Please do not quote without author’s permission. Comments welcomed alrapp@san.rr.com

The Economics of Short-Term Performance Obsession

Alfred Rappaport*

September 16, 2004 Version

My thanks to Martha Amram, Jack Bogle, Bruce Johnson, David Larcker, Marty Leibowitz, Michael Mauboussin, Larry Siegel, Shyam Sunder, Allan Timmerman, Jack Treynor, and Linda Vincent for helpful discussions and comments.

* Leonard Spacek Professor Emeritus, Kellogg Graduate School of Management, Northwestern University

1

Abstract

According to economic theory discounted cash flows set prices in all well-functioning capital markets. In practice, investment managers attach substantial weight to short-term performance, particularly earnings, to select stocks. Analysts fixate on quarterly earnings while often neglecting the longer-term drivers of shareholder value. Corporate executives point to the behavior of the investment community to rationalize their own obsession with short-term accounting earnings. The fascination with earnings is not surprising. Investment managers fear that failure to achieve acceptable performance relative to a target benchmark might lead to large fund withdrawals and ultimately their dismissal. Consequently, they focus on short-term stock price performance (which they believe is strongly influence by quarterly earnings) and on controlling tracking error. CEOs and other senior corporate executives concerned with their reputations and the company’s stock price focus on publicly reported short-term performance measures, particularly earnings. Short-termism is the disease and earnings and tracking error the carriers.

The earnings accrual process estimates future cash flows from existing, but incomplete, contracts between the company and its customers (receivables, product warranties, unrealized gains or losses on long-term sales contracts), employees (defined-benefit pension plans and other postretirement benefits, stock options), suppliers (payables), and government (taxes, environmental obligations). The estimated present value of these accrual estimates typically represents less than 5 percent of the stock price and astute analysts view these estimates with understandable caution. Since the overwhelming majority of a company’s value depends on cash flows from future sales and purchase contracts, analysts must go beyond the current financials to find the preponderance of a company’s value.

The paper goes on to examine whether stock prices are likely to be allocatively efficient when short-term earnings, tracking error, and a host of non-economic criteria dominate investment decisions. The question then arises whether investment managers who identify stocks they believe to be mispriced on a discounted cash-flow basis can earn excess returns. The answer depends on two basic factors that shape returns: the size of the estimated mispricing and the time it takes for the stock price to converge

2

toward the target value. Underlying these factors lurks the ever-present risk that new information will trigger unfavorable price changes.

Corporate executives observe the high turnover in shares and the investment community’s obsession with earnings and often conclude that operating the company for the long-term is not rewarded by higher stock prices. They therefore exploit the discretion allowed in calculating earnings and forego or delay value-creating investments in order to meet quarterly earnings expectations. Both of these activities compromise shareholder value. The paper clarifies why maximizing long-term cash flow, even in an earnings-dominated market, is the most effective means of creating value for continuing shareholders.

The final section charts a course for alleviating the obsession with short-term performance and improving allocative efficiency. Recommended initiatives are outlined for corporate performance reporting, investment manager incentives, and corporate executive incentives.

3

THE ECONOMICS OF SHORT-TERM PERFORMANCE OBSESSION

A company’s value depends upon its long-term ability to generate cash to fund value-creating growth and pay dividends to its shareholders. Even so, investment managers commonly base their stock selections on short-term earnings and portfolio tracking error rather than discounted cash flow--the standard for valuing financial assets in well-functioning capital markets. Financial analysts fixate on quarterly earnings at the expense of fundamental research. Corporate executives, in turn, point to the behavior of the investment community to rationalize their own obsession with earnings. Short-termism is the disease--earnings and tracking error are the carriers.

This gap between theory and practice prompts four basic questions: Why do investment managers focus on quarterly earnings? Can stock prices be allocatively efficient when short-term

earnings and tracking error dominate investment decisions? Can investment managers earn excess returns if they buy and

sell stocks they believe the market has mispriced on a discounted cash-flow basis?

Is corporate management’s focus on short-term earnings self-serving or also in the best interests of its shareholders?

These questions, which are addressed in the first four sections of the paper, focus on why investment and corporate managers are obsessed with short-term performance and the implications for allocative efficiency in capital and corporate markets. The final section presents a three-pronged program--improving corporate performance reporting, incentives for investment managers, and incentives for corporate managers—for reducing short-term performance obsession.

Why Do Investment Managers Focus on Quarterly Earnings?

The limitations of earnings

The accountant’s bottom line neither approximates a company’s value nor its change in value over the reporting period. Nor was it ever intended to. Valuation is the investor’s job, not the purpose of financial reporting.1

4

Earnings are relevant to valuation to the extent they help investors and analysts estimate the magnitude, timing, and uncertainty of future cash flows. But two factors severely limit the usefulness of earnings for forecasting cash flows. First, companies manage earnings because they have considerable latitude in estimating the amount and timing of accruals such as for restructuring, employee pension costs, stock option grants and sometimes even revenue. Second, and much more significant, accruals deal with only a small fraction of the cash flows investors need to value stocks. This critical but generally ignored limitation calls for a brief explanation.

Earnings are an amalgam of facts (realized cash flows) and assumptions about future outcomes (accruals). The cash-flow portion of earnings consists of the cash a company receives for current-period sales minus the cash it disburses to suppliers and employees for products and services used during the period. Revenue and expense accruals (excluding arbitrary depreciation and amortization charges) reflect the company’s estimates of subsequent-period cash receipts and payments that will arise from current-period sales and purchase transactions, respectively. Contracts between the company and its customers (receivables, unrealized gains or losses on long-term sales contracts, product warranties), employees (defined-benefit pension plans and other postretirement benefits, stock options), suppliers (payables), and government (taxes, environmental obligations) govern the amounts that companies record.

The crucial point is that accruals encompass only existing, incomplete contracts while the overwhelming majority of a company’s value derives from cash flows attributable to future sales and purchase contracts. The estimated present value of existing contracts typically account for less than 5 percent of a company’s share price and astute analysts view that amount with understandable caution given management’s considerable latitude in establishing accrual amounts.2 Not only do revenue and expense accruals convey information about a relatively small fraction of a company’s cash flow value, the earnings figure which combines realized cash flows and uncertain accruals masks even this limited information. For example, a positive earnings surprise does not necessarily signal an increase in value. Companies can boost their earnings without creating value not only with accounting shenanigans, but also by value-destroying underinvestment which investors can’t easily detect. Additionally, shareholder value increases only if a company earns a rate of return on its new investments that is greater than its cost of capital. Earnings, on the other hand, can increase not only when a company is

5

investing at above its cost of capital, but also when it is investing below its cost of capital.

To find the preponderance of a company’s value analysts must go beyond financial statements. To evaluate the sustainability and potential growth of sales and cash flow, they must weigh factors such as industry market growth potential, the company’s competitive position, the likely behavior of competitors, technological change, and quality of management.

The appeal of earnings

The fascination with quarterly earnings is not all that puzzling. In fact, it might be perfectly rational in a market dominated by agents responsible for other peoples’ money but also looking out for their own interests.

Most investment professionals recognize that discounted cash-flow is the appropriate model for valuing financial assets including equities. But they believe that estimating distant cash flows is too time-consuming, costly, and speculative. Because they have much less information about a company’s operations and prospects than insiders do, investors tend to attach substantial weight to short-term performance. Short-term performance is particularly significant for younger companies, where expectations about future growth are much more sensitive to current performance than for companies with established operating histories. Earnings-based decision making also frees investment managers and analysts from having to develop longer-term cash-flow forecasts.

CEOs and other senior corporate executives concerned with their reputations and the company’s stock price also focus on publicly reported short-term performance measures, particularly earnings. As a consequence, investment and corporate managers have a mutually-reinforcing obsession with short-term performance, and earnings the most widely accepted metric.

Why else are investors and corporate managers so fixated on earnings? Sizeable stock price responses to earnings surprises suggest that it’s short-term earnings rather than long-term cash-flow prospects that fuel prices. But it’s not clear whether prices respond mechanically to earnings announcements, to new information about longer-term prospects conveyed by components of earnings, to both or to neither.3 What is clear is that portfolio managers who were able to accurately and consistently forecast year-ahead earnings could have earned extraordinary returns.

Hagin4 examines the annual percent changes in earnings and annual shareholder returns for 800 widely-traded companies over the twenty-five

6

year period from 1977 to 2002. With remarkable consistency companies with the worst annual earnings changes have the worst returns and those with the best earnings changes have the best annual returns. Companies in the bottom quintile of annual earnings changes have a -15.2% average annual return below the average for all companies, while companies in the top quintile yielded an average annual excess return of 11.6%.

It’s easy for investors who observe sizeable price responses to earnings surprises to conclude that it is better to employ “irrational” earnings analysis than the “rational” discounted cash-flow model, which they view as theoretically valid but practically disconnected from expected returns. When there is greater risk in going against the market’s apparent pricing model than there is reward, it’s best to join the market. The cumulative effect of this thinking becomes a self-fulfilling prophecy.

The final factor that appears to favor short-term earnings over long-term cash flows is the relatively short holding period for stocks. The annual portfolio turnover in professionally-managed funds is now greater than 100 percent for an average holding period of less than a year. The average holding period was about seven years until the mid-1960s.5 The shorter the holding period the more the beliefs of others, rather than long-term fundamentals, become central to investment decisions. High turnover sets the stage for short-term earnings-based decision making or momentum-motivated trading that is not even concerned with earnings. Welcome to Keynes’ beauty contest.

