to firm
FCFE
EBITDA
EBIT (1 – t) (NOPLAT is a slightly modified version of this estimate and it removes any nonoperating items that might affect the reported EBIT.)
Use in Valuation Discounting free cash flow to the firm at the cost of capital will yield the value of the operating assets of the firm. To this, you would add on the value of nonoperating assets to arrive at firm value. FCFF - Interest (1 - t) - Principal repaid + New Discounting free cash flows to equity at the cost of equity will yield the value of equity in a debt issued - Preferred dividend business. FCFF + EBIT(t) + Capital expenditures + If you discount EBITDA at the cost of capital to value an asset, you are assuming that there are Change in working capital no taxes and that the firm will actively disinvest over time. It would be inconsistent to assume a growth rate or an infinite life for this firm. FCFF + Capital expenditures Depreciation + If you discount after-tax operating income at the cost of capital to value a firm, you are Change in working capital assuming no reinvestment. The depreciation is reinvested back into the firm to maintain existing assets. You can assume an infinite life but no growth.
Growth in FCFE versus Growth in FCFF Will equity cash flows and firm cash flows grow at the same rate? Consider the starting point for the two cash flows. Equity cash flows are based on net income or earnings per share—measures of equity income. Firm cash flows are based on operating income (i.e., income prior to debt payments). As a general rule, you would expect growth in operating income to be lower than growth in net income, because financial leverage can augment the latter. To see why, let us go back to the fundamental growth equations laid out in Chapter 11:
We also defined the return on equity in terms of the return on capital:
When a firm borrows money and invests in projects that earn more than the after-tax cost of debt, the return on equity will be higher than the return on capital. This, in turn, will translate into a higher growth rate in equity income at least in the short term. In stable growth, though, the growth rates in equity income and operating income have to converge. To see why, assume that you have a firm whose revenues and operating income are growing at 5 percent a year forever. If you assume that the same firm's net income grows at 6 percent a year forever, the net income will catch up with operating income at some point in time in the future and exceed revenues at a later point in time. In stable growth, therefore, even if return on equity exceeds the return on capital, the expected growth will be the same in all measures of income.1
FIRM VALUATION: THE COST OF CAPITAL APPROACH The value of the firm is obtained by discounting the free cash flow to the firm at the weighted average cost of capital. Embedded in this value are the tax benefits of debt (in the use of the after-tax cost of debt in the cost of capital) and expected additional risk associated with debt (in the form of higher costs of equity and debt at higher debt ratios). Just as with the dividend discount model and the FCFE model, the version of the model used will depend on assumptions made about future growth.
Stable Growth Firm As with the dividend discount and FCFE models, a firm that is growing at a rate that it can sustain in perpetuity—a stable growth rate—can be valued using a stable growth model.
The Model A firm with free cash fl ows to the firm growing at a stable growth rate can be valued using the following equation:
270
where FCFF1 = Expected FCFF next year WACC = Weighted average cost of capital gn = Growth rate in the FCFF forever
The Caveats There are two conditions that need to be met in using this model. First, the growth rate used in the model has to be less than or equal to the growth rate in the economy—nominal growth, if the cost of capital is in nominal terms, or real growth, if the cost of capital is a real cost of capital. Second, the characteristics of the firm have to be consistent with assumptions of stable growth. In particular, the reinvestment rate used to estimate free cash flows to the firm should be consistent with the stable growth rate. The best way of enforcing this consistency is to derive the reinvestment rate from the stable growth rate:
If reinvestment is estimated from net capital expenditures and change in working capital, the net capital expenditures should be similar to those other firms in the industry (perhaps by setting the ratio of capital expenditures to depreciation at industry averages) and the change in working capital should generally not be negative. A negative change in working capital creates a cash inflow, and while this may, in fact, be viable for a firm in the short term, it is dangerous to assume it in perpetuity.2 Even if industry averages are used to compute the reinvestment, it is always prudent to estimate what return on capital is imputed in that reinvestment (obtained by dividing the growth rate in perpetuity by the reinvestment rate). The cost of capital should also be reflective of a stable growth firm. In particular, the beta should be close to 1—the rule of thumb presented in the earlier chapters that the beta should be between 0.8 and 1.2 still holds. While stable growth firms tend to use more debt, this is not a prerequisite for the model, since debt policy is subject to managerial discretion.
Limitations Like all stable growth models, this one is sensitive to assumptions about the expected growth rate. This is accentuated, however, by the fact that the discount rate used in valuation is the WACC, which is significantly lower than the cost of equity for most firms. So, if keeping the growth rate below the risk free rate was good practice with equity valuation models, it is even more so with firm valuation. Furthermore, the model is sensitive to assumptions made about capital expenditures relative to depreciation. As n oted in chapter 12, if the inputs for reinvestment are not a function of expected growth the free cash flow to the firm can be inflated (or deflated) by reducing (increasing) capital expenditures relative to depreciation.
