Earnings Management

EM2EarningsQuality.pdf

Home >Business & Finance homework help >Accounting homework help >Earnings Management

5. Earnings Management

Earnings management is the purposeful intervention in the external financial reporting process, with the intent of obtaining some private gain (as opposed to, say, merely facilitating the neutral operation of the process). (Schipper 1989, p. 92)

Earnings management occurs when managers use judgment in financial reporting and in structuring transactions to alter financial reports to either mislead some stakeholders about the underlying economic performance of the company or to influence contractual outcomes that depend on reported accounting numbers. (Healy and Wahlen 1999, p. 368)

Given these definitions of earnings management and our definition of earnings quality—that earnings reflect current performance, that earnings data are useful for predicting future performance, and that the earnings data accurately annuitize intrinsic firm value—clearly, earnings management decreases earn- ings quality.

Before discussing research into who manages earnings, what they man- age, and when, we first take a step back and try to put earnings management in the United States in perspective. How bad is earnings quality in the United States relative to elsewhere in the world? Next, we provide evidence on structuring transactions to manipulate earnings and discuss managing earn- ings by managing accruals. Finally, we provide evidence from research that has investigated companies’ incentives to engage in earnings management. To identify when managers face strong incentives to engage in earnings management requires close scrutiny of the financials by the analyst.

Global Earnings Quality The financial reporting system in the United States is based on more rules and has greater regulatory and informal monitoring structures than the systems in most other countries. But do these systems and structures result in less earnings management? Several studies have investigated this issue.

Leuz, Nanda, and Wysocki (2003) developed an earnings management score to measure earnings management in various countries. Their aggregate score uses four measures of earnings management: (1) the volatility of earn- ings relative to the volatility of cash flows, (2) the correlation between cash flows and accruals, (3) the extent of discretion in accruals based on the absolute magnitude of accruals relative to the absolute value of cash flows,

Earnings Quality

and (4) the extent of loss avoidance. A high aggregate score implies “high” earnings management. Figure 5.1 shows the aggregate score for various countries. The United States ranked as the country with the least amount of earnings management. Leuz et al. also showed that earnings management scores are lower in countries with large stock markets, dispersed ownership, strong investor rights, and strong legal enforcements (i.e., the United States, Australia, and the United Kingdom).

In a related vein, Hung (2000) found that countries with more accrual- related accounting standards have more useful (i.e., value-relevant) earnings when the country has strong shareholder protection (as in the United States). If earnings are more value relevant when accruals are used in the proper manner (i.e., to reduce mismatching problems in cash flows) and less value relevant when accruals are used to manipulate earnings, then an implication of Hung’s results is that the United States is likely to experience less earnings management than other economies.

Another approach to investigating the quality of financial reporting in the United States versus elsewhere is through the Form 20-F reconciliations mandated by the U.S. Securities and Exchange Commission (SEC). Foreign

Figure 5.1. Earnings Management around the World

Source: Leuz, Nanda, and Wysocki (2003).

Aggregate Earnings Management Score

Austr ia

Unite d Stat

Sin gap

ore

In dones

Fin lan

Gree ce

Ger m

an y

Thail an

South K

orea Jap

Pak ist

Portu gal

Belg iu

Neth er

lan ds

Ita ly

Hong K ong

Den m

ar k

Ta iw

an In

dia

M ala

ysia

Switz er

lan d

Spain

Fra nce

South A

fri ca

Phili ppin

Unite d K

in gdom

Swed en

Norw ay

Ire lan

Austr ali

Can ad

Earnings Management

companies listing American Depositary Receipts in the United States are required to reconcile their financial statements—based on their domestic generally accepted accounting principles—to U.S. GAAP. The reconciliation components can provide insight into which accounting system elicits earnings information that is the more relevant to investors. Amir, Harris, and Venuti (1993) documented that U.S. GAAP earnings are more value relevant to investors than a company’s domestic GAAP earnings. Their results suggest that financial statement quality is generally higher in the United States than elsewhere (see also Barth and Clinch 1996).

Note that the studies we discuss in the remainder of this chapter primarily used U.S. data. Thus, the results may well represent a lower bound on the extent of earnings management in other countries.

Manipulation of Real Transactions Earnings are composed of cash flows and accruals, and the manipulation of either component will affect the earnings number. A manager can take real economic actions that affect cash flows. Examples are cutting research and development (R&D) expenditures and boosting sales by offering products at a discount. The manipulation of real transactions is not a GAAP violation as long as the company properly accounts for the transaction. And these actions generally do not result in a qualified audit opinion or an enforcement action by the SEC. Nonetheless, such actions can have a significant impact on earnings quality and devastating effects on the company’s future performance, and the transactions are a form of earnings management.

Using publicly available data to document the extent to which companies engage in real transactions to manipulate earnings is difficult. Observing that a company enters into a transaction that receives favorable accounting treat- ment is not evidence that the company entered into the transaction just because of its accounting consequences. To show that a company engaged in a transaction purely for accounting purposes, one must know what the com- pany would have or should have or could have done as an alternative.

Several papers have provided compelling evidence about the manipula- tion of real transactions. Imhoff and Thomas (1988) documented that the use of capital leases dropped sharply after the effective date of Statement of Financial Accounting Standards (SFAS) No. 13, Accounting for Leases (FASB 1976), which mandated capitalization for leases with particular terms.19

Significant substitution into operating leases occurred. Hand (1989) found that companies undertake debt-to-equity swaps to mitigate what would be

19FASB statements can be found at www.fasb.org/st/.