Short-horizon investors expect to derive substantially all of their proceeds from selling shares at the end of their investment horizon and very little from cash dividends. With dividend yields averaging about 2 percent and assuming a one-year horizon, about 98 percent of total cash proceeds can be expected to come from selling shares. Without a dividend or cash-flow anchor short-horizon investors focus on forming an expectation about the end-of-horizon selling price. This expectation however depends on the impossible task of assessing the expectations of countless other investors with varying investment horizons and then backward inducting to price. Faced with this impossible task and under pressure to show acceptable short-term performance, investment managers turn to past and expected short-term metrics, particularly earnings, to project end-of horizon prices.6

Contrary to those who believe that stocks are priced on a company’s short-term outlook, most stocks require more than ten years of value-creating cash flows to justify their price. Therefore, knowingly or unknowingly short-horizon investors make bets on prices that reflect long-term company prospects.

7

Can Stock Prices Be Allocatively Efficient When Short-term Earnings and Tracking Error Dominate Investment Decisions?

The last section presented the reasons for the popularity of earnings. This section reviews the major models investors use for their buy and sell decisions. The purpose is to assess the relative efficiency of stock prices in allocating capital to companies with the most promising economic prospects. Misallocation of capital, whether by the equity market or by corporate executives, takes a toll on economic growth and shortchanges everyone, including investors, employees, and consumers.

Informational and fundamental efficiency

Behavioral economists distinguish between informational efficiency and fundamental efficiency.7 In an informationally efficient, or no “free-lunch,” market stock prices fully reflect all relevant information thus preventing investors from earning excess returns using available information. As evidence of informational efficiency, researchers point to the notable scarcity of investment strategies or professional money managers that outperform the market over long time periods.

But how can we reconcile the enormous amounts spent on investment research with an informationally efficient market?8 Grossman and Stiglitz argue that prices cannot perfectly reflect available information.9 Since research is costly, investors who expend resources to obtain information expect to receive some compensation in the form of excess returns. This logic holds in a world of economically rational individuals who invest their own funds--a world of principals without agents.

In the existing agent-dominated market, however, it is perfectly rational for active fund managers to incur costs, even when they face very long odds of achieving excess returns, as long as fund shareholders, not managers, bear the costs. The result is what I call “subsidized informational efficiency.” This sets on its head the conventional wisdom that informational efficiency depends on market participants disbelieving it. Paradoxically, active investment managers contribute to informational efficiency not by maximizing long-term returns but by closely tracking their benchmarks thereby constraining their ability to outperform the benchmarks.

Stock prices reflect information relevant to the models investors employ. Investment managers have little incentive to pursue private information that contributes to more allocatively-efficient prices unless such

8

information is also relevant to their decision models. In other words, active managers can produce an informationally-efficient market without necessarily making it highly allocatively-efficient.

How does fundamental efficiency differ from informational efficiency? In an informationally efficient market there are no free lunches, in a fundamentally efficient market “prices are right.” Fundamental efficiency is not an empirically refutable hypothesis because in a sea of uncertainty and heterogeneous beliefs, the right price is indeterminate. Black10 defines “an efficient market as one in which price is within a factor of two of value, i.e., the price is more than half of value and less than twice value.” But even this definition, with its permissive range for efficiency, is not operational as long as we can’t know the right price.

The spectrum of beliefs that exists not only about prospective cash flows, but also about the other essential determinant of value, the discount rate, exacerbates the trouble with fundamental efficiency. Benartzi and Thaler propose that the surprisingly large equity premiums over the past century are due to the combination of loss aversion (individuals are about twice as sensitive to losses as similar-sized gains) and the tendency of investors to evaluate their portfolios frequently, thereby behaving as if their investment time horizons are short.11 If Benartzi and Thaler are correct, investors with longer horizons using a lower discount rate would place higher values on stocks than investors with shorter horizons.

The right price for a stock is not only unknowable today, we cannot determine it at a later date because future prices will not be based on today’s information, but on revised information. When market observers contend that stocks were mispriced in the past, they typically exhibit hindsight bias by relying on information that only became available subsequent to the alleged mispricing. The many event studies conducted since the late 1960s address the informational efficiency of stock price changes rather than the fundamental efficiency of stock price levels. No surprise here since tests of fundamental efficiency necessarily presuppose the implausible—knowledge of the right price. In sum, prices are neither right nor wrong, there are only transacting investors who are betting they are wrong.

Allocative efficiency

The most basic function of capital markets is to allocate scarce resources to firms with the most promising long-term prospects. Perfect resource allocation, like its equivalent fundamental efficiency, is impossible because it too assumes flawless foresight. Allocative efficiency, or how well

9

market prices allocate resources, depends on the skills of informed buyers and sellers with competing estimates of discounted cash-flow values. Reasonable estimates of value require not only the use of an economically sound model but also well-conceived inputs not burdened by common behavioral pitfalls such as overconfidence and anchoring.

Treynor distinguishes “between ideas whose implications are obvious” and those “that require reflection, judgment, and special expertise for their evaluation and are the only meaningful basis for long-term investing.”12 The long-term economic implications of company announcements about events such as major acquisitions, FDA approval of a drug, the appointment of a new CEO, or a government anti-trust action are highly uncertain at announcement date. Increased trading volume and price volatility signal that there are wide differences of opinion about the significance of the news.

The always uncertain future affords the most prescient investors opportunities to earn excess returns by betting against current market prices. Their competitive advantage lies in their superior ability to correctly anticipate the longer-term valuation implications of currently available information before others. In other words, there’s no free lunch, but there are occasional early-bird specials for the most skillful investors.

This nearly informationally-efficient market dominated by investors who employ sound valuation models, will be difficult to attain in an environment largely populated by agents with imperfectly aligned incentives. This becomes evident by examining the pervasive use of non-DCF models.

Non-DCF models

The quarterly performance of mutual fund and pension fund managers is evaluated relative to a benchmark like the Standard & Poor’s 500 Index as well as relative to peers. Understandably, given such an evaluation structure, investment managers focus on short-term relative performance and are hypersensitive to tracking error. “Closet indexers” prefer the safety of performing acceptably close to the index to the more personally risky strategy of trying to maximize long-run returns. They argue that failure to achieve acceptable benchmark performance in the short run could lead to large fund withdrawals and their dismissal. Under no circumstances do investment managers want to be “wrong and alone.”

Managers constrain their chances of outperforming their benchmarks when their portfolios closely mimic the weightings of their benchmark

10

indexes. If prices are in fact driven by short-term earnings, then mean-variance portfolio optimization which depends on price data reinforces the influence of earnings.

Some investment managers select stocks based on near-term investor sentiment and play the earnings expectations game. The components that make up earnings can convey information, but those who fixate on the bottom line are chasing a largely value-irrelevant signal.

In the search for mispriced stocks, investment managers employ fundamental analysis which allegedly takes a long-term view of the company’s prospects. However, because forecasting cash flows is considered too speculative and costly, much of what is known today as fundamental analysis entails the use of shortcut metrics like price/earnings, price/sales, and price/book multiples that sidestep direct forecasts. Investment managers also frequently employ computer models that incorporate a variety of metrics and screening devices to select stocks.

The P/E multiple is by far the investment community’s most widely used investment yardstick. Instead of comparing the stock price with its discounted cash-flow value, the P/E ratio compares price with the past year’s reported earnings or next year’s estimated earnings. The dangers of drawing conclusions from a noisy, one-year earnings figure are well documented. Additionally, P/E multiples ignore risk and the resulting differences in discount rates.13

While conceding these shortcomings, analysts still attempt to identify investment opportunities by comparing P/E multiples of companies within the same industry by taking into account differences that warrant higher or lower multiples. Such relative valuation makes no effort to independently estimate the absolute value of stocks and thereby makes no direct contribution to allocatively-efficient prices.14 The pervasive use of relative valuation for selecting stocks among portfolio managers comes as no surprise since their performance is measured relative to a market index and competing funds.

Technical analysis, on the other hand, makes no pretense of being concerned with company fundamentals or prospective cash flows. Instead, it studies patterns of stock price movements and volume in search of profitable buy and sell signals.

Index funds are another important and growing segment of the market that makes no independent contribution to allocatively-efficient prices since indexing requires no valuation. Equity index funds now represent about 15 percent of all equity fund assets in the mutual fund industry.

11

Restrictions on short-selling are a barrier to allocatively-efficient prices because they limit the ability of pessimistic, would-be short-sellers to reflect their opinions in prices. The restrictions, however, would only affect allocative efficiency if would-be short-sellers employ discounted cash-flow analysis.

Finally, there are investors who do not base their decisions on expected returns. In the words, of Fama and French15 they treat equity investments as “consumption goods.” Examples include socially responsible funds, employees who hold large undiversified positions in their employer’s stock to demonstrate loyalty, and investors who enjoy holding growth stocks and dislike distressed (value) stocks.