ILLUSTRATION 15.1: Valuing a Firm with the Stable Growth FCFF Model—Telesp (Brazil) Telesp provides local telecommunication services to the Brazilian state of Sao Paulo. In 2010, the company had operating income (EBIT) of 3,544 million BR and faced an effective tax rate of 30%. In 2010, the firm reported capital expenditures of 1,659 million BR and depreciation of 1,914 million BR and an increase in working capital of 1,119 million BR. Consequently, its reinvestment in 2010 can be computed as follows:
The return on capital generated by the company in 2010 was computed using the operating income for the year and the book value of capital invested at the end of the previous year (2009):
The expected growth rate that emerges from these inputs is:
271
While this would be too high a growth rate for stable growth in a currency with low expected inflation, the risk-free rate in BR in May 2011 was 7%. In conjunction with a beta of 0.8 and an equity risk premium for Brazil of 8% (composed of a mature market premium of 5% and an additional country risk premium of 3% for Brazil), this yields a cost of equity of 13.40%. Incorporating a pretax cost of debt of 9.50% and a debt ratio of 20% (based on current market values for equity and debt) results in a cost of capital of 12.05% for Telesp:
The value for the operating assets can then be estimated as follows:
Adding the cash and marketable securities (1,557 million BR) and subtracting the debt (5,519 million BR) at the end of 2010 yields a value for the equity:
The company's market capitalization in May 2011 was 21,982 million BR, making it fairly priced.
General Version of the FCFF Model Rather than break the free cash flow model into two-stage and three-stage models and risk repeating what was said in the preceding chapter, we present the general version of the model in this section. We follow up by examining a range of companies—a traditional manufacturing firm (Gerdau Steel), a firm with operating leases (Target), and a firm with substantial R&D investments (Amgen)—to illustrate the differences and similarities between this approach and the FCFE approach.
The Model The value of the firm, in the most general case, can be written as the present value of expected free cash flows to the firm:
where FCFFt = Free cash flow to firm in year t WACC = Weighted average cost of capital
MARKET VALUE WEIGHTS, COST OF CAPITAL, AND CIRCULAR REASONING To value a firm, you first need to estimate a cost of capital. Every textbook is categorical that the weights in the cost of capital calculation be market value weights. The problem, however, is that the cost of capital is then used to estimate new values for debt and equity that might not match the values used in the original calculation. One defense that can be offered for this inconsistency is that if you bought all of the debt and equity in a publicly traded firm, you would pay current market value and not your estimated value, and your cost of capital reflects this. For those who are bothered by this inconsistency, there is a way out. You could do a conventional valuation using market value weights for debt and equity, but then use the estimated values of debt and equity from the valuation to reestimate the cost of capital. This, of course, will change the values again, but you could feed the new values back and estimate cost of capital again. Each time you do this, the differences between the values you use for the weights and the values you estimate will narrow, and the values will converge sooner rather than later. How much of a difference will it make in your ultimate value? The greater the difference between market value and your estimates of value, the greater the difference this iterative process will make. In the valuation of Telesp, we began with a market value of 21,982 million BR and estimated a value of 21,939 million BR. If we substituted back this estimated value and iterated to a solution, we would arrive at an estimate of value of 21,946 million BR.3
If the firm reaches steady state after n years and starts growing at a stable growth rate gn after that, the value of the firm can be written as: 272
where WACC = Cost of capital (hg: high growth; st: stable growth)
Firms Model Best Suited For Firms that either have very high or very low leverage or are in the process of changing their leverage are best valued using the FCFF approach. The calculation of FCFE is much more difficult in these cases because of the volatility induced by debt payments (or new issues), and the value of equity, which can a small slice of the total value of the firm for highly levered firms, is more sensitive to assumptions about growth and risk. It is worth noting, though, that in theory the two approaches should yield the same value for the equ ity. Getting them to agree in practice is an entirely different challenge and we will return to examine it later in this chapter.
Problems There are three problems that we see with the free cash flow to the firm model. The first is that the free cash flows to equity are a much more intuitive measure of cash flows than cash flows to the firm. When asked to estimate cash flows, most of us look at cash flows after debt payments (free cash flows to equity), because we tend to think like business owners and consider interest payments and the repayment of debt as cash outflows. Furthermore, the free cash flow to equity is a real cash flow that can be traced and analyzed in a firm. The free cash flow to the firm is the answer to a hypothetical question: What would this firm's cash flow be if it had no debt (and associated payments)? The second is that its focus on predebt cash flows can sometimes blind us to real problems with survival. To illustrate, assume that a firm has free cash flows to the firm of $100 million but that its large debt load makes its free cash flows to equity equal to –$50 million. This firm will have to raise $50 million in new equity to survive, and if it cannot, all cash flows beyond this point are put in jeopardy. Using free cash flows to equity would have alerted you to this problem, but free cash flows to the firm are unlikely to reflect this. The final problem is that the use of a debt ratio in the cost of capital to incorporate the effect of leverage requires us to make implicit assumptions that might not be feasible or reasonable. For instance, assuming that the market value debt ratio is 30 percent will require a growing firm to issue large amounts of debt in future years to reach that ratio. In the process, the book-to-debt ratio might reach stratospheric proportions and trigger covenants or other negative consequences. In fact, we count the expected tax benefits from future debt issues implicitly in the value of equity today.