Earnings Quality

an unexpected and transitory decline in earnings per share (EPS). Bartov (1993) showed that companies time asset sales to smooth intertemporal earnings changes and also to prevent debt covenant restrictions. Dechow and Sloan (1991) showed that R&D-intensive companies are more likely to cut R&D expenditures just before the chief executive officer’s retirement, and Bushee (1998) showed that companies that have little institutional ownership are more likely to cut R&D when earnings are abnormally low.20

None of these actions violate GAAP, but they certainly affect earnings quality.

Accrual Manipulation The second way for managers to produce a desired earnings number is to manage accruals. In this method, the company does not change its activities but, rather, opportunistically reports income for an existing activity. Examples that increase income are reducing the allowance for doubtful accounts, capi- talizing rather than expensing costs, and avoiding write-offs of assets.

Accruals create the opportunity for earnings management because they require managers to make forecasts, estimates, and judgments. The greater the degree of discretion in an accrual, the greater the opportunity for earnings management. Consider the discretion available in the typical accrual adjust- ments to the following asset and liability accounts. • Accounts receivable. Managers forecast expected product returns and the

proportion of customers who will not pay (high discretion). • Inventory. Managers capitalize some costs in inventory and expense other

costs as periodic expenses. They forecast expected demand in order to determine future sales prices and whether a write-down is necessary (high discretion).

• Other current assets. This account is typically a catchall category for capitalized costs (high discretion).

• Property, plant, and equipment (PP&E). Managers capitalize a multitude of costs and depreciate them in arbitrary ways. Managers must also forecast future demand to determine whether an impairment has occurred (high discretion).

• Accounts or interest payable. These accounts are amounts owed in dollars to suppliers or debtors (low discretion).

20We provide further details about the Bushee study in Chapter 6 in the discussion of the role of institutional investors as monitors of earnings quality.

Earnings Management

• Pension liabilities and postretirement benefits. Managers must forecast the expected return on plan assets, obtain actuarial assumptions on life expectancies, and so forth (high discretion).21

• Long-term debt. Value is based on the amount received at issuance of long- term debt, and the premium or discount is amortized under specific rules (low discretion). Accruals management affects only the timing of the recognition of earn-

ings. An overstatement of earnings in one period implies an understatement of earnings in another. For example, an understatement of bad-debt expense in one period (overstated earnings) will result in a write-off of accounts receivable in excess of the allowance (understated earnings) in another. Hence, earnings management strategies based on accruals management are rational only if the expected costs associated with the reversal do not dominate the expected benefits of the initial accruals management. In many cases, the only rational explanation for observed accruals management is managerial optimism: Man- agers must believe that the accrual reversals will go undetected in future periods when earnings are sufficiently high to absorb the reversals.

Earnings management by opportunistically reporting accruals is not nec- essarily a violation of GAAP. In many cases, companies can choose among accounting methods. In the spirit of high-quality financial reporting, the right choice is the one that best reflects the economics of the underlying transaction (such as using accelerated depreciation for long-term assets that lose more value early in their service lives). In many cases, however, the company is free to choose among methods without economic justification (e.g., using straight- line depreciation instead of accelerated depreciation).

One result found in many studies is that high accruals, in absolute magni- tude, are a potential “red flag” that companies are engaging in earnings man- agement. Dechow, Sloan, and Sweeney (1996) found that companies that were subject to enforcement actions had higher accruals than a control group. Richardson, Tuna, and Wu (2003) found that companies that restate earnings have high accruals in periods before the restatement. Their sample consisted of 338 companies (452 company-year observations) that restated earnings because of accounting abuses. The authors ranked the companies in deciles based on total accruals relative to all companies in the Compustat database. Decile 10 companies had the highest level of accruals (scaled by assets). The results are presented in Figure 5.2. A much larger proportion of companies

21Amir and Benartzi (1998) found that income manipulation affects the selection of the expected rate of return (ERR) on pension assets. They also documented that the allocation of assets between stocks and bonds is a better predictor of future fund performance than the ERR. Amir and Gordon (1996) found that companies manage pension-related estimation parameters, such as discount rates and trends in health care costs, and showed that the management was done to avoid violating debt covenants and to increase the company’s bargaining power in plan renegotiations.

Earnings Quality

that subsequently restated earnings had high accruals. The restatements were probably reversing prior overstated accruals.

Our analysis of SEC Accounting and Auditing Enforcement Release (AAER) No. 1 through AAER No. 1745 provides evidence as to which specific accounts are managed. We focused on releases that involved manipulations of annual financial statements and identified 294 separate companies that had manipulated 426 different accounts. Figure 5.3 summarizes the accounts that the SEC most frequently alleged were manipulated.

Overstatement of revenues is the most common type of earnings manage- ment; 70 percent of the cases involved overstated revenue.22 Further analysis of the revenue overstatements highlights that the SEC is concerned about revenue recognition even if the overstatement of revenue does not affect bottom-line earnings. Some companies are judged on revenue growth, which creates incentives for the companies to manage revenues. One example of boosting revenues without boosting the bottom line is the reporting of barter advertising revenue on a gross basis. The Emerging Issues Task Force (EITF) recently took action to limit when companies can report advertising barter

Figure 5.2. Frequency of Future Earnings Restatement by Accrual Decile

Source: Richardson, Tuna, and Wu (2003).