The pervasiveness of short-termism and non-DCF models makes it difficult to conclude that prices are allocatively-efficient. If non-DCF managers of institutional funds dominate market activity, how likely is it that prices are set by buyers and sellers who employ DCF? Who then are the guardians of allocative efficiency? All the same, it’s prudent not to dismiss the possibility that the aggregation of many investors with diverse decision rules and information sets can somehow discover allocatively-efficient prices in an Adam Smith invisible-hand type fashion.16

The wisdom of collectives

A complete discussion of allocatively-efficient prices must consider the documented ability of groups to estimate current and predict future states. Mauboussin17 presents an example of each. In a 1907 Nature article, “Vox Populi” Francis Galton documents the results of a contest he conducted for 787 participants to guess the weight of an ox. The median guess was within 0.8 percent of the correct weight of 1,198 pounds and the mean of the guesses was within 0.01 percent. The distribution of the guesses shows that errors cancel out and lead to the extraordinarily accurate result. The result has been replicated many times since Galton including Jack Treynor’s oft-cited number-of-jellybeans-in-a-jar experiments.18

While Galton’s contestants estimate a current state, prediction markets like the Iowa Electronic Markets (IEM) forecast future states. IEM traders buy or sell futures “contracts” based on their predictions about an upcoming election. Since 1988, IEM has been more accurate than national polls 75 percent of the time, with an average prediction error of 1.37 percent versus 2.0-2.5 percent for polls.

Some believe an analogous process takes place in the stock market. I am far less confident of invisible-hand assertions in the presence of the

12

visible hands of earnings- and benchmark-obsessed investors. To understand why, consider some important differences between Galton and IEM and the stock market.

The results in all three settings depend on aggregation of information by individuals who are motivated to be “right.” While Galton and IEM contestants address a one-time outcome, stock prices are continuously changing and have no predictable endpoint. Furthermore, current prices are themselves a source of information and hence can influence subsequent stock prices. What participants are estimating is unambiguous in the case of Galton and IEM, but far less obvious in the case of stock market pricing.

1Notes

? “Financial accounting is not designed to measure directly the value of a business enterprise, but the information it provides may be helpful to those who wish to estimate its value.” Financial Accounting Standards Board, Statement of Financial Accounting Concept No. 1 (1978).2 Ordinarily net working capital (short-term receivables minus payables) accounts for a miniscule part of the share price. Furthermore, these are relatively low-uncertainty accruals. For most companies unfunded pension and postretirement benefit plans represent the greatest burden on future cash flows. David Zion of Credit Suisse First Boston estimated the unfunded balances to total 5% of the S&P 500 market capitalization at the end of 2003. 3 Sloan (1996) finds that the extent to which current earnings performance persists into the future depends on the relative mix of cash flow and accrual components of current earnings. Specifically, the accrual component of earnings is less persistent than the cash-flow component. However, Sloan finds that stock prices behave as if investors fixate on earnings, failing to exploit the information in the cash flow and accrual components of current earnings.4 See Hagin (2004).5 Ellis (2004), p. 265 estimates at 100 percent turnover the typical portfolio manager is making four multi-million dollar “to-buy or not-to-buy” and “to-sell or not-to-sell” decisions every business day of the year. Lower transaction costs and capital gains tax rates may have also contributed to shorter holding periods.6 I thank Shyam Sunder for discussions that helped clarify the link between the difficulty short-horizon investors face with backward induction and their turn toward projecting short-term metrics like earnings. See Hirota and Sunder (2004) for a more complete discussion. 7 See Shleifer and Vishny (1997), Shleifer (2000), and Barberis and Thaler (2003).8 Bogle (2004) estimates mutual fund intermediations costs (advisory fees, marketing expenditures, sales loads, brokerage commissions, transaction costs, custody and legal fees, and securities processing expenses) total at least 2 ½ percent of assets and consume nearly 40 percent of the 6 ½ percent historical real rate of return on equities.9 See Grossman and Stiglitz (1980).10 Black (1986).11 Benartzi and Thaler (1995).12 Treynor (1976).13 For a more detailed discussion of the shortcomings and complexities of P/E, see Leibowitz (2004) and Rappaport (2003).14 Professor Aswath Damodaran of New York University examined 550 equity research reports between 1999 and 2001 and found that 85 percent are based on multiples and comparables.15 Fama and French (2004).16 Mauboussin (2002) and Surowiecki 2004) offer more optimistic views on the pricing ability of decentralized self-organizing systems.17 Mauboussin (2003).18 Treynor (1987).

13

The stock market aggregates information, but it’s just not obvious that it’s the information needed to discover allocatively-efficient prices.

While partially-informed groups can make surprisingly wise collective choices, they are also capable of making costly errors. When participants act independently their “errors” cancel out, but when herding behavior prevails the wisdom of collectives collapses and markets tend toward excesses. During fast-rising markets it’s called “momentum investing” sometimes leading to a bubble and in sharply declining markets it’s called “panic”.

Can Investment Managers Earn Excess Returns If They Buy and Sell Stocks They Believe the Market Has Mispriced on a DCF Basis?

The efficient markets literature assumes that when stock prices diverge from informed estimates of discounted cash-flow values, arbitrageurs buy or sell to bring prices back into line. More recently, behavioral economists have argued that arbitrage is risky and costly, thereby severely limiting the opportunity to exploit mispricings.

Barberis and Thaler specify three factors that limit arbitrage—fundamental risk, noise trader risk, and implementation costs.19 They illustrate fundamental risk with the following example. Suppose an investor buys Ford shares after excessively pessimistic traders drive its price well below her best estimate of fundamental value. She faces the risk that bad news will drive Ford stock down even further. She can hedge this risk by shorting General Motors when she purchases Ford shares. While this protects the arbitrageur from bad news for the industry, it still exposes her to Ford-specific surprises such as what occurred with the Firestone tire debacle. Finally, there is also the possibility that the imperfect substitute security, General Motors in this case, may be similarly mispriced.

Noise trader risk materializes when traders become even more pessimistic, driving Ford shares to fall even further. Noise traders buy and sell without reference to news that might affect the fundamental value of a stock. Arbitrageurs may therefore be forced to liquidate their positions prematurely. Short sellers can actually help fuel bubbles. They offer a safety net to holders of a bubble stock because short sellers are forced to buy shares back if the price rises sufficiently, thus ending up playing the “greater fool.” In brief, noise may cause prices to diverge from value, but it is difficult to arbitrage the gap.

19 Barberis and Thaler (2003).

14

Barberis and Thaler include identifying mispricings, trading commissions, bid-ask spreads, market impact costs, short-sale fees and constraints as costs that make arbitrage less attractive. Not surprisingly, professional arbitrage such as that conducted by hedge funds is concentrated in the bond and foreign exchange markets where investors can estimate value with greater confidence than in the stock market.

If arbitrage is not feasible, then investors seeking to exploit mispricings must trade on their ability to translate available information into better estimates of value than the current price. This process is also risky and costly.

If short-term earnings information dominates stock price changes, why should long-term investors base their decisions on a company’s cash-flow prospects? The simple answer is that stock prices ultimately depend on a company’s ability to generate cash flow. As Peter Bernstein points out, while art collectors can only hope that other collectors will step up in the future to justify today’s purchase price, owners of financial assets such as stocks and bonds must depend on the company to produce cash flows rather than on the whims of other investors. The market reveals its broad agreement by bestowing relatively large market capitalizations to companies with demonstrated cash-generating capability and lower capitalizations to those with less impressive track records and prospects.

Here’s the crucial observation. If stock prices ultimately depend on company cash flows it is unnecessary to “prove” that prices actually reflect the cash-flow expectations of buyers and sellers at any particular time. Instead, estimate the cash-flow expectations implied by the current price and assess whether your expectations are sufficiently different to warrant purchasing or selling shares.

Two basic factors shape the returns from a stock you believe to be mispriced. First, the greater the estimated mispricing relative to the current stock price, the greater the potential return. Of course, a stock may turn out to be mispriced, but not by as much as you believe. The second factor is the time it takes the stock price to converge toward the target price. The shorter the time, the greater the return. The longer it takes, the lower the return.20 For price to move towards the investor’s target price either other investors come to agree with the investor’s assessment of a company’s prospects or the prospects are eventually reflected in the information investors tap to 20 Rappaport and Mauboussin (2001) illustrate with a stock priced at $80 per share, a 20 percent discount from its estimated fundamental value of $100. Assuming a ten percent cost of equity and no change in expectations, fundamental value in one year would rise to $110. If the $80 stock rises to $110 in a year, the annual return would be 37.5 percent and the excess annual return 27.5 percent. If it takes two years to reach the target price, the excess return drops to 13.0 percent and to 8.5 percent for three years.

15

make their trading decisions. For example, if using a DCF analysis you conclude that a stock is undervalued, in an earnings-driven market you must rely on future reported earnings to correct the mispricing. As long as short-term earnings analysis and noise dominate price changes, prices may not converge quickly toward the target estimate of value. As John Maynard Keyes cautioned over seventy-five years ago, “markets can remain irrational longer than you can remain solvent.” Finally, investors employing DCF analysis, just as arbitrageurs, face the risk that new, unanticipated information triggers unfavorable price changes. This risk is countered by unanticipated information that produces favorable price changes.

Only individuals with brains, resources, a long investment horizon, and no agency conflicts are promising candidates for exploiting mispricings. To improve their chances of success, the market’s fascination with the short-term and its obsession with earnings will have to be reduced. The final section of the paper charts a course for accomplishing this.

Is Corporate Management’s Focus on Short-term Earnings Self-serving or Also in the Best Interests of Its Shareholders?