ILLUSTRATION 15.2: Valuing Target—Dealing with Operating Leases In 2010, Target reported $5,252 million in pretax operating income on revenues of $67,390 million. While its high growth days are behind it, there is some potential for growth, and we will attempt to value the firm using a two-stage FCFF model. The first step in this valuation is to recognize that the financial statement numbers for Target are skewed by the failure to consider lease commitments as debt. Using the annual report for 2010, we obtained the lease commitments for the next five years and beyond, which we discount at Target's pretax cost of debt of 4.5% (estimated based on its S&P bond rating of A) to convert the commitments to debt:
Year
Commitment Present Value @ 4.5%
1
$190.00
$ 181.82
2
$189.00
$ 173.07
3
$187.00
$ 163.87
4
$147.00
$ 123.27
5
$141.00
$ 113.15
6–23
$172.22
$1,680.51
Debt value of leases =
$2,435.68
Note that Target reported a lump sum of $3,100 million for commitments beyond year 5, which we have converted into annual commitments of $172.22 million a year for 18 years (a judgment call based on the annual average commitment for years 1–5). We will adjust the stated debt and operating income to reflect the decision to treat lease commitments as debt:
To estimate the expected growth rate, we estimate the return on capital and reinvestment rate for Target in 2010, again staying true to the decision to capitalize leases:
273
Note that we computed the present value of lease commitments at the end of 2009 by going back to the annual report for that year, extracting the lease commitments, and computing the present value of the commitments using the pretax cost of debt at the end of 2009. Target pulled back on reinvestment in 2010, but we expect the reinvestment rate to bounce back to 40% (close to the average for the past five years) in the next five years, yielding an expected growth rate of 4.30% each year for that period:
To compute the cost of capital over this period, we estimate a beta of 1.05 for Target (based on the average beta across general retailers) and use an equity risk premium of 5% (the risk-free rate is 3.5%):
Here again, we computed debt to capital ratios, with operating leases treated as part of debt, and the market capitalization for Target of $34,346 million. The resulting free cash flows to the firm for the following five years are reported in the table, with the present value computed using the cost of capital:
At the end of year 5, we assume that Target will be a mature firm, with a growth rate of 3% in perpetuity and a return on capital equal to its cost of capital. The resulting reinvestment rate and terminal value are estimated in the following calculations:
Adding the present value of the terminal value to the sum of the present value of the free cash flows to the firm for the next five years, we arrive at the value of the operating assets:
Adding the cash balance ($1,712 million) and subtracting debt inclusive of the operating leases ($18,162 million) yields a value of equity of $40,636 million. Dividing by the number of shares (689.13 million) results in a value per share of $58.97, about 20% higher than the prevailing market price of $49 in May 2011. As a final part of the analysis, we examine the effect that treating leases as debt has on the valuation. As the following table makes clear, staying with the current accounting treatment of operating leases as operating expenses would result in a higher return on capital, a higher cost of capital, and a slightly higher value of equity per share. Operating income
Operating Expense Financial Expense $5,346.00 $5,252.00
274
Debt
$16,814.00
$19,250.00
ROIC
11.39%
10.75%
Reinvestment rate
40%
40%
Expected growth rate 4.56%
4.30%
Debt-to-capital ratio
31.41%
34.59%
Cost of capital
6.92%
6.74%
Value of firm
$56,731.00
$58,795.00
Value of equity
$41,005.00
$40,633.00
Value/share
$59.50
$58.97
While the value per share effect is small in the case of Target, it will be larger for firms with more substantial lease commitments (relative to conventional debt). A key number to track is the excess return (return on capital – cost of capital) earned by the firm. For Target, converting leases to debt lowers the excess return slightly from 4.47% (11.39% minus 6.92%) to 4.01% (10.75% minus 6.74%), which also lowers the value per share. The greater the change in the excess returns from the lease adjustment, the greater will be the impact of converting leases to debt on value per share.
ILLUSTRATION 15.3: Valuing Amgen in March 2009: The Effect of R&D Capitalization In Illustration 9.2, we used Amgen to illustrate the effects of capitalizing R&D, using a 10-year amortizable life for R&D. Using data through 2008, we estimated the capital invested in R&D and the amortization as follows:
Using the financial statements from 2008, we compute the adjusted operating income and return on capital at the firm, using Amgen's effective tax rate of 20% in 2008.
Note that the capitalized R&D used in the return on capital computation was based on the R&D expenses through 2007 and that the adjusted after-tax earnings reflect the tax benefits of R&D expensing. We used the restated numbers to estimate the value of the firm and equity per share. The valuation, where we assume 10 years of high growth, is summarized in Figure 15.1.
Figure 15.1 Valuing Amgen—March 2009
275
The transition period, as in the prior chapter, exists primarily so as to allow us to adjust our high growth inputs to stable growth levels. The cost of capital for instance, which is 11.23% for the next five years, drops in linear increments to the stable growth cost of capital of 8.23%; the compounded cost of capital is therefore used to discount cash flows in those years. Our estimate of value of equity per share is $67.16 a share, well above the prevailing stock price of $47.47 in March 2009. An intriguing question is how the capitalization of R&D expenses affected value. To investigate, we compare the valuation fundamentals for Amgen, with conventional accounting, and with R&D treated as capital ex penses in the following table:
Valuation Fundamentals—With and Without R&D Capitalization Conventional Capitalized R&D After-tax ROC
20.44%
17.17%
Reinvestment rate 14.47%
34.13%
Growth rate
2.96%
5.86%
Value per share
$48.24
$67.17
We then revalue the firm, using both sets of fundamentals. As the table indicates, the value per share would have been $48.24 if we had used conventional accounting numbers. Clearly, capitalization matters, and the degree to which it matters will vary across firms. In general, the effect will be negative for firms that invest large amounts in R&D, with little to show (yet) in terms of earnings and cash flows in subsequent periods. It will be positive for firms that reinvest large amounts in R&D and report large increases in earnings in subsequent periods. In the case of Amgen, capitalizing R&D has a positive effect on value per share, because of its track record of successful R&D.