22A study by the SEC involving 38 enforcement releases indicated that common revenue recognition problems include recognizing revenue on consignment sales, sales to related parties, bill-and-hold transactions, sales in which ownership had not passed to the customer, shipments not ordered by customers, and nonqualifying barter transactions. Other overstatements related to fictitious sales and delayed recognition of returns. See www.pobauditpanel.org/downloads/appendixf.pdf.

Number of Restatements

0 Low High3 6 92 54 7 8

Accrual Decile

Operating Accruals Investing Accruals Total Accruals

Earnings Management

transactions on a gross basis (EITF Issue No. 99-17, Accounting for Advertising Barter Transactions). Their action was concurrent with the SEC’s expressing concerns about these transactions. Another example is Qwest Communications International’s accounting for bandwidth swaps so as to allow the company to show continued sales growth.23 Subsequent restatements of the revenue rec- ognition had no effect on bottom-line earnings.

Figure 5.3 shows that capitalizing costs as long-term assets or overstating PP&E occurred in about 20 percent of the companies. This observation is consistent with Richardson, Sloan, Soliman, and Tuna’s (2003) proposition that long-term operating asset and liability accounts (such as PP&E and long- term receivables) are subject to manipulation, despite claims by some that current accruals are easier to manipulate. Richardson et al. suggested that a growing gap between earnings and free cash flow (FCF) can be viewed as a red flag about earnings quality. Figure 5.4 shows the earnings-to-FCF gap for WorldCom from the early 1990s through 2001. The plot illustrates a growing gap through the year 2000, which investors now know was the result of capitalizing rather than expensing current costs.

Figure 5.3. Types and Frequency of Earnings Manipulations

Note: Based on 294 AAERs that involved the manipulation of 426 different accounts.

23See, AAER No. 1879 at www.sec.gov/divisions/enforce/friactions.shtml.

Overstated Inventory/

Understated Cost of Goods Sold

Overstated PP&E

Misstatement of Accrued Liabilities

Understated Expense

(other than cost of goods

sold)

Understated Accounts

Receivable Discounts/ Allowances Other Than Bad Debts

Overstated Intangibles and Other Long-Term

Assets

Overstated Revenue and

Accounts Receivable/ Understated

Bad-Debt Allowance

Overstated Security

Valuation

Percent

Earnings Quality

Understating expenses (other than cost of goods sold) accounted for almost 30 percent of the allegations (see Figure 5.3); overstating inventory (and understating cost of goods sold) accounted for approximately 10 percent. The SEC alleged in approximately 8 percent of the cases that the companies were creating hidden reserves. These cases involved the overstatement of liabilities, such as a restructuring accrual, to be drawn down in future periods to boost operating performance.

Nelson, Elliott, and Tarpley (2003) provided a different perspective on the various accounts most frequently managed. Their sample consisted of 515 attempts at earnings management obtained from a survey of 253 experienced auditors from one of the Big Five accounting firms. They reported the catego- ries of earnings management attempts shown in Table 5.1.

Their analysis indicates that 272 of 515 (53 percent) of the earnings management attempts were made to increase income and that the auditor adjusted for these income-increasing attempts in 52 percent of the cases. The most common type of earnings management attempt was an attempt to adjust an expense or loss through the manipulation of reserves or through the capitalization or deferral of costs (269 cases). These attempts include record- ing inappropriate reserves for such items as restructuring charges, provisions for loan losses, write-offs of inventory, and asset impairments. These items are generally classified as “special items” in the income statement, and the measurement of the charge involves significant estimation and judgment, which creates opportunities for earnings management.

Figure 5.4. WorldCom Earnings and Free Cash Flows, December 1992– December 2001

Dollars (millions)

6,000

2,000

−2,000

4,000

−4,000

−6,000

−8,000

−10,000 12/92 12/0112/94 12/97 12/0012/93 12/9612/95 12/98 12/99

Earnings Net Operating and Investing Cash Flows

Earnings Management

In the Nelson et al. sample, the second most common type of earnings management attempt was adjusting revenue or other gains (114 attempts). The cases involved cutoff manipulation (booking sales after the fiscal year- end), recording bill-and-hold sales, and recording sales with a right of return. Of these attempts, 86 of 114 (75 percent) would have increased income, and the auditor required an adjustment in 62 percent of the cases.

The results of SEC enforcement actions and audited earnings manage- ment attempts suggest that an analyst should carefully scrutinize accounts receivable (including bad-debt allowances), inventory, capitalized costs for PP&E, intangible assets (such as software development), and liabilities that are difficult to estimate. All these accounts require management to make forecasts and estimates, and the company can use this discretion to boost current or future earnings.

Beyond focusing on the level of specific accruals, how can an analyst detect earnings management? Accounting research has put considerable effort into producing models to identify legitimate (“nondiscretionary”) accruals versus managed (“discretionary”) components. A common model used to detect discre- tionary accruals or earnings management is the Jones (1991) model. Using time- series data for a company, this model estimates expected accruals as follows:24

Accrualst = � + �1(�Revenue) + �2(PP&E) + �t. (5.1)

Table 5.1. Number of Earnings Management Attempts and the Percentage Adjusted by Auditor

Approach Total

Attempts Attempt to

Increase Income

Percent of All Attempts

Adjusted by Auditor

Percent of Increase Attempts

Adjusted by Auditor

Attempt to adjust an expense or other recorded loss 269 133 42% 48%

Attempt to adjust revenue or other recorded gain 114 86 56% 62%

Business combination 67 12 40% 67% Other approaches 65 41 34% 41% Total/Percent adjusted 515 272 44% 52%

Source: Table 1 of Nelson, Elliott, and Tarpley (2003).