Corporate executives point to the behavior of market participants to justify their short-term focus and their belief that investing for the long-term is not rewarded by higher stock prices.21 This bias is reinforced by incentive compensation plans that reward short-term financial performance. Even equity incentives such as stock options and restricted stock do not alter the short-term orientation of executives if they believe that near-term performance largely influence stock prices. To the contrary, incentives for options-laden executives to misrepresent publicly-reported financial information increased during the 1990s.

There is compelling evidence that managers are obsessed with earnings. A survey by Graham, Harvey, and Rajgopal (GHR)22 shows that the vast majority of 400 financial executives viewed earnings as the most important performance measure that they report to outsiders. The two key earnings benchmarks are quarterly earnings for the same quarter last year and the analyst consensus estimate for the current quarter. Executives believe that meeting earnings expectations helps maintain or increase the

21 “Reported earnings follow the rules and principles of accounting. The results do not always create measures consistent with underlying economics. However, corporate management’s performance is generally measured by accounting income, not underlying economics. Therefore, risk management strategies are directed at accounting, rather than economic, performance.” Enron in-house risk-management manual, p. 132.22 Graham, Harvey, and Rajgopal (2004).

16

stock price, provides assurance to customers and suppliers, and boosts the reputation of the management team. Failure to meet earnings targets is seen as a sign of managerial weakness and, if repeated, can lead to a career-threatening dismissal.

Missed earnings targets may also signal the presence of more serious problems because investment managers and analysts justifiably assume that all companies manage earnings and have considerable discretion in accounting that enables them to meet quarterly earnings expectations in most situations.23

The willingness of executives to forego or delay value-creating activities to meet quarterly earnings targets is evidence of the crucial importance they attach to meeting expectations. GHR report a startling 80 percent of survey respondents would decrease discretionary spending on R&D, advertising, maintenance, and hiring to meet earnings benchmarks and more than half would delay a new project even if it entailed giving up value. As GHR aptly observe, “getting managers to admit such value-decreasing actions in a survey perhaps suggests that our evidence represents only the lower bound of such behavior.”

Managers with career concerns and short-term incentive compensation arrangements are predictably obsessed with earnings. The question is whether their focus on earnings contributes to or compromises shareholder value in a market where stock prices respond to earnings information.24

The idea that management’s primary responsibility is to maximize long-term shareholder value is widely accepted in principle, but imperfectly implemented in practice. Maximizing long-term value means that management’s primary commitment is to continuing shareholders rather than to day traders, momentum investors, and other short-term oriented market players. To succeed managers must develop and effectively execute strategies that maximize the company’s long-term cash-flow potential.

Managing for short-term earnings compromises shareholder value in two ways. First, companies delay or forego value-creating investments to meet consensus earnings expectations. While such actions improve the current period’s reported earnings, they reduce the company’s future earnings potential and value. Second, companies exploit the discretion 23 Degeorge, Patel, and Zeckhauser (1999) present impressive empirical evidence of earnings management.24 The tension between earnings and shareholder value is lowest among companies with relatively small amounts of capital expenditure and R&D outlays. In such companies the impact of current operating decisions shows up in earnings promptly rather than with a lag. The reverse is true in capital- and knowledge-intensive industries such as manufacturing, pharmaceuticals, and software where revenues lag investments by several years or more.

17

accounting rules allow in the calculation of earnings by pushing revenues into the current period and deferring expenses to future periods. Borrowing from the future to satisfy today’s reported earnings expectations inevitably catches up with the borrowing company.

When a company can no longer deliver on expectations, the market hammers the stock price. Jensen, who in 1978 famously declared that “there is no other proposition in economics which has more solid empirical evidence supporting it than the Efficient Market Hypothesis,” points to the hundreds of billions of dollars of market value erosion due to overvalued equity.25 He cites WorldCom, Enron, Nortel, and eToys as companies which pushed earnings management beyond acceptable limits to meet expectations and ended up destroying part or all of their value.

Maximizing long-term cash flows rather than managing for short-term earnings, even in an earnings-dominated market, is the most effective means of creating value for continuing shareholders. Again, the governing objective of managing in the interests of continuing shareholders justifies this conclusion. The conclusion even holds for companies that engage in significant transactions in their own stock.

Companies ordinarily create a disproportionately large part of their value from operations. But they also issue new shares and repurchase outstanding shares. In some cases these financial transactions can create or destroy significant shareholder value.

One possible argument in favor of managing for short-term earnings is that an earnings-addicted market will price shares more favorably and benefit continuing shareholders when new shares are sold to incoming shareholders at a higher price. There are three flaws in this argument.

First, reporting better earnings, whether accomplished by real decisions that compromise value or accounting gimmicks, will sooner or later catch up with the company when it can no longer meet market expectations and the value of continuing shareholders’ shares fall significantly. Second, when an acquiring company offers shares to the selling company’s shareholders, the attractiveness of the offer is not evaluated by short-term earnings consequences but by comparing the expected value of selling shareholders’ interest in the combined enterprise with the current share price. Third, if a company needs to raise funds but management believes its shares are undervalued, alternatives to equity financing like debt financing or limiting dividend payouts can usually be tapped.

25 Jensen (2004).

18

Not only is earnings management of questionable value for issuing new shares, it can destroy significant value when companies employ it as the criterion for share buybacks. A company should repurchase shares only when its stock is trading below management’s best estimate of value and when no better investment opportunities are available. Companies that follow this guideline serve the interests of continuing shareholders who, if management’s assessment that shares are undervalued is correct, gain at the expense of shareholders who voluntarily tender their shares.

Spurred by the belief that investors mechanically apply a multiple to current earnings to establish value and the fact that management’s compensation is partially tied to earnings performance, some companies repurchase shares even when they believe shares are fairly valued or overvalued. When an immediate boost to earnings-per-share rather than value creation dictates share buyback decisions, wealth is transferred from continuing shareholders to exiting shareholders. Especially widespread are buyback programs that offset the earnings-per-share dilution from employee stock option programs. In these cases, the exercise of options by employees rather than undervaluation dictates the number of shares and the prices at which they are repurchased.

A Three-Pronged Attack on Short-Term Performance Obsession

To reduce short-term performance obsession and improve allocative efficiency, changes in three areas are needed--corporate performance reporting, incentives for investment managers, and incentives for corporate managers.

Corporate Performance Reporting

Relevant, transparent, and timely information is vital to the allocative efficiency of markets. In the present unforgiving climate for accounting shenanigans, companies have an unprecedented opportunity to meaningfully improve the form and content of their financial statements. Not only is better disclosure an antidote to earnings obsession, but it is also an act of enlightened corporate self-interest that can reduce investor uncertainty, decrease the company’s cost of capital, and restore confidence in corporate reporting.

A Corporate Performance Statement shown below would:26 Separate cash flows and accruals.

26 The discussion of the Corporate Performance Statement is adapted from Rappaport (2004).

19

Classify accruals by levels of uncertainty. Provide a range and the most likely estimate for each accrual. Exclude arbitrary, value-irrelevant accruals. Detail assumptions and risks for each line item.

Separating realized cash flows from forward-looking accruals provides a historical baseline for estimating a company’s future cash flow prospects and enables analysts to evaluate the reasonableness of accrual estimates. While accruals ordinarily account for only about 5 percent of a company’s price, separating cash flows and accruals provides investors a platform for assessing the remaining portion of the price. Transparent accruals also discourage companies from producing unrealistic estimates or

20

engaging in outright fraud. Most important, separating cash flows and accruals helps restore confidence in the integrity of corporate reporting.

The Corporate Performance Statement calculates free cash flow as revenue minus operating expenses (excluding noncash charges such as depreciation, amortization, deferred taxes, and asset and liability revaluations) minus all investments, including working capital changes. Investments include those that appear on the balance sheet, like production facilities, equipment, real estate, patents, and trademarks, as well as expenditures companies ordinarily expense for such things as research and development, software development, and branding activities. Whether a company records an expenditure as an operating expense or a capitalized asset does not affect free cash flow since it is subtracted in either case.

The Corporate Performance Statement focuses on a company’s operations. It is designed to replace the traditional income statement, but cannot entirely replace the traditional cash-flow statement because cash flows from financing activities--new issues of stocks, stock buybacks, new borrowing, repayment of previous borrowing, and interest payments--are excluded. A new cash-flow statement can begin with the “free cash flow” line of the Corporate Performance Statement and add and subtract the various financing activities to calculate the increase or decrease in cash.

The Corporate Performance Statement separates accruals into three increasing levels of uncertainty--low, medium, and high. Low-uncertainty or “check is in the mail” accruals included in revenue and operating expenses are relatively low-risk because a company normally expects to convert the corresponding receivables and payables into cash over the next accounting period. Medium and high-uncertainty accruals, on the other hand, have longer cash-conversion cycles and wider ranges of plausible outcomes.

Companies typically develop estimates for medium-uncertainty accruals, like allowances for uncollectible receivables and warranty obligations from historical experience and modify the assumptions for changes in current conditions. For example, restructuring charges reflect estimates of future-period outlays for such things as severance pay, cancelled leases, and litigation. Deferred-tax accruals result from temporary timing differences between pretax book income and taxable income for items such as depreciation expense. Estimated unrealized gains or losses from incomplete long-term construction, energy, and R&D contracts usually depend on assumptions about future prices, costs, and a host of other factors.