ILLUSTRATION 15.4: Valuing an Emerging Market Company with Developed Market Exposure: Gerdau Steel (Brazil) in March 2009 Gerdau Steel is a Brazilian steel company that derived about 51% of its revenues in Brazil in 2008 and the rest in North America. We chose to value Gerdau Steel in U.S. dollars, partly because of the difficulties we face in estimating risk-free rates and risk premiums in Brazilian reais (R $). To estimate the cost of capital in U.S. dollar terms, we start with the U.S. Treasury bond rate of 3%. In March 2009, the equity risk premium that we were using for mature markets (like the United States) was 6% and the additional country risk premium for Brazil was 4.75%. For Gerdau Steel, we use the average unlevered beta of 1.01 for steel companies listed globally, using the argument that steel is a commodity that is bought and sold on a world market. Since Gerdau has a very high market debt to equity ratio (138.89%), the resulting levered beta is 1.94 (with 34% being the marginal tax rate for Brazil):
To reflect the fact that Gerdau Steel derives almost half its revenues in emerging markets, we estimated a lambda to measure exposure to Brazilian country risk, using two approaches: 1. Revenue-based approach: Dividing Gerdau's Brazilian revenue proportion (51 percent) by the average revenue proportion for a Brazilian company (72 percent) yields a lambda of 0.79. 2. Price-based approach: Regressing the weekly returns on Gerdau stock, between January 2007 and January 2009, on the weekly returns on the Brazilian government dollar-denominated bond yields a lambda of 0.625:
We use the latter estimate to compute a US$ cost of equity for Gerdau Steel of 17.61%:
To estimate the cost of debt for Gerdau, we began with the interest coverage ratio for the firm, using the 2008 income statement:
276
This interest coverage ratio, in conjunction with Table 8.1 (from Chapter 8), yields a rating of A– and a default spread of 3% (based on March 2009 spreads). Adding the default spread for Brazil (3%) at the time, we get a pretax cost of debt of 9% for Gerdau:
Finally, incorporating Gerdau's current market debt to capital ratio of 58.45%, we estimate a US$ cost of capital of 10.79%:
We use the 2008 financial statements and exchange rates at the time of the statements to estimate the cashflows in R$ and then convert these cash flows to U.S. dollars.
Base year numbers: In the 2008 financial year, Gerdau reported operating income of R$ 8,005 million, after depreciation of R$1,896 million. During the year, acquisitions and internal investments combined to create capital expenditures of R$6,818 million and noncash working capital increased by R$1,083 million. Gerdau earned an after-tax return on capital of 18.68%, based on a marginal tax rate (for Brazil) of 34%, and start-of-the-year book value for equity of R$17,449 million, book value of debt of R$ 15,979 million and a cash balance of R$ 5,139 million:
Forecasted growth and cash flows: We do not believe that either the return on capital or the reinvestment rate is sustainable in the long term. Consequently, we use a reinvestment rate of 60% and a return on capital of 16% to estimate the expected growth rate of 9.60%, in R$, for the next five years.
We use this expected growth rate to estimate expected cash flows for the next five years, in R$, in the following table:
Again, the reinvestment each year is the consolidated value of net capital expenditures, acquisitions, and investments in working capital, and amounts to 60% of after-tax operating income each year.
Conversion to U.S. dollars: To convert the cash flows in R$ to U.S. dollars, we start with the prevailing exchange rate (in March 2009) of R$ 2.252/$ but forecast exchange rates for future years based on expected inflation rates of 2% in U.S. dollars and 5% in BR. The resulting expected exchange rates and cash flows in U.S. dollars are reported in the following table:
The difference in expected inflation results in R$ depreciating in value relative to the U.S. dollar over the five-year period.
Stable growth: In stable growth, we assume that Gerdau will grow 3% a year, in dollar terms, and that its return on capital in stable growth will converge on its cost of capital (also in dollar terms). To estimate the dollar cost of capital in stable growth, we assume that the stock will have a beta of 1.20 and that the country risk premium will decline to 3%. Using a debt ratio of 50% and a cost of debt of 8%, we estimate a cost of capital of 8.68%.4 To estimate the terminal value, we first compute the after-tax operating income in dollar terms in year 5:
We then compute the reinvestment rate and terminal value:
277
Firm and equity valuation: To complete the analysis, we first discount the expected cash flows in US dollars at the cost of capital of 10.79%:
To get to firm value, we add in dollar value of the cash holdings of the firm ($2,404 million) and subtract out the dollar value of debt ($9.788 million), with the conversion at today's exchange rate. Since Gerdau ha s consolidated holdings, we subtract out the estimated market value of the minority interest in these holdings of $2,599 million (in dollar terms) and then divide by the number of shares outstanding (1,681.12 million) to arrive at a dollar value per share of $10.12:5
Converted at the exchange rate of 2.252 R$/$, we arrive at an estimate of value of R$ 22.79/ share, making it significantly undervalued at the price of R$9.32/share at which it was trading in March 2009. As with the prior two valuations, it is worth exploring the effect of the choice we made to value Gerdau Steel in U.S. dollars. We could have valued Gerdau Steel in BR by adjusting the U.S. dollar cost of capital for differential inflation:
Making a similar adjustment to the stable period cost of capital yields a BR cost of capital of 11.89%. Finally, we adjust the stable growth rate to reflect the higher inflation rate in BR:
The terminal value in $R can then be estimated:
The value of Gerdau Steel's operating assets in BR can then be computed by discounting the BR cash flows back at the high growth period BR cost of capital of 14.05%:
Converting the value into US$ at the prevailing exchange rate of 2.252 yields a dollar value for the operating assets of $26,315 million, very close to our dollar-based estimate of $26,996 million.
278
fcffginzu.xls: This spreadsheet allows you to estimate the value of a firm using the FCFF approach.