24Some papers (e.g., DeFond and Jiambalvo 1994) estimated this model for a sample of companies in a single industry rather than for a single company across time periods. Such cross- sectional estimation assumes that the relationships among revenues and PP&E and accruals are relatively stable within the group of companies.

Earnings Quality

All the variables in this model are scaled by assets. Expected accruals, which are assumed to be nondiscretionary accruals, are equal to � + �1(�Revenuej) + �2(PP&Ej), where �, �1, and �2 are estimates from the regression (Equation 5.1). Discretionary accruals are equal to �, the residual from the regression.25

The main criticism of this model is that it does not control for growth. If accruals do not move one for one with sales and PP&E as these variables grow, the estimate of discretionary accruals will be measured with error. Our reason for including the model here is that it highlights the importance of appropriate benchmarks for evaluating accruals. When examining accruals, analysts need to consider changes in fundamental variables, such as revenues and long-lived assets, that reflect changes in the underlying operations of the company; these fundamental changes clearly have an impact on nondiscretionary accruals. A company’s change in accruals relative to changes for other companies in the same industry is also an important consideration. Other companies provide a useful benchmark for pinpointing what is nondiscretionary.

Identifying the discretionary component of accruals by using regression analysis can be difficult in practice. In many circumstances that an analyst faces, focusing on the level of total accruals is a simple and easy way to identify discretionary accruals. Total accruals and discretionary accruals are highly positively correlated. Dechow, Richardson, and Tuna (2003) showed that the correlation between estimated discretionary accruals (from accrual models) and total accruals is higher than 80 percent.

Incentives for Earnings Management Earnings management requires opportunity and motive. In this section, we analyze situations that create motives for earnings management. By understand- ing when companies have incentives to manage earnings, readers of financial statements can assess when a company is likely to engage in such behavior. Note that the earnings management decision represents a trade-off between costs and benefits. This discussion focuses on the benefits of earnings management.

Dechow, Sloan, and Sweeney (1996) studied 92 AAERs that accused companies of engaging in earnings manipulation between 1978 and 1990. Thirty-nine of the AAERs provided at least one explanation for the earnings management; the remaining fifty-three AAERs provided no explanation.

25 Dechow, Sloan, and Sweeney (1995) suggested the following “modification” to the original Jones model. If manipulation is predicted to occur in year j, then discretionary accruals in that year are calculated as Accrualsj – [� + �1(�Revenuej –�Accounts receivablej) + �2(PP&Ej)]. This modification increases the power of the model to detect revenue overstatements because it does not assume, as the original Jones model did, that all changes in revenue are nondiscretionary.

Earnings Management

Table 5.2 contains the results of their study. They suggested that three main factors create incentives for earnings management. First, capital market transactions are associated with earnings management. In more than half of the cases examined, managers manipulated earnings either to issue securities at higher prices or to profit from insider trading. Second, the desire to report upwardly trending earnings per share, presumably to meet the expectations of analysts and investors, is a motivation for engaging in earnings manage- ment. Third, contractual incentives are motives for earnings management— for example, manipulating earnings to increase the size of earnings-based bonuses. The remainder of this chapter discusses empirical research into each of these incentives.

Capital Market Incentives. Research has documented the existence of earnings management in connection with • seasoned equity offerings, • initial public offerings, • mergers and management buyouts, and • insider equity transactions. The incentive in these instances is to affect the company’s stock price. Stock is currency in these transactions, and price matters. Moreover, short-term stock price matters. Even if managers know that the effects of the earnings management will eventually reverse, they have incentives to manage earnings in the current period to manipulate the transaction price. The greater stock price associated with the managed earnings, even if only in the short run, reduces the cost of acquiring new capital or the effective price of an acquisi- tion, or it increases the manager’s personal wealth. (An underlying assump- tion of this research is that investors do not rationally anticipate the earnings management. If they did, managers would have no reason to engage in the behavior and bear the costs associated with potentially getting caught.)

Table 5.2. Motivations for the Manipulation of Earnings

Motivation Number Issue securities at higher prices 22 Report upwardly trending EPS 11 Increase the size of earnings-based bonuses 7 Profit from insider trading 6 Other 3

Totala 49

aEight AAERs provided more than one explanation.

Earnings Quality

An important aspect that the research into capital market incentives for earnings management highlights is that earnings management may involve artificially decreasing instead of artificially increasing earnings. Management buyouts are an example of a capital market transaction that provides incentives for a manager to manage earnings in order to decrease the company’s stock price—at least in the short run—and reduce the buyout price.

■ Seasoned equity offerings. SEOs provide an incentive for companies to manage earnings to increase stock prices prior to the offerings. And research findings are consistent with the intuition that some companies engage in earnings management prior to SEOs (Rangan 1998; Teoh, Welch, and Wong 1998b). Figure 5.5 shows returns for four groups of companies classified by their pre-issue levels of discretionary current accruals (DCA) from Teoh, Welch, and Wong (1998b).26 Companies were classified into quartiles based on discretionary current accruals in the fiscal year prior to the offering. The

26The authors considered current assets and liabilities (excluding the current portion of long- term debt) to be easier to manipulate than long-term accruals. Thus, their measure of discretionary accruals was changes in current accrual assets and current accrual liabilities that were not related to changes in sales and PP&E, where the estimated relationships between the accruals and the other measures were assumed constant within an industry.

Figure 5.5. Stock Returns Following SEOs

Source: Teoh, Welch, and Wong (1998b). The sample consists of 1,248 SEOs between 1970 and 1989.