The cost of defined-benefit pension and employee stock-option plans are examples of high-uncertainty accruals. Pension expense, for example,

21

should reflect the change in the present value of the company’s obligations minus the change in the present value of expected returns on pension-fund assets. The calculation requires a panoply of assumptions, including projected employee turnover, future pay increases, estimated retirement dates, future market discount rates and expected return on plan assets.

The traditional income statement, with its single-point accrual estimates, ignores the wide variability of possible outcomes--particularly for medium and high-uncertainty accruals. The Corporate Performance Statement complements the most-likely figure for each accrual with optimistic and pessimistic estimates. These estimates, coupled with management’s disclosure of the associated probabilities for each, enable investors to form their own expectations more confidently.

Low, medium, and high-uncertainty expense accruals are estimates of future cash flows a company needs to satisfy commitments to customers, employees, and suppliers arising from earlier arms-length market transactions. In sharp contrast, companies record depreciation and amortization charges after the outlay of cash for investments. Faced with the unknowable magnitude and timing of future cash flows that capitalized assets will generate, accountants use arbitrary depreciation methods to assign expenses over their expected useful lives. This is clearly a case of accounting ritual trumping relevance. Therefore these value-irrelevant charges are not included on the Corporate Performance Statement.

Companies are required to record an impairment charge if the value of acquired goodwill, other intangible assets or long-lived fixed assets have shrunk below their balance-sheet amounts, typically estimating asset values based on speculative assumptions about future sales and costs. Even if a company can reasonably establish values for individual assets, intangibles such as customer goodwill, employee training, software development and R&D ordinarily can’t be sold separately from the company itself. Investors are interested in the going-concern value of a company’s various businesses and its consolidated value, not the assigned value of individual assets. Asset impairment charges are therefore also excluded from the Corporate Performance Statement.

Nonrecurring gains and losses, charges from discontinued operations, and the effect of accounting changes are disclosed in the management discussion and analysis section but excluded from the Corporate Performance Statement because they offer no meaningful help in forecasting the sustainability and growth potential of a company’s cash flows.

The Statement presents no bottom line because no single number can reasonably encapsulate a company’s performance. The traditional earnings

22

bottom line misleadingly suggests that aggregating amounts based on past transactions and on uncertain assumptions about future transactions somehow yields an economically meaningful number.

Finally, the Corporate Performance Statement includes a “management discussion and analysis” section in which management presents the critical assumptions supporting each accrual estimate and the company’s business model along with key financial and non-financial performance indicators that drive value, such as customer-retention rates, time to market for new products and quality improvements.

Will the Corporate Performance Statement prove too costly? If the information isn’t already available for internal purposes there should be concern about senior management’s grasp of the business and the board’s exercise of its oversight responsibility. Board members, particularly members of the audit and compensation committees, should, at minimum, know how much of the company’s reported performance comes from realized cash flows and how much from accrual estimates, as well as the risk that “most likely” revenue and expense accruals will prove to be materially misstated.

Companies commonly use short-term financial measures such as operating earnings and return on invested capital to measure operating-unit performance and to determine incentive pay for senior executives. The Corporate Performance Statement exposes the shortcomings of earnings and makes it easier for boards to champion executive-compensation plans that reward management for creating long-term value rather than achieving short-term earnings results. Specific recommendations are presented in the section dealing with incentives for corporate managers.

The Corporate Performance Statement also offers benefits to the accounting profession. Auditors are vulnerable to bias. While they owe their first loyalty to shareholders, they are also eager to please their corporate clients. To report better earnings, companies look to accelerate revenue and delay expenses. This is the subject of most negotiations between auditors and their client companies and the source of virtually all the recent accounting scandals.

The Corporate Performance Statement makes these negotiations unnecessary. Auditors verify reported free cash flow by reviewing the company’s internal controls and conducting substantive testing. For accruals, the Statement shifts the auditor’s task from judging the reasonableness of single-point estimates to evaluating management’s three-level (optimistic, pessimistic and most likely) estimates. This places greater reliance on the auditor’s knowledge of business and industry dynamics. A

23

skill set emphasizing competitive strategy analysis rather than one largely based on familiarity with detailed accounting rules will attract better talent to a profession that has lost considerable ground in the competitive labor market over the past two decades.

By eliminating the earnings bottom line, the corporate performance approach to financial reporting removes many high-decibel but irrelevant controversies. It would shield accounting standard-setters from intense lobbying and government intervention. For example, the argument that companies should not deduct the cost of employee stock options from earnings because options are too difficult to value vanishes. There are no reported earnings and the wide variability of possible outcomes is explicitly recognized in the Corporate Performance Statement. The Financial Accounting Standards Board and the International Accounting Standards Board can then focus on developing the most informative means of estimating accruals and reporting their potential variability, criteria for the line items to include in the Corporate Performance Statement, and disclosure guidelines for the management discussion and analysis section.

Incentives for investment managers

Even if improved disclosure along the lines of the Corporate Performance Report became accepted practice, earnings obsession will persist as long as investment managers have inadequate incentives to use the new information and shift their analytic orientation toward valuing a company’s long-term prospects. They, of course, need to be convinced that the shift would improve their performance and compensation.

Investment managers must overcome two fears before accepting the possibility that discounted cash-flow analysis could boost their performance. The first fear is that others in the investment community will not follow and short-term earnings continue to significantly influence stock prices. Second, there is the ever-present concern that forecasting highly uncertain cash flows is simply impractical.

If investors shift their focus from short-term earnings to long-term cash flows, achieving excess returns would depend on their ability to interpret the valuation implications of information better than other market participants. This is a daunting task and only the most skillful investors are likely to succeed. Active management in an allocatively-efficient market is not for the fainthearted, but the criterion for success--making superior judgments about the future performance of companies--is unambiguous.

24

Even if short-term earnings continue to influence stock price changes, the ability to make superior judgments about a company’s future prospects remains the key to successful long-term investing. In an earnings-dominated market, it just takes time for the “earnings evidence” to materialize and be reflected in stock prices. The longer it takes, the more difficult it is to earn excess returns.

The second concern—the reluctance to forecast cash flows—can be easily dispelled. You can, in fact, employ the DCF model without assuming the burden of making independent cash-flow forecasts. Think about it this way: It’s difficult for any individual to forecast an uncertain future better than the collective wisdom of the market can. So instead of forecasting cash flows, begin with the current price and determine the expectations for a company’s future cash flows that justify the price. Estimating price-implied cash-flow expectations is critical because only investors that correctly anticipate changes in a company’s prospects that aren’t already reflected in the current price can hope to earn excess returns. While this “expectations investing” approach blunts a major objection to DCF valuation, it too is no easy ticket to generating alphas.27

Investment managers are likely to cling to short-term accounting metrics, high turnover, and benchmark tracking in the absence of significant changes in existing performance evaluation and compensation arrangements. Most commonly, fund managers are paid a management fee based on a percentage of the market value of the assets they manage. Less frequently, they are paid fees based on performance relative to their benchmark and/or a group of peers.

Because net inflows are positively correlated with a fund’s recent performance, asset-based fees encourage managers to focus on short-term returns that increase the assets they manage. Such behavior by investment managers will continue until investors realize that chasing “hot” funds is a losing game. But what better criterion for choosing an active fund can they substitute? Investing in a fund is essentially an act of faith. It takes more years than either the investor or the manager will be around to assess with statistical confidence how much of the fund return is due to skill and how much to chance.

Critics frequently recommend reducing the management fee and including a meaningful performance incentive in the compensation mix as a way to better align fund manager and shareholder interests. The impact of incentives on behavior is complex. Poorly-designed incentives can worsen rather than defuse manager-shareholder conflicts. For example, a 27 A detailed presentation of the process can be found in Rappaport and Mauboussin (2001).

25

management fee that just covers the cost of servicing the account might encourage a manager to focus disproportionately on the performance fee, and hence, take on unacceptable risk. On the other hand, because relative performance dominates retention-dismissal decisions, managers hesitate to take on risk levels that trigger unacceptably high tracking error.

Simply changing the mix of management and performance fees does not come to grips with the major agency costs of delegated investment. More fundamental changes are needed in the structure of funds, design of performance fees, and the benchmarks chosen for performance measurement.

Stein reports that open-end funds, which enable investors to liquidate shares at net asset value at any time, account for 96 percent of total mutual-fund assets and 93 percent of all funds.28 The open-end structure exposes even managers with outstanding track records to large withdrawals if they perform poorly in the short-run or if equity prices experience a sustained downturn.29 This risk discourages managers from making trades that are only attractive in the long-run. At first glance open-end funds appear to be a losing proposition for managers and shareholders alike. Managers operate under the tight leash of short-term relative performance and the lurking risk of dismissal. Shareholders forego the potential of larger alphas when skilled managers focus on tracking error.

Why then is the open-end organizational form so dominant? Perhaps investors worry that closed-end managers with long-term contracts may turn out to be incompetent or even dishonest. Unlike an open-end investment that can be liquidated at its underlying net asset value at the first hint of trouble, closed-end investors have no recourse but to liquidate at a painful market discount to asset value which materializes when investors lose confidence in fund managers.