NET DEBT VERSUS GROSS DEBT In valuing the companies in this chapter, we used total debt outstanding (gross debt) rather than net debt where cash was netted out against debt. What is the difference between the two approaches, and will the valuations from the two approaches agree? A comparison of gross and net debt valuations reveals the differences in the way we approach the calculation of key inputs to the valuation, summarized as follows: Levered beta Cost of capital
Gross Debt Unlevered beta is levered using gross debt to market equity ratio.
Debt-to-capital ratio used is based on gross debt. Treatment of cash and debt Cash is added to value of operating assets and gross debt is subtracted to get to equity value.
Net Debt Unlevered beta is levered using net debt to market equity ratio. Debt-to-capital ratio used is based on net debt. Cash is not added back to operating assets and net debt is subtracted to get to equity value.
While working with net debt in valuation is not difficult to do, the more interesting question is whether the value that emerges will be the same as the value that would have been estimated using gross debt. In general the answer is no, and the reason usually lies in the cost of debt used in the net debt valuation. Intuitively, what you are doing when you use net debt is break the firm into two parts—a cash business, which is funded 100 percent with riskless debt, and an operating business funded partly with risky debt. Carrying this to its logical conclusion, the cost of debt you would have for the operating business would be significantly higher than the firm's current cost of debt. This is because the current lenders to the firm will factor in the firm's cash holdings when setting the cost of debt. To illustrate, assume that you have a firm with an overall value of $1 billion—$200 million in cash and $800 million in operating assets—with $400 million in debt and $600 million in equity. The firm's cost of debt is 7 percent, a 2 percent default spread over the risk-free rate of 5 percent; note that this cost of debt is set based on the firm's substantial cash holdings. If you net debt against cash, the firm would have $200 million in net debt and $600 million in equity. If you use the 7 percent cost of debt to value the firm now, you will overstate its value. Instead, the cost of debt you should use in the valuation is 9 percent:
In general, we would recommend using gross debt rather than net debt for two other reasons. First, the net debt can be a negative number if cash exceeds the gross debt. If this occurs, you should set the net debt to zero and consider the excess cash just as you would cash in a gross debt valuation. Second, maintaining a stable net debt ratio in a growing firm will require that cash balances increase as the firm value increases.
Will Equity Value Be the Same under Firm and Equity Valuation? This model, unlike the dividend discount model or the FCFE model, values the firm rather than equity. The value of equity, however, can be extracted from the value of the firm by subtracting the market value of outstanding debt. Since this model can be viewed as an alternative way of valuing equity, two questions arise: Why value the firm rather than equity? Will the values for equity obtained from the firm valuation approach be consistent with the values obtained from the equity valuation approaches described in the previous chapter? The advantage of using the firm valuation approach is that cash flows relating to debt do not have to be considered explicitly since the FCFF is a predebt cash flow, while they have to be taken into account in estimating FCFE. In cases where the leverage is expected to change significantly over time, this is a significant time saver, since estimating new debt issues and debt repayments when leverage is changing can become increasingly messy the further into the future you go. The firm valuation approach does, however, require information about debt ratios and interest rates to estimate the weighted average cost of capital. The value for equity obtained from the firm valuation and equity valuation approaches will be the same if you make consistent assumptions about financial leverage. Getting them to converge in practice is much more difficult. Let us begin with the simplest case—a no-growth, perpetual firm. Assume that the firm has $166.67 million in earnings before interest and taxes and a tax rate of 40 percent. Assume that the firm has equity with a market value of $600 million, with a cost of equity of 13.87 percent, and debt of $400 million, with a pretax cost of debt of 7 percent. The firm's cost of capital can be estimated as follows:
Note that the firm has no reinvestment and no growth. We can value equity in this firm by subtracting the value of debt:
279
Now let us value the equity directly by estimating the net income:
The value of equity can be obtained by discounting this net income at the cost of equity:
Even this simple example works because of the following three assumptions made implicitly or explicitly during the valuation: 1. The values for debt and equity used to compute the cost of capital were equal to the values obtained in the valuation. Notwithstanding the circularity in reasoning—you need the cost of capital to obtain the values in the first place—it indicates that a cost of capital based on market value weights will not yield the same value for equity as an equity valuation model, if the firm is not fairly priced in the first place. 2. There are no extraordinary or nonoperating items that affect net income but not operating income. Thus, to get from operating to net income all we do is subtract interest expenses and taxes. 3. The interest expenses are equal to the pretax cost of debt multiplied by the market value of debt. If a firm has old debt on its books, with interest expenses that are different from this value, the two approaches will diverge. If there is expected growth, the potential for inconsistency multiplies. You have to ensure that you borrow enough money to fund new investments to keep your debt ratio at a level consistent with what you are assuming when you compute the cost of capital.
fcffvsfcfe.xls: This spreadsheet allows you to compare the equity values obtained using FCFF and FCFE models.
FIRM VALUATION: THE ADJUSTED PRESENT VALUE APPROACH The adjusted present value (APV) approach begins with the value of the firm without debt. As debt is added to the firm, the net effect on value is examined by considering both the benefits and the costs of borrowing. To do this, it is assumed that the primary benefit of borrowing is a tax benefit, and that the most significant cost of borrowing is the added risk of bankruptcy.
Mechanics of APV Valuation We estimate the value of the firm in three steps: 1. Estimate the value of the firm with no leverage. 2. Consider the present value of the interest tax savings generated by borrowing a given amount of money. 3. Evaluate the effect of borrowing the amount on the probability that the firm will go bankrupt, and the expected cost of bankruptcy.