Fama−French Excess Return (%)

−5

−15

−25

−35

−45

−10

−20

−30

−40

−50 −12 606 24 423 21 3918 36−3 15 33−6 12 30 5754514845−9 9 27

Event Month

Conservative Quartile 1 Quartile 2 Quartile 3 Aggressive Quartile 4

Earnings Management

graphed returns are the logged cumulative sum of monthly portfolio excess returns relative to the Fama–French (1993) model, normalized so that the offering month return is zero.27

Figure 5.5 shows two important patterns. First, all the companies making SEOs had increasing returns prior to the offering. The similar return perfor- mance across the groups suggests that investors did not recognize any differ- ence in the quality of earnings among the four groups prior to the offering. This result is important because the companies were classified by using publicly available data. Thus, investors could have identified the aggressive and conservative companies as Teoh, Welch, and Wong did. Second, the most aggressive earnings managers experienced the worst underperformance after the SEO. Eventually, the low quality of earnings for the Aggressive Quartile 4 portfolio was revealed, probably as the discretionary accruals were reversed.

Teoh, Welch, and Wong (1998b) interpreted the observed patterns as evidence that companies manage earnings through accruals prior to an SEO and that the market is fooled by the earnings management. Investors overreact to the managed earnings and are subsequently “disappointed” by poor earnings performance when the accruals reverse. Rangan supported these findings. Shi- vakumar (2000), however, argued that these tests are misspecified; he believed that investors rationally anticipate the earnings management around an SEO and “undo” its effects. An alternative explanation for the apparent overreaction to earnings prior to an SEO and the subsequent disappointment is that the high- accrual companies, which the studies identified as earnings managers, are really high-growth companies, for which the models of accruals do not fit the data well. High-growth companies are likely to have the most optimistic forecasts of future growth, and the market is subsequently disappointed when that growth is not realized. In a follow-up paper, Teoh and Wong (2002) found that discretionary accruals in the offering year predict analysts’ errors in annual earnings forecasts for as many as four fiscal years after the new issue. The discretionary accruals also predict the analysts’ errors in five-year growth forecasts made in the offering year. Analysts and investors appear to suffer from similar biases.

■ Initial public offerings. A company’s initial public offering (IPO) also provides incentives for earnings management, and the opportunity may be even greater than the opportunities around an SEO. IPOs typically have a shorter operating history than SEOs. Even for old companies, the amount of publicly available historical financial information is smaller for IPOs than for SEOs. Thus, it is more difficult for market participants to estimate for IPOs the nondiscretionary or unmanaged portion of accruals and the managed

27The Fama–French classic model of expected returns consists of three factors: (1) the market return (return on a value-weighted market index less a risk-free rate of return), (2) company size (small minus large), and (3) company book-to-market value (low minus high).

Earnings Quality

portion. Because the probability of fooling investors is greater with IPOs, the expected benefits of managing earnings are greater and are more likely to outweigh the costs.

As for SEOs, research evidence consistently indicates that companies manage earnings by using discretionary accruals prior to IPOs. The accruals eventually reverse, however, which explains (at least in part) the long-run underperformance of IPOs (Teoh, Welch, and Wong 1998a; Teoh, Wong, and Rao 1998; DuCharme, Malatesta, and Sefcik 2001). Aharony, Lin, and Loeb (1993), however, did not find much evidence that IPO issuers make unusual accounting choices that would be consistent with earnings management.

■ Mergers and management buyouts. Around a merger, both the acquirer and the target have incentives to manage earnings to increase stock prices, and evidence shows that both parties do. Erickson and Wang (1999) found evidence consistent with income-increasing earnings management by acquiring companies in stock-for-stock mergers between 1985 and 1990. Easterwood (1998) found evidence that takeover targets—primarily targets of hostile takeovers—use discretionary accruals to inflate earnings (compared with the discretionary accruals by nontarget companies) during the quarters immediately preceding and following a takeover attempt. Christie and Zimmerman (1994) also showed that targets more frequently select income- increasing accounting methods than do nontargets for depreciation, inventory, and investment tax credits. The authors suggested that efficiency, however, rather than opportunism is the primary driver in the accounting choice. The evidence is not conflicting: Companies may be more likely to manage discretionary accruals, which may go unnoticed, than to manage a publicized accounting choice, such as its depreciation method, which investors can easily observe.

In a management buyout, in contrast to a takeover, managers have incentives to minimize the purchase price, potentially through income- decreasing earnings management that will negatively affect the company’s stock price. Two papers on management buyouts during the 1980s provided evidence of earnings management prior to buyouts. Perry and Williams (1994) found abnormal negative behavior of discretionary accruals for buyout com- panies in the year before managers announced their buyout intentions. Wu (1997) reported that earnings changes are significantly smaller than industry median changes for buyout companies in the year preceding the buyout. DeAngelo (1986), however, did not find evidence of accrual management for management buyouts during the 1973–82 period. DeAngelo’s findings may be a result of the different time period examined, the different research method used, or the use of a less powerful earnings management proxy.

Earnings Management

■ Insider equity transactions. Insiders have incentives to increase earnings to artificially inflate stock prices, even if only in the short run, around the time they anticipate selling shares. Beneish and Vargus (2002) provided two kinds of evidence related to earnings management at the time of insider equity transactions. First, they documented that it exists. When income- increasing accruals occurred contemporaneously with unusual insider trading activity, the accrual-related earnings had lower persistence to the one-year- ahead accounting period. The lower persistence was at least partly a result of accruals management. Second, they documented that a hypothetical long– short trading strategy that takes positions based on the direction of companies’ accruals and contemporaneous insider trading activity earns positive hedge returns. This evidence does not suggest that investors can earn abnormal returns by forming portfolios on the basis of insider trades and accruals; investors do not have information about a contemporaneous insider trade until after the trade is made. The evidence does suggest, however, that public information about insider trading activity, which is disclosed on or before the 10th day of the month after the event, can help analysts and investors assess the probability that the accruals prior to the insider’s sale or purchase were likely to have been managed.