Stein argues that while the prevalence of open-end funds may appear to represent a socially efficient outcome, it is not. Instead, he hypothesizes that “the gains from being able to undertake longer-horizon trades in the closed-end form outweigh the potential losses that come from being unable to control wayward managers.”30 Unfortunately, as Stein explains, it will be very difficult to sustain the efficient outcome where all funds are closed-end

28 Stein (2004).29 Bernstein (2004) cites the case of Bill Miller the legendary manager of the Legg Mason Value Trust. According to Morningstar data, Miller beat his S&P 500 benchmark, and was top quartile every year but one, from 1992 to 1999. During the falling markets between 1999 and 2002 Miller lost 42 percent of his assets despite beating his benchmark and being in the top quartile during this period.30 I am not aware of any studies that compare the long-term performance of open-end versus closed-end funds.

26

because the best-performing managers will likely move to the open-end form to gain more assets under management and higher compensation. Other high-quality as well as lower-quality managers will then also migrate to the open-end model to avoid losing their investors. This brings us right back to the current state of affairs in which benchmark-sensitive managers afraid of being wrong and alone focus on short-term relative performance and the earnings number reigns supreme.

The challenge for closed-end funds is to develop incentives that encourage the best managers to remain and poor performers to leave. Assuming that funds hire portfolio managers with established track records, the performance evaluation horizon can be reasonably extended to perhaps three to five years or more. This gives managers the freedom to be wrong and alone in the short run and time to demonstrate their skills over a longer time period. For managers who seriously underperform, the fund should be able to exercise a buyout provision.

What can closed-end funds do to keep high-ability managers once they have demonstrated a sustained period of exceptional performance? First, make total compensation including variable performance competitive with the amount managers could earn if they moved to an open-end fund. Second, pay annual bonuses on the basis of rolling three-to-five year performance. Third, extend the performance measurement period and motivate long-term value creation, by deferring some payouts and placing them “at risk” against future performance. Finally, require portfolio managers to make a meaningful investment in the fund.

The near-universal use of benchmarks like the S&P 500, the Russell 2000, or the Wilshire 5000 reduces differences in returns between the best and worst performers because unskilled managers mask their shortcomings by closely tracking benchmarks and even skilled managers succumb to tracking, thereby concealing their skill. Benchmark tracking, along with a herd-like focus on short-term earnings, culminates in a no-free-lunch, informationally-efficient market, but also could produce mispricing opportunities for those willing to take the road less traveled.31

The performance measurement problem for individual stock funds is not benchmarking per se, but rather short-horizon benchmarking, restrictive tracking-error constraints, and benchmarks that do not provide managers sufficient investment scope to demonstrate their skills. Short-horizon benchmarking encourages managers to focus on their tracking-error risk instead of owners’ risk-adjusted long-term returns. Closet indexing is exacerbated when consultants or sponsors impose tight tracking-error constraints on managers. According to Grinold’s “fundamental law of active

27

management” information ratios are a function of not only skill but also the breadth of opportunities available to managers. To achieve breadth, managers must increase the number of independent alpha bets they make. Decisions are “independent” when supporting forecasts depend on different and uncorrelated sources of information. When skilled managers with a track record are identified they should be given greater breadth rather than being confined by a style box.

The fund is essentially betting that the potential for increased information ratios from extending the performance evaluation horizon, relaxing tracking error restrictions, and allowing greater breadth outweighs the risk that some managers will prove to have no value-adding skills. These initiatives to increase information ratios are far more risky for open-end than for closed-end funds because short-run underperformance can trigger significant fund withdrawals from open-end funds. The widespread adoption of longer investment horizons would however reduce the value-relevance of short-term earnings and ultimately earnings obsession.

The closed-end format for mutual funds, the four incentive features for managers outlined earlier, extended time horizons and greater breadth represent promising steps for reducing the agency costs of delegated investment, reducing earnings obsession, and increasing allocative efficiency.

Nonetheless, incentive contracting to alleviate the agency costs of delegated investment is complex and outcomes are difficult to anticipate. A competitive market solution is certainly preferable. Consider the following possibility.

Individual investors seeking professional management and diversification for the equity portion of their portfolios can choose a low-cost market index fund and/or a higher-cost fund with a modest level of active management due to benchmark tracking. A better asset allocation

31 Bernstein (2000) argues that market-index benchmarks are floating crap games whose membership constantly changes through mergers and decisions to drop and add index companies. More importantly, because most indexes are market-weighted, the hottest stocks acquire the greatest weight. When a relatively small number of stocks account for a significant percentage of an index as occurred during the 1990s technology bubble, the tracking portfolio may well become much riskier than clients would prefer. Managers face a Hobson’s choice of either living with a level of risk incompatible with the fund’s objectives or generating an unacceptably high tracking error. Bernstein recommends that required return and tolerable risk become the performance benchmarks for calculating alphas and information ratios. Required return and risk are determined by the fund’s obligations to participants whether they are part of a corporate-sponsored pension plan or mutual fund shareholders hoping to build wealth to finance education costs or their retirement years. The problem is that the rate required to defease pension fund liabilities or fulfill the expectations of mutual fund investors will invariably be greater or less than a market-based estimate of expected rate of return. A bogey based on a fund’s requirements rather than market opportunities cannot serve as an effective measure of performance.

28

choice would be between an index fund and a long-horizon fund that is truly actively managed. This would enable investors to allocate their equity dollars between the two funds based on their tolerance for market-deviating returns while at the same time avoiding high fees for minimal active management. For example, an investor whose equity portfolio is entirely in a traditionally-benchmarked fund with a targeted tracking error can replicate the tracking error by an allocation between an index fund and an actively-managed long-horizon fund saving approximately 2 percent annually in fees for the portion allocated to the index fund.

Incentives for Corporate Executives32

Many commentators point to the deliberately deceptive accounting practices of Enron, WorldCom, Adelphi and other recent business failures and contend that the underlying cause is management’s infatuation with shareholder value. This claim fails to capture the essence of the shareholder value approach. The actions taken by these companies added no value, but were dishonest attempts to create the appearance that value was added. Shareholder value did not fail management, management failed shareholder value.

Most CEOs champion the goal of maximizing shareholder value without embracing the essential determinant of value—risk-adjusted, long-term cash-flows. Instead, they are obsessed with Wall Street’s earnings-expectations machine and short-term share price. Sacrificing the company’s long-term prospects to meet quarterly earnings expectations in an attempt to temporarily boost the stock price represents the very antithesis of sound shareholder value management. The driving force behind such behavior can usually be traced to executive compensation schemes.

In the early 1990s, as corporate boards endorsed shareholder value they became convinced that the surest way to align the interests of managers and shareholders was to make stock options a large component of executive compensation. By the end of the decade, stock options accounted for more than half of total CEO compensation in the largest U.S. companies. Options and stock grants also constituted almost half the remuneration of directors. But short-term thinking and earnings obsession did not decrease, instead they increased.

To discover what went wrong one only has to examine the principal features of the standard option plan: the exercise price equals market price at 32 This section deals only with incentives for CEO’s and other top executives that can affect the company’s stock price. For a detailed discussion of incentives for operating units see Rappaport (1999).

29

date of grant and stays fixed over the entire ten-year term, and the typical vesting period is three or four years. Four factors limit the ability of standard options to promote long-term value-maximizing behavior by corporate executives.33

Performance targets are too low. Holding periods are too short. Underwater options undermine motivation and retention. Options can induce too little or too much risk-taking.

CEOs widely declare their overriding commitment to achieving superior returns for shareholders. Standard stock options, however, reward performance well below superior-return levels. In a rising market, options can reward even mediocre performance. That’s because executives profit from any increase in share price—even one substantially below competitors or the broader market. The standard option is structured as if the opportunity cost of equity were zero. Since rising markets are fueled not only by corporate performance but also by factors beyond management control such as changing interest rates, some executives enjoy huge windfalls simply by being in the right place at the right time. No board of directors should approve an incentive plan that provides significant option gains for a level of performance that could become grounds for dismissing the CEO. Nor should institutional investors judged by the alphas they deliver for fund owners remain passive as corporate boards reward executives who not only fail to produce positive “corporate alphas” but achieve returns well below their peers, i.e. negative alphas.

Relatively short vesting periods coupled with the belief that earnings fuel stock prices encourage executives to manage earnings, exercise their options early, and cash out shares opportunistically. These actions significantly diminish the intended long-term incentives options and stock holdings provide. The practice of accelerating vesting for CEOs upon retirement adds yet another incentive to be short-term.

The standard stock option loses its power to motivate and retain executives when options are hopelessly underwater. Options fall underwater more frequently than is commonly believed. Hall reports that about one-third of all options held by U.S. executives in publicly-traded companies were underwater in 1999 at the height of the 1990s bull market.34 Board responses such as increasing cash compensation, granting restricted stock, 33 Rappaport (1999). Another factor which affects shareholders but not executives’ motivation is the gap between the cost of option grants to shareholders and their value to executives. .The value of options to undiversified, risk averse executives is substantially lower than the cost to shareholders. For excellent overviews of the shortcomings of traditional options see Hall and. Murphy (2003) and Hall, (2003).34 Hall (2003).

30

offering more options or lowering the exercise prices of existing options are shareholder-unfriendly tactics that rewrite the rules in midstream. They can undermine the option incentive by turning it into a heads-I-win, tails-I-win arrangement.

Without equity-based incentives executives tend to be excessively risk-averse in order to avoid failure and dismissal. The standard option however does not necessarily induce greater risk-taking. To preserve unrealized option gains, executives may bypass positive, but risky, value-creating investments. On the other hand, when their options are hopelessly underwater executives with little to lose may pursue overly risky investments in a desperate attempt to resuscitate the stock price and the value of their options.