Value of Unlevered Firm The first step in this approach is the estimation of the value of the unlevered firm. This can be accomplished by valuing the firm as if it had no debt (i.e., by discounting the expected free cash flow to the firm at the unlevered cost of equity). In the special case where cash flows grow at a constant rate in perpetuity,
where FCFF1 is the expected after-tax operating cash flow to the firm, ρu is the unlevered cost of equity, and g is the expected growth rate. In the more general case, you can value the firm using any set of growth assumptions you believe are reasonable for the firm. The inputs needed for this valuation are the expected cash flows, growth rates, and the unlevered cost of equity. To estimate the unlevered cost of equity, we can draw on our earlier analysis and compute the unlevered beta of the firm:
280
where βunlevered = Unlevered beta of the firm βcurrent = Current equity beta of the firm t = Tax rate for the firm D/E = Current debt/equity ratio This unlevered beta can then be used to arrive at the unlevered cost of equity.
Expected Tax Benefit from Borrowing The second step in this approach is the calculation of the expected tax benefit from a given level of debt. This tax benefit is a function of the tax rate and interest payments of the firm and is discounted at the cost of debt to reflect the riskiness of this cash flow. If the tax savings are viewed as a perpetuity,
The tax rate used here is the firm's marginal tax rate, and it is assumed to stay constant over time. If you anticipate the tax rate changing over time, you can still compute the present value of tax benefits over time, but you cannot use the perpetual growth equation cited earlier. In addition, you would have to modify this equation if the current interest expenses do not reflect the current cost of debt.
Estimating Expected Bankruptcy Costs and Net Effect The third step is to evaluate the effect of the given level of debt on the default risk of the firm and on expected bankruptcy costs. In theory, at least, this requires the estimation of the probability of default with the additional debt and the direct and indirect cost of bankruptcy. If πa is the probability of default after the additional debt and BC is the present value of the bankruptcy cost, the present value (PV) of expected bankruptcy cost can be estimated:
This step of the adjusted present value approach poses the most significant estimation problems, since neither the probability of bankruptcy nor the bankruptcy cost can be estimated directly. There are two basic ways in which the probability of bankruptcy can be estimated indirectly. One is to estimate a bond rating and use the empirical estimates of default probabilities for the rating. For instance, Table 15.2, extracted from a study by Altman, summarizes the probability of default over 10 years by bond rating class in using the 1999 to 2008 time period.6 Table 15.2 Ratings and Probability of Default Source: Altman (2009).
Rating Probability of Default AAA
0.07%
AA
0.51%
A+
0.60%
A
0.66%
A–
2.50%
BBB
7.54%
BB+
10.00%
BB
16.63%
B+
25.00%
B
36.80%
B–
45.00%
CCC
59.01%
CC
70.00%
C
85.00%
D
100.00%
281
The other way is to use a statistical approach such as a probit to estimate the probability of default, based on the firm's observable characteristics, at each level of debt. The bankruptcy cost can be estimated, albeit with considerable error, from studies that have looked at the magnitude of this cost in actual bankruptcies. Research that has looked at the direct costs of bankruptcy concludes that they are small7 relative to firm value. The indirect costs of bankruptcy can be substantial, but the costs vary widely across firms. Shapiro (1989) and Titman (1984) speculate that the indirect costs could be as large as 25 to 30 percent of firm value but provide no direct evidence of the costs.
ILLUSTRATION 15.5: Valuing a Company Using APV: The Leveraged Acquisition of J. Crew J. Crew is a U.S. retailer that sells clothes made under its brand name through its own stores and online. In 2010, the firm was acquired in a leveraged deal by Mickey Drexler, its CEO, and two private equity firms—TPG and Leonard Green—for $2.7 billion, with about $1.85 billion coming from debt (with a rating of BB and a pretax cost of debt of 7%). To assess the value of the deal using the APV approach, we first value the firm as an all-equity funded (unlevered) firm. To estimate the value, we first computed a cost of equity using an unlevered beta of 1.00 for specialty retailers, in conjunction with a risk-free rate of 3.5% and mature market premium of 5%:
J. Crew generated $230 million in operating income on revenues of $1,722 million in 2010. We assume a 35% tax rate and a growth rate of 3.5% in perpetuity, with a return on capital of 14%, resulting in the following:
To estimate the tax benefits from debt, we assume that a debt schedule by which the dollar debt would be repaid in equal annual increments to a debt level of $500 million in year 10 and beyond. Using the 35% tax rate and the pretax cost of debt, we compute the interest expenses and tax benefits each year, and discount these benefits back to today using the pretax cost of debt as the discount rate.
Note that the value of tax benefits in perpetuity is computed in two steps. First, we compute the present value of $12.25 million in tax savings in perpetuity ($12.25/.07 = $175 million). Next, we discount that value back to today at the pretax cost of debt ($175/1.0710 = $88.96 million) As the final piece of the analysis, we assume that bankruptcy costs (BC), direct and indirect, would amount to 30% of firm value and that the high debt level taken in the deal increases the probability of bankruptcy (πBC) to 20%. The expected bankruptcy cost is then:
The value for J. Crew can now be computed using all three components:
At $2.7 billion, the private equity investors are paying too much for the firm unless they can increase operating income substantially.
Cost of Capital versus APV Valuation 282
In an APV valuation, the value of a levered firm is obtained by adding the net effect of debt to the unlevered firm value.