■ Summary. The key point of this section is that any event that involves stock being purchased or sold or otherwise used as currency provides managers with incentives to manage earnings to influence investor perception of firm value. We pointed out that companies that are issuing new financing and that also have high accruals are the most overvalued by investors. Whether this overvaluation stems from earnings management or from investors’ misunderstanding of the growth potential of high-accrual companies is an open question. In either case, analysts should understand that these companies are likely to have low-quality earnings that should be scrutinized carefully.

Managing Earnings to Forecasts and Other Targets. Identifying the targets toward which companies manage earnings can help identify companies that are likely candidates for managing earnings. Companies that report earn- ings that barely meet or beat a target may have managed the earnings to do so.

The literature related to earnings targets is extensive and varied. One target that has been considered is simply zero. This research claims that companies do not want to report negative earnings and will manage earnings just enough to get over the zero threshold. In other words, managers believe there is a stigma to reporting negative earnings. Reporting positive $0.01 per share may be worse than reporting positive $0.02, but reporting –$0.01 is significantly worse than reporting $0.00.

Earnings Quality

Hayn (1995) and Burgstahler and Dichev (1997b) showed that the fre- quency with which companies report earnings that are barely less than zero is lower than the frequency with which companies report earnings just greater than zero. The histogram of earnings observations in Figure 5.6 illustrates this result. One interpretation of the observed pattern in earnings realizations is that earnings that are just above the target are managed.

Although the pattern in Figure 5.6 is striking, it is not a priori evidence of earnings management. In fact, researchers who have attempted to verify that the companies with earnings barely over the zero threshold were managing earnings have not been able to do so by using statistical models of discretion- ary accruals. Dechow, Richardson, and Tuna (2003) suggested other potential explanations for the observed pattern. For example, managers and employees may simply work harder to improve company performance when they are close to the zero threshold. And sample selection biases could also play a role. Beaver, McNichols, and Nelson (2003) also proposed alternative explanations for the pattern. They claimed that asymmetrical tax treatment of profits and losses and conservatism related to the reporting of special items can cause the kink at zero. So, although certainly not all companies with earnings just over the zero threshold have managed earnings to get there, low but positive earnings are a potential red flag for earnings management.

Figure 5.6. Distribution of Earnings Scaled by Market Value (Truncated), Data for 1988–2000

Notes: The sample was 47,847 company-years. Bolded classes represent income/market value that is just above zero Source: Dechow, Richardson, and Tuna (2003). Burgstahler and Dichev (1997b) and Hayn (1995) contain similar figures.

Number of Company-Years

Net Income Class

1,800

1,600

1,400

1,200

800

400

1,000

600

200

0 −30 290−10−20 10 20

Earnings Management

Researchers also have investigated whether companies manage earnings to meet analysts’ consensus forecasts. Figure 5.7 provides a plot of forecast errors, which represent the difference between the actual earnings realization and the consensus forecast, for annual forecasts (the figure is similar when quarterly data were used). A positive forecast error indicates that the actual earnings beat the forecast. The plot reveals an obvious pattern: A larger-than- expected number of companies meet or barely beat analysts’ forecasts. There are more small positive forecast errors than small negative forecast errors. This pattern would not be expected if analysts were equally likely to overesti- mate or underestimate earnings and if companies did not engage in earnings manipulation to exceed a target.

Again, the observed pattern in forecast errors is suggestive (but not conclusive) that companies engage in earnings management to meet or beat forecasts. The question of whether companies manage earnings to meet analysts’ forecasts is complicated, however, because the target (unlike the number zero) is not exogenous. A company might meet analysts’ forecasts because analysts are good at forecasting earnings, because the company’s earnings are easy to forecast, or because managers provide “guidance” to analysts about expected earnings. Nonetheless, the fact that a company consistently just meets the consensus forecast is a potential red flag that earnings quality is low.

Figure 5.7. Distribution of Analyst Forecast Errors

Notes: The distribution is as reported in I/B/E/S. The interval width is 1 cent; for example, the number at –30 includes all company-years when the consensus forecast was missed by 30 cents. Source: Dechow, Richardson, and Tuna (2001).

Number

Forecast Error Interval

6,000

5,000

4,000

3,000

1,000

2,000

0 −30 120−12−24 6−6−18 18 3024

Earnings Quality

This red flag is especially relevant for high-growth companies. The stock prices of these companies are particularly hard hit by investors when the companies subsequently miss the consensus forecast (Skinner and Sloan 2002). Not every company that barely meets its forecast, however, is equally likely to have a precipitous stock price decline, which the Skinner and Sloan study called an earnings “torpedo.”28 Torpedoes are more likely when a company does not meet revenue growth expectations but still manages to meet earnings expectations. Such a company is potentially sacrificing future earnings to boost current earnings by understating expenses.

Kasznik (1999) found, in a related study, that companies manage earn- ings toward their own annual earnings forecasts. Kasznik documented abnor- mally high positive discretionary accruals when earnings would otherwise have been below management forecasts. He interpreted this finding as evidence that managers manage reported earnings toward their own fore- casts, but he also recognized that the abnormal level of discretionary accruals might have motivated the issuance of the forecast or that the two could have been simultaneously determined as part of an overall reporting strategy.