Companies can go a long way toward overcoming the shortcomings of standard options by implementing a discounted indexed-options plan with extended time horizons. Indexed options have an exercise price tied to either an index of their competitors or a broader market index such as the Standard & Poor’s 500. While broader market indexes are easily tracked, they do not reflect the particular factors that impact the company’s industry and consequently are not appropriate benchmarks for measuring and rewarding management performance. In the discussion to follow indexed option plans are based on a peer group index. Some companies cannot find suitable competitors to include in a peer index. This problem is particularly troublesome for companies that have diversified into a wide range of products and markets. For those companies, discounted equity-risk options, is presented as an alternative following the discussion of indexed options.

As the index increases or decreases, the exercise price of the options increases or decreases by the same percentage. Indexed options are therefore only worth exercising if the company’s shares outperform the index. Indexed options do not reward underperforming executives simply because the market is rising. Nor do they penalize superior performers because the market is steady or declining. If the index declines, then so does the exercise price, which keeps executives motivated even in a sustained bear market.35 Indexed options reward superior performers in all markets. They overcome two of the criticisms directed at standard options--performance targets are too low and underwater options driven by a declining market undermine executive motivation and retention.

35 Some believe that a disadvantage of indexed option is that executives profit when they outperform the index even if the stock price falls below the exercise price at grant date. To counter this objection, boards can require that options only be exercised if the company’s stock is trading above its price at grant date or if it has appreciated at some minimum rate of return.

31

Companies can address the other two criticisms of standard options—holding periods are too short and the options induce too little or too much risk-taking—by extending vesting periods and requiring executives to hold meaningful equity stakes they obtain through option exercise or purchase of shares. Citigroup, for example, requires 100 of its top executives to retain at least 75 percent of their stock, including shares acquired by the exercise of options, as long as they remain with the company. Boards can also limit the sale of stock by CEOs over a two- or three-year period after retirement to ensure a longer-term focus.

Despite its advantages only Level 3 Communications, a telecommunications company, has adopted an indexed option plan. The rejection of indexed options because the cost must be expensed while the cost of standard options need only be disclosed in a footnote underscores the rampant obsession with earnings. Stock options do not become more or less costly depending on whether the disclosure is made in a company’s income statement or in its footnotes. Still, the requirement to expense indexed options has discouraged companies from adopting such plans. The impending Financial Accounting Standards Board requirement that companies expense standard options will level the playing field and perhaps end situations where earnings consequences, rather than economic substance, dictate the choice of executive compensation plans.

While accounting concerns are misplaced concerns, CEOs understandably shun indexed options because of their more demanding performance standard. To compensate executives for bearing greater risk, boards will have to offer more options so that high-performing CEOs will do better with indexed options than they would have with standard options.36

Median stock price returns are less than average returns because a relatively few stocks with extraordinarily high returns inflate the average. Consequently, more than 50 percent of indexed options will underperform and fall underwater. One response to the underwater-option problem is that executives that underperform don’t deserve incentive compensation. However, a management team without continuing incentives to create value is not in the best interests of shareholders.

One way to resolve the dilemma is to lower the exercise price allowing executives to profit at a performance level modestly below the index.37 Suppose the index rises 10 percent from 100 to 110 during the year. 36 For a detailed analysis of how to determine the additional shares needed see Rappaport (1999). Concerns over dilution should not focus on the number of options granted but rather on the number that can be exercised in the absence of superior performance. Because executives can be rewarded for weak performance under standard plans, there is a greater risk of dilution with standard plans than with indexed plans.

32

A 1 percent discount would reduce the yearend index from 110 to 108.9. The exercise price would therefore rise by only 8.9 percent instead of 10 percent. Discounted index options make gains accessible to more executives, motivate the best-performing executives to remain with the company, and encourage sub-par performers to leave.

For companies unable to develop a peer index, I propose “discounted equity-risk options” (DEROs).38 DEROs call for a significantly higher level of threshold performance than standard fixed-price options. But unlike indexed options, DEROs does not require the construction of an index. More specifically, the exercise price rises by the yield to maturity on the ten-year Treasury note plus a fraction of the expected equity-risk premium minus dividends paid to holders of the underlying shares. Assume a company’s shares are trading at $100 at grant date, yield on the most-recently issued ten-year Treasury note is 4.50 percent, the equity-risk premium is estimated at four percent, and the company must earn 50 percent of the premium before the options become profitable. The exercise price rises by 6.50 percent over the next year to $106.50 before consideration of dividends. If dividends paid during the year totaled $1.50 per share, the year-ahead exercise price would be $105.

Treasury notes provide a nominal return consisting of a real return, a return for expected inflation, and an inflation risk premium. Equity investors expect an extra return, the equity premium, above the Treasury yield as compensation for assuming greater risk. While the prospective equity premium is a crucial input for establishing investment hurdle rates and asset allocation policy, various approaches (surveys, extrapolation from historical excess returns, forward-looking estimates) employed by researchers lead to estimates ranging from near-zero by Arnott and Bernstein39 to as high as 6 or 7 percent. Siegel (2002) estimates the average equity premium over the past 200 years at 3.50 percent and 5 percent since 1926. He forecasts an equity premium in the range of 2 to 3 percent which is in line with most other forecasts.

37 For a discussion of discounted index options see Rappaport (1999).38 A number of companies including IBM, Yahoo, Office Depot, and EDS have adopted premium-priced option plans that target a higher level of performance than standard fixed-priced options. IBM, for example, announced in February 2004 that its annual grants to senior executives will have a strike price 10 percent above the market price at grant date and remain fixed over the ten-year life of the options. The share price must rise a meager one percent annually for executives to realize gains. Because strike prices are fixed, premium-priced options hold no guarantee that the level of required performance will turn out to be superior. During a sustained period of rising markets such as occurred in the late 1990s, premiums of 25 to 50 percent on ten-year options still reward below-average performance if the stocks of peers are appreciating at a double-digit rate annually.39 Arnott and Bernstein (2002).

33

Choosing an equity-premium rate for a DERO plan becomes a much less daunting task if corporate directors recognize that nobody can reliably predict future return spreads between stocks and Treasury notes, that “expert” forecasts tend to cluster in a relatively narrow range, and that longer vesting periods for options and holding periods for company shares mitigate forecast risk as noise decreases with time. But any forecast “error” pales by comparison to the failure of standard options to incorporate any shareholder opportunity cost, not even the risk-free rate.40

Equity investors expect a minimum return consisting of the risk-free rate plus the equity-risk premium. Following this line of reasoning the exercise price of options should rise at a rate no less than at this cost of equity. However, the threshold level of performance required by cost-of-equity adjusted options, like indexed options, causes many executives to hold underwater options that can encourage them to take unwarranted business risks or simply lose financial motivation. Properly designed incentives consider the delicate tradeoff between setting performance at levels that compensate shareholders for taking on equity risk and the need to keep executives motivated. That’s the purpose of the discounted equity-risk component of DEROs. If the board decides that only a fraction of the estimated equity risk premium will be incorporated into the exercise price growth rate, it is betting that the value added by management will more than offset the costlier options granted. Finally, dividends are deducted from the exercise price to remove the incentive for companies to hold back distributions to shareholders when there are no value-creating investment opportunities.

Many companies have followed Microsoft’s July 2003 decision to shift from stock options to restricted stock grants. The impending FASB standard which will require expensing of option costs, thus placing options on a level playing field with stock grants, has fueled additional interest in stock grants. Following the lead of other companies or choosing a plan because of its more favorable earnings consequences are improbable means for finding the most efficient form of executive compensation. Nonetheless, because of their growing popularity it is useful to examine how well restricted stock grants promote long-term value-maximizing behavior compared to stock options. Discounted index options and DEROs rather than less efficient standard options are the appropriate standards of comparison with restricted stock.

40 Individual stocks can be more or less risky than the market. High-technology firms, for example, can choose to increase the equity-risk premium or reduce the number of options granted to offset the greater value of high-volatility options.

34

Stock grants motivate key executives to stay with the company until the restrictions lapse, typically in three or four years, and they can cash in their shares. These grants are a compelling incentive for CEOs and other executives who influence the stock price to play it safe, protect existing value, and avoid getting fired. Not surprisingly, restricted stock plans are commonly referred to as “pay for pulse” rather than for performance.

Restricted stock grants are essentially options with an exercise price of zero. Because standard options have a positive exercise price (market price at grant date) they are substantially more risky and less valuable than restricted shares. To provide equal value executives generally get only one share of restricted stock for every three or four options. Indexed options and DEROs raise the performance bar further and would require an even greater number of options for each share of restricted stock.

Are executives better off with restricted stock? There is minimal downside since they realize gains even if the stock price falls below the grant price. On the other hand, because of the larger number granted, options provide a greater upside than stock grants if the company’s stock price performs well. Suppose the board offers the CEO a choice of 20,000 restricted shares or three and a half standard options for each restricted share for a total of 70,000 options. The company’s stock is trading at $40 per share. If the stock price rises 40 percent to $56 the CEO has the identical pre-tax gain.41 Assuming a ten-year option, the breakeven annual rate of stock price appreciation is less than 4 percent by the end of the tenth year. In other words, a CEO would realize greater gains from options if the stock price increased an average of 4 percent or more annually and would be better off with restricted stock for stock-price appreciation of less than 4 percent all the way down to a near-100 percent decline from the price at date of grant.