In the cost of capital approach, the effects of leverage show up in the cost of capital, with the tax benefit incorporated in the after-tax cost of debt and the bankruptcy costs in both the levered beta and the pretax cost of debt. Will the two approaches yield the same value? Not necessarily. The first reason for differences is that the models consider bankruptcy costs very differently, with the adjusted present value approach providing more flexibility in allowing you to consider indirect bankruptcy costs. To the extent that these costs do not show up or show up inadequately in the pretax cost of debt, the APV approach will yield a more conservative estimate of value. The second reason is that the APV approach considers the tax benefit from a dollar debt value, usually based on existing debt. The cost of capital approach estimates the tax benefit from a debt ratio that may require the firm to borrow increasing amounts in the future. For instance, assuming a market debt-to-capital ratio of 30percent in perpetuity for a growing firm will require it to borrow more in the future, and the tax benefit from expected future borrowings is incorporated into value today. Generally speaking, the cost-of-capital approach is a more practical choice when valuing ongoing firms that are not going through contortions on financial leverage; it is easier to work with a debt ratio than with dollar-debt levels. The APV approach is more useful for transactions that are funded disproportionately with debt and where debt repayment schedules are negotiated or known; this is why it has acquired a footing in leveraged-buyout circles. Finally, there is a subtle distinction in how the tax benefits from debt are incorporated in value in the two approaches. While the conventional APV approach uses the pre-tax cost of debt as the discount rate to estimate the value of the tax savings from debt, there are variations on the APV that discount the tax savings back at the cost of capital or the unlevered cost of equity that yield values that are closer to those obtained in the cost of capital approach.
APV WITHOUT BANKRUPTCY COSTS There are many who believe that adjusted present value is a more flexible way of approaching valuation than traditional discounted cash flow models. This may be true in a generic sense, but APV valuation in practice has significant flaws. The first and most important is that most practitioners who use the adjusted present value model ignore expected bankruptcy costs. Adding the tax benefits to unlevered firm value to get to levered firm value makes debt seem like an unmixed blessing. Firm value will be overstated, especially at very high debt ratios, where the cost of bankruptcy is clearly not zero.
EFFECT OF LEVERAGE ON FIRM VALUE Both the cost of capital approach and the APV approach make the value of a firm a function of its leverage. It follows directly, then, that there is some mix of debt and equity at which firm value is maximized. The rest of this chapter considers how best to make this link.
Cost of Capital and Optimal Leverage In order to understand the relationship between the cost of capital and optimal capital structure, we rely on the relationship between firm value and the cost of capital. The earlier section noted that the value of the entire firm can be estimated by discounting the expected cash flows to the firm at the firm's cost of capital. The firm value can then be written as follows:
and is a function of the firm's cash flows and its cost of capital. If we assume that the cash flows to the firm are unaffected by the choi ce of financing mix, and the cost of capital is reduced as a consequence of changing the financing mix, the value of the firm will increase. If the objective in choosing the financing mix for the firm is the maximization of firm value, we can accomplish it, in this case, by minimizing the cost of capital. In the more general case where the cash flows to the firm are a function of the debt-equity mix, the optimal financing mix is the mix that maximizes firm value.8
ILLUSTRATION 15.6: WACC, Firm Value, and Leverage Assume that you are given the costs of equity and debt at different debt levels for Strunks Inc., a leading manufacturer of chocolates and other candies, and that the cash flows to this firm are currently $200 million. Strunks is in a relatively stable market, and these cash flows are expected to grow at 6% forever and to be unaffected by the debt ratio of the firm. The cost of capital schedule is provided in the following table, along with the value of the firm at each level of debt.
283
Note that:
The value of the firm increases as the cost of capital decreases, and decreases as the cost of capital increases. This is illustrated in Figure 15.2. While this illustration makes the choice of an optimal financing mix seem easy, it obscures problems that will arise in its practice. First, we typically do not have the benefit of having the entire schedule of costs of financing prior to an analysis. In most cases, the only level of debt at which we have information on the cost of debt and equity financing is the current level. Second, the analysis assumes implicitly that the level of operating income of the firm is unaffected by the financing mix of the firm and, consequently, by the default risk (or bond rating) for the firm. While this may be reasonable in some cases, it will not be in others. Firms that borrow too much might find that there are indirect bankruptcy costs that affect revenues and operating income.
Figure 15.2 Cost of Capital and Firm Value Source: Applied Corporate Finance, Third Edition, by Aswath Damodaran, copyright © 2010 by John Wiley & Sons, Inc. This material is used by permission of John Wiley & Sons, Inc.
Steps in Cost of Capital Approach We need three basic inputs to compute the cost of capital—the cost of equity, the after-tax cost of debt, and the weights on debt and equity. The costs of equity and debt change as the debt ratio changes, and the primary challenge of this approach is in estimating each of these inputs. Let us begin with the cost of equity. We argued that the beta of equity will change as the debt ratio changes. In fact, we estimated the levered beta as a function of the market debt to equity ratio of a firm, the unlevered beta, and the firm's marginal tax rate:
Thus, if we can estimate the unlevered beta for a firm, we can use it to estimate the levered beta of the firm at every debt ratio. This levered beta can then be used to compute the cost of equity at each debt ratio.