A final issue to consider when thinking about how targets provide incentives for earnings management, and thus affect earnings quality, is why companies care about meeting targets. Trying to meet targets solely through earnings manipulation is not rational if investors assume that earnings that barely meet or beat a target are managed. In that case, rational investors will ascribe less value to the earnings that are assumed to be managed, which in turn means managers have no reason to manage the earnings in the first place. Some recent evidence indicates, however, that investors do not dis- count earnings that are just over a target (see, for example, Burgstahler and Eames 2003; Bartov, Givoly, and Hayn 2002; Kasznik and McNichols 2002;

28The earnings announcement of F5 Networks provides an interesting example of the effects of missing an analyst forecast. On Tuesday, 25 July 2000, F5 announced for the first time that it would record tax expense and that its EPS (after taxes) would be 17 cents. The stock price was $42.50. The consensus analyst forecast, according to First Call, was 18 cents, and a –1 cent earnings surprise was reported. The financial press reported that missing the analysts’ consensus sent the stock price “reeling” in after-hours trading; the opening price on 26 July was below $39. That day, First Call announced that it had aggregated analysts’ estimates of pretax and post-tax earnings for F5 in error. The new pretax estimate was 19 cents, and the after-tax average was 14 cents. As a result, the company beat the consensus by 3 cents. Its share price rose to $46. In this example, investors appear to have been reacting more to whether the company beat the consensus than the fact that it had an EPS of 17 cents.

Earnings Management

Matsumoto 2002; Dhaliwal, Gleason, and Mills 2002; Abarbanell and Lehavy 2003).29 The evidence suggests that there are some real benefits to meeting forecasts and that, whether the benefits are real or perceived, managers attribute value to meeting forecasts.

Contracting Incentives for Earnings Management. Various types of contracting arrangements provide companies with a direct incentive to man- age earnings. Debt contracts, compensation contracts, and tax regulations or other regulations—which we will loosely refer to as “contracts”—may contain provisions that are a function of GAAP earnings and thus provide incentives for earnings management. A company’s regulatory environment also can contain earnings management incentives—not only because some regula- tions, such as capital requirements, are explicitly based on earnings but also because a company’s profitability may be associated with regulatory pressure.

A significant distinction of contract-based incentives for earnings manage- ment is that the motivation is obvious. The contract specifies a particular earnings-based number the company must meet to avoid costs or to gain some benefit. The incentive to manage earnings is direct; no assumptions or predic- tions are needed about who is rational and what managers, investors, or contracting parties believe about the rationality of the other parties. If the contracting parties detect or undo the manipulation, it will not matter.

Consider capital regulations as an example. Banks that violate capital requirements incur both out-of-pocket and opportunity costs. For a signifi- cantly undercapitalized bank, regulators can require recapitalization or can force the institution into conservatorship or receivership. Banks that fail to meet minimum capital requirements must submit a comprehensive capital restoration plan to regulators, which is costly to prepare and implement. In addition, during the time that a bank is undercapitalized, its regulators can restrict dividends and management fees. The regulators can also exercise control over the bank’s operations by placing limits on branching, expansion, and new services. Even for banks with capital above the minimum require- ment, greater capital creates a competitive advantage. Well-capitalized insti- tutions face fewer regulatory constraints on operations than poorly capitalized institutions, enjoy more timely approval for expansion and growth from federal banking agencies, and pay lower Federal Deposit Insurance

29One explanation for the fact that investors do not discount earnings that are just over a target is that it takes time for them to distinguish manipulators from nonmanipulators. Balsam, Bartov, and Marquardt (2002) investigated the stock price reactions to 10-Q filings and the use of discretionary accruals to meet or just beat analyst forecasts. Companies that met a forecast by using discretionary accruals had a negative stock price reaction at the 10-Q filing date, which is presumably when the market recognizes the larger-than-expected accruals.

Earnings Quality

Corporation premiums.30 Thus, even if one assumes that bank regulators are perfectly able to anticipate and detect earnings management and, therefore, understand that capital is above the requirement only because earnings are of low quality, banks still have incentives to manage earnings to be above the requirement in order to avoid regulatory costs that are a function of strict capital requirements.

When thinking about the incentives that a contracting arrangement pro- vides for earnings management, analysts need to consider the specific nature of the arrangement and the type of incentives it provides. Not all contracting arrangements provide incentives to maximize reported earnings. For example, although for most financial institutions regulatory capital constraints provide incentives to inflate earnings to meet the constraint, institutions with capital that is significantly below the requirement may have incentives to engage in behavior that creates volatility; they are hoping for a big upside realization while not viewing a big downside realization as costly. That is, the company may try to maximize the value of the “put option” to the federal government that results from the existence of deposit insurance by increasing earnings volatility (see Ronn and Verma 1986). Contracts with creditors often include multiple covenants, and a company may have to manage particular compo- nents of earnings to meet all the covenants. As for taxes, companies have incentives to manipulate components of taxable income (not GAAP income), which often requires manipulation of real transactions—such as selling cer- tain securities that are in a gain or loss position—rather than accruals man- agement. Earnings management incentives derived from compensation contracts can be particularly complex. Bonus schemes, for example, can have a minimum threshold that must be met followed by a range over which the bonus is a linear function of earnings up to a predetermined maximum. Such nonlinearities in the relationship between compensation and earnings affect a manager’s incentives to manage earnings. And political costs often cause incentives for companies to manage earnings downward to look less profitable and more in need of “subsidies” or favorable antitrust decisions, union con- cessions, or import relief. In summary, understanding which companies have direct incentives to manage earnings is important because of contracting relationships and what the contract provides an incentive to do.