CEOs who choose restricted stock signal their expectation of low value-creation prospects. While many executives favor restricted stock, why would shareholders choose a restricted stock plan that rewards executives more generously than a stock options plan when the share price underperforms and provides lesser payoffs when the price performs at superior levels? The case against restricted stock as an incentive becomes even stronger when the comparison is made to indexed options or DEROs instead of standard options.

Hall suggests three advantages of restricted stock relative to options.42 The first is that restricted stock cannot fall underwater. Shareholders pay a

41 Holders of restricted stock must pay income taxes at ordinary rates when the shares vest. If executives sell shares to pay the tax bill the retention and motivational incentives are correspondingly reduced.42 Hall (2003).

35

very high price for this benefit. Specifically, they guarantee the grant-date value of shares even if there is zero price appreciation. More importantly, shareholders forego the positive incentives that more challenging performance targets in option plans provide.

Second, the value of equity compensation to risk-averse, undiversified executives is lower than the cost to shareholders. Hall estimates that executives holding restricted stock value it at 80 to 90 percent of its cost to shareholders, while they value standard at-the-money options at only 50 to 75 percent.43 He argues that because executives attach greater value to restricted stock it is a more cost-efficient compensation mechanism than stock options. Once again, this advantage must be weighed against the significantly greater incentive benefits of a well-structured stock option plan.

43 Comprehensive modeling and numerical results for the value of stock options from a manager’s perspective can be found in Lambert, Larcker, and Verreecchia (1991).

36

The third claimed advantage is that stock grants are much less complex to value and much more transparent. Options are more difficult to value, but CEOs need not be versed in the details of the Black-Scholes or binomial option-pricing models to be motivated by the potential financial payoffs. The arithmetic for calculating payoffs for a range of performance possibilities is straightforward.

In an effort to blunt the criticism that restricted stock plans are a giveaway, many companies offer performance shares which require not only that the executive remain on the payroll, but that the company achieve

37

predetermined performance goals. The most common are financial measures such as earnings-per-share growth, revenue targets, and return on capital employed. The number of shares payable to executives ordinarily depends on the extent to which the goals are met.44 While shareholders would probably favor performance shares over restricted stock, performance share incentives introduce a new set of problems. An incentive to maximize two- or three-year earnings, revenue, or return on capital can undermine the objective of maximizing long-term value. Earnings-per-share growth, for example, does not necessarily create shareholder value. Performance share 44 Some companies employ stock price appreciation or total return to shareholders as the criterion in performance plans. These plans essentially replicate the results of stock options.

ReferencesArnott, Robert D. and Peter L. Bernstein. 2002. “What Risk Premium is Normal? Financial Analysts Journal (March-April).

Barberis, Nicholas and Richard Thaler. 2003. “A Survey of Behavioral Finance.” In Handbook of the Economics of Finance. Elsevier Science B.V.: 1053-1116.

Benartzi, Shlomo and Richard H. Thaler. 1995. “Myopic Loss Aversion and the Equity Premium Puzzle.” Vol. 110, no. 1: 73-92.

Bernstein, Peter L. 2000. “A Modest Proposal: Portfolio Management Practice for Modern Times.” Economics and Portfolio Strategy (April 15).

Bernstein, Peter L. 2004. “Who Does What With Which to Whom, Or How Efficient Is the Market?” Economics and Portfolio Strategy (March 1).

Black, Fischer. 1986. “Noise.” The Journal of Finance, Vol. 41, no. 3 (July): 529-543.

Bogle, John C. 2004. “As the Index Fund Moves from Heresy to Dogma…What More Do We Need To Know?” The Gary M. Brinson Distinguished Lecture at Washington State University (April 13).

Degeorge, Francois, Jayendu Patel and Richard Zeckhauser. 1999. “Earnings Management to Exceed Thresholds.” The Journal of Business, Vol. 72, no. 1: 1-33.

Ellis, Charles D. 2004. Capital: The Story of Long-Term Investment Excellence. Hoboken, N. J.: John Wiley & Sons.

Fama, Eugene F. and Kenneth R. French. 2004. “Disagreement, Tastes and Asset Pricing.” Tuck Business School Working Paper no. 2004-03, (May).

Financial Accounting Standards Board. 1978. Statement of Financial Accounting Concept No. 1.

38

plans, unlike restricted stock, demand performance, but just not the right performance.

Conclusion

Recent governance reforms, including the Sarbanes-Oxley Act, fail to address the root cause of recent corporate scandals: the widespread obsession with short-term performance. There is no greater impediment to good corporate governance and long-term value creation than earnings

Graham, John R., Campbell R. Harvey and Shiva Rajgopal. 2004. “The Economic Implications of Corporate Financial Reporting.” Working paper. (January 20).

Grinold, Richard C. and Ronald N. Kahn. 2000. Active Portfolio Management, Second edition. New York: McGraw-Hill.

Grossman, Sanford J. and Joseph E. Stiglitz. 1980. “On the Impossibility of Informationally Efficient Markets.” The American Economic Review, Vol. 70, no. 3 (June): 393-408.

Hagin, Robert L. 2004. Investment Management: Portfolio Diversification, Risk, and Timing—Fact and Fiction. Hoboken, N.J.: John Wiley & Sons.

Hall, Brian J. 2003. “Six Challenges in Designing Equity-Based Pay.” Journal of Applied Corporate Finance, Vol. 15, no. 3 (Spring): 21-33.

Hall, Brian J. and Kevin J. Murphy. 2003. “The Trouble With Stock Options.” Journal of Economic Perspectives, Vol. 17, no. 3 (Summer).

Hirota, Shinichi and Shyam Sunder. 2004. “Price Bubbles Sans Dividend Anchors: Evidence from Laboratory Stock Markets.” Working Paper (May)

Jensen, Michael C. 2004. “Agency Costs of Overvalued Equity.” Negotiation, Organization and Markets, Harvard University Working Paper no. 04-26 (April).

Lambert, Richard A. 1991. “Portfolio Considerations in Valuing Executive Compensation.” Journal of Accounting Research, Vol. 29, no. 1. (Spring): 129-149.

Leibowitz, Martin L. 2004. Franchise Value: A Modern Approach to Security Analysis. Hoboken, N. J.: John Wiley & Sons.

Mauboussin, Michael J. 2002. “Revisiting Market Efficiency: The Stock Market as a Complex Adaptive System.” Journal of Applied Corporate Finance, Vol. 14, no. 4 (Winter): 47-55.

39

obsession. Nor is there a greater enemy of stock market allocative efficiency. Alleviating earnings obsession will not eliminate the occasional madness of crowds, but sensible investors looking for excess returns will bet against the madness and hasten the return to sanity. The potential payoff from reducing short-term performance obsession in the investment and corporate communities is substantial.

Earnings obsession and benchmark tracking are byproducts of short-termism, which in turn is triggered by difficult-to-monitor corporate and investment managers who as agents must provide frequent feedback on their ___________________. 2003.”Within and Without: Using the Collective to Find, Solve, and Predict.” The Consillient Observer, Vol. 2, no. 19 (October 21).

Rappaport, Alfred. 1999. “New Thinking on How to Link Executive Pay With Performance.” Harvard Business Review, (March-April):91-101.

__________. 2003. “How to Avoid the P/E Trap.” The Wall Street Journal, (March 4): R2.

__________. 2004. “Beyond Quarterly Earnings.” The Wall Street Journal, (March 8). R2.

Rappaport, Alfred and Michael J. Mauboussin. 2001. Expectations Investing: Reading Stock Prices for Better Returns. Boston, MA: Harvard Business School Press.

Shleifer, Andrei. 2000. Inefficient Markets: An Introduction to Behavioral Finance. Oxford, U.K.: Oxford University Press.

Shleifer, Andrei and Robert W. Vishny. 1997. “The Limits of Arbitrage.” The Journal of Finance, Vol. 52, no. 1 (March): 35-55.

Siegel, Laurence B. 2003. Benchmarks and Investment Management. Charlottesville, Virginia: Research Foundation of AIMR.

Sloan, Richard G. 1996. “Do Stock Prices Fully Reflect Information in Cash Flows and Accruals About Future Earnings?” The Accounting Review, Vol. 71, no. 3 (July): 289-315.

Stein, Jeremy C. 2004. “Why Are Most Funds Open-End? Competition and the Limits of Arbitrage.” Working paper (January).

Surowiecki, James. 2004. The Wisdom of Crowds. New York, N. Y.: Doubleday.

Treynor, Jack L. 1976. “Long-Term Investing.” Financial Analysts Journal, Vol. 32, no. 3 (May-June): 56-59.

40

performance to investor-owners. Disclosure and incentives that more closely align the interests of managers and owners are promising, but by no means perfect, ways of alleviating short-term performance obsession. Predicting the behavior of people who operate in a complex web of organizational relationships and an uncertain future is risky business. The incentives proposed in this paper are designed to defer a portion of investment and corporate managers’ compensation until at least some of the uncertainty surrounding their performance can be resolved. The expectation is that the recommended changes will produce more owner-friendly management behavior and a more allocatively-efficient market.

_____________. 1987. “Market Efficiency and the Bean Jar Experiment.” Financial Analysts Journal, Vol. 43: 50-53.

41