The cost of debt for a firm is a function of the firm's default risk. As firms borrow more, their default risk will increase and so will the cost of debt. If we use bond ratings as our measure of default risk, we can estimate the cost of debt in three steps. First, estimate a firm's dollar debt and interest expenses at each debt ratio; as firms increase their debt ratio, both dollar debt and interest expenses will rise. Second, at each debt level, compute a financial ratio or ratios that measure default risk and use the ratio(s) to estimate a rating for the firm; again, as firms borrow more, this rating will decline. Third, a default spread, based on the estimated rating, is added to the 284
risk-free rate to arrive at the pretax cost of debt. Applying the marginal tax rate to this pretax cost yields an aftertax cost of debt. Once we estimate the costs of equity and debt at each debt level, we weight them based on the proportions used of each to estimate the cost of capital. While we have not explicitly allowed for a preferred stock component in this process, we can have preferred stock as a part of capital. However, we have to keep the preferred stock portion fixed, while changing the weights on debt and equity. The debt ratio at which the cost of capital is minimized is the optimal debt ratio. In this approach, the effect on firm value of changing the capital structure is isolated by keeping the operating income fixed and varying only the cost of capital. In practical terms, this requires us to make two assumptions. First, the debt ratio is decreased by raising new equity and retiring debt; conversely, the debt ratio is increased by borrowing money and buying back stock. This process is called recapitalization. Second, the pretax operating income is assumed to be unaffected by the firm's financing mix and, by extension, its bond rating. If the operating income changes with a firm's default risk, the basic analysis will not change, but minimizing the cost of capital may not be the optimal course of action, since the value of the firm is determined by both the cash flows and the cost of capital. The value of the firm will have to be computed at each debt level and the optimal debt ratio will be the one that maximizes firm value.
ILLUSTRATION 15.7: Analyzing the Capital Structure for Disney: May 2009 The cost of capital approach can be used to find the optimal capital structure for a firm, as we will for Disney in May 2009. Disney had $14,962 million in interest-bearing debt on its books and adding the present value of operating lease commitments of $1,720 million to this value, we arrive at a total market value for the debt of $16,682 million. The market value of equity at the same time was $45,193 million; the market price per share was $24.34, and there were 1,856.752 million shares outstanding. Proportionally, 26.96% of the overall financing mix was debt, and the remaining 73.04% was equity. The unlevered beta for Disney's stock in May 2009, estimated by breaking it down into its constituent businesses and weighting the unlevered betas for each business, was 0.7333.
The Treasury bond rate at that time was 3.5%. Using an estimated equity risk premium of 6%, we estimated the cost of equity for Disney to be 8.91%:
Disney's bond rating in May 2009 was A, and based on this rating, the estimated pretax cost of debt for Disney is 6%. Using a marginal tax rate of 38%, we estimate the after-tax cost of debt for Disney to be 3.72%.
The cost of capital is calculated using these costs and the weights based on market value:
DISNEY'S COST OF EQUITY AND LEVERAGE The cost of equity for Disney at different debt ratios can be computed using the unlevered beta of the firm, and the debt equity ratio at each level of debt. We use the levered betas that emerge to estimate the cost of equity. The first step in this process is to compute the levered beta at each debt ratio, using this unlevered beta and Disney's marginal tax rate of 38%:
We continue to use the Treasury bond rate of 3.5% and the market premium of 6% to compute the cost of equity at each level of debt. If we keep the tax rate constant at 38%, we obtain the levered betas for Disney in the following table:
285
In calculating the levered beta in this table, we assumed that all market risk is borne by the equity investors (this may be unrealistic) especially at higher levels of debt and that the firm will be able to get the full tax benefits of interest expenses even at very high debt ratios. We will also consider an alternative estimate of levered betas that apportions some of the market risk to the debt:
The beta of debt can be based on the rating of the bond, estimated by regressing past returns on bonds in each rating class against returns on a market index or backed out of the default spread. The levered betas estimated using this approach will generally be lower than those estimated with the conventional model.9 We will also examine whether the full benefits of interest expenses will accrue at higher debt ratios. DISNEY'S COST OF DEBT AND LEVERAGE There are several financial ratios that are correlated with bond ratings, and we face two choices. One is to build a model that includes several financial ratios to estimate the synthetic ratings at each debt ratio. In addition to being more labor and data intensive, the approach will make the ratings process less transparent and more difficult to decipher. The other is to stick with the simplistic approach that we developed in Chapter 8, of linking the rating to the interest coverage ratio, with the ratio defined as:
We will stick with the simpler approach for three reasons. First, we are not aiming for precision in the cost of debt, but an approximation. Given that the more complex approaches also give approximations, we will tilt in favor of tra nsparency. Second, there is significant correlation not only between the interest coverage ratio and bond ratings but also between the interest coverage ratio and other ratios used in analysis, such as the debt coverage ratio and the funds flow ratios. In other words, we may be adding little by adding other ratios that are correlated with interest coverage ratios, including EBITDA/fixed charges, to the mix. Third, the interest coverage ratio changes as a firm changes its financing mix and decreases as the debt ratio increases, a key requirement since we need the cost of debt to change as the debt ratio changes. To make our estimates of the synthetic rating, we will use the lookup table that we introduced in Chapter 8 for large market capitalization firms (since Disney's market capitalization is greater than $5 billion) and use the default spreads from early 2009 to estimate the pre tax cost of debt. The following table reproduces those numbers:
Interest Coverage Ratios, Ratings and Default Spreads—Early 2009 Interest Rating Typical Default Spread Coverage Ratio >8.5
AAA
1.25%
6.5–8.5
AA
1.75%
5.5–6.5
A+
2.25%
4.25–5.5
A
2.50%
3.0–4.25
A–
3.00%
2.5–3.0
BBB
3.50%
2.25–2.5
BB+
4.25%
2.0–2.25
BB
5.00%
1.75–2.0
B+
6.00%
1.5–1.75
B
7.25%
1.25–1.5
B–
8.50%
0.8–1.25
CCC
10.00%
0.65–0.8
CC
12.00%
0.2–0.65
C
15.00%