■ Capital/regulatory constraints. In general, the evidence is strong that financial institutions manage earnings to meet or beat regulatory capital requirements. The bulk of this research focuses on management of loan-loss provisions or insurance reserves in financial institutions (see, e.g., Ahmed,

30See the discussions of costs associated with violating bank capital requirements in Rose (1996) and Moyer (1990).

Earnings Management

Takeda, and Thomas 1999; Beatty, Chamberlain, and Magliolo 1996), primarily because researchers can confidently measure the discretionary component of accruals. They can observe ex post whether reserves were too high or too low. Evidence about other accounts, such as the valuation allowance against deferred tax assets (Schrand and Wong 2003), is more limited. All of the research reaches the same conclusion: Capital requirements provide strong incentives for earnings management. The costs of not meeting the requirement are significant, and all earnings, managed or not, help the institution meet the requirement.

The research about earnings management to meet capital and regulatory requirements clearly makes an important point: Although managing capital is important, companies also have other incentives for managing earnings and the various incentives may not all lead to the same optimal reporting strategy. Companies may want to manage regulatory capital, GAAP earnings, and taxable income simultaneously (see, for example, Collins, Shackelford, and Wahlen 1995; Chen and Daley 1996). The ultimate earnings management decision will represent a trade-off between the costs and the benefits associ- ated with all the earnings management incentives.

■ Debt covenants. Debt contracts commonly contain covenants based on accounting numbers or ratios. DeAngelo, DeAngelo, and Skinner (1994) and Sweeney (1994) found somewhat conflicting evidence on earnings management related to debt covenants. Both papers predicted that companies would engage in income-increasing behavior to avoid costly violation of debt covenants. DeAngelo et al., however, did not find evidence of such behavior for a sample of “financially troubled” companies (defined as those that reported at least three annual losses and reduced cash dividends in the six- year period of 1980–1985). The accrual behavior of the companies with binding covenants and those with nonbinding covenants was not significantly different. DeAngelo et al. provided some evidence that the more troubled companies with binding constraints had more negative accruals, but the accruals were related to inventory write-offs, which are probably a result of real troubles, not discretionary decisions.

In contrast, Sweeney found evidence that as companies approach violation of their debt covenants, they respond with more income-increasing account- ing changes. And Dichev and Skinner (2002) revealed unusual patterns in reported current ratios and net worth around the thresholds for these vari- ables set in private debt agreements.

Jaggi and Lee (2002) offered evidence that potentially reconciles the conflicting results. They found that some financially distressed companies use income-increasing discretionary accruals and some use income-decreasing

Earnings Quality

discretionary accruals to manage earnings. They observed income-decreasing accruals by companies that were not able to obtain waivers of the debt covenants and that were forced to renegotiate or restructure their debt, but they observed income-increasing accruals by companies that were able to obtain debt waivers.

One caveat about all of this research is that measuring discretionary accruals by using statistical models for companies in financial distress is difficult because the estimated model parameters may not be representative of the relationships within a particular distressed company or be stable through time.

■ Executive compensation contracts. A question that has been researched quite extensively is whether managers manipulate accruals to maximize earnings-based bonuses. Initial evidence by Healy (1985) supported this claim. He noted that earnings-based bonus plans frequently have upper and lower bounds. When earnings are below the lower bound, no bonus is awarded; when earnings are above the upper bound, no additional bonus is paid; and when earnings are between the bounds, the bonus is a function of earnings. For such arrangements, Healy predicted that managers had an incentive to increase earnings in order to increase the bonus only when unmanaged earnings were between the bounds. When unmanaged earnings were below the lower bound or above the upper bound, managers had an incentive to decrease earnings and “reserve” them for future periods when earnings were within the bonus range. His evidence is consistent with this prediction for a sample of 1,527 company-year observations for 1930–1980.

More recent research using different sample periods has had difficulty replicating Healy’s findings. For example, Gaver, Gaver, and Austin (1995) investigated bonus plans for 1980–1990 and found that companies with unman- aged earnings above the upper bound used income-decreasing discretionary accruals, which is consistent with Healy’s findings. They found, however, that companies with unmanaged earnings below the lower bound used income- increasing discretionary accruals, which is opposite to the results reported by Healy. The results found by Gaver et al. suggest that companies manage earnings to smooth the earnings series rather than to maximize bonuses. Holthausen, Larcker, and Sloan (1995) used a confidential dataset for which the lower and upper bounds of the contract were known. They found evidence of earnings management to decrease earnings and create reserves when earnings were greater than the upper bound but found no evidence that earnings were managed downward when earnings were below the lower bound.

In summary, although it seems intuitive that earnings-based bonuses would provide an incentive to engage in earnings management, recent

Earnings Management

research has not been able to document this effect or the extent to which it might occur. One explanation for the lack of evidence relates to the research methods: That is, earnings management may exist, but researchers do not have adequate information about the nature of the incentives to design statistically powerful enough tests to detect it. It could be, however, that, despite the intuitive appeal of a link between earnings-based compensation contracts and earnings management, certain forces mitigate the compensa- tion-related incentive. For example, boards of directors may be able to adjust bonuses to exclude “managed” earnings either informally or through formal contract provisions, such as making the bonus a function of only those earnings components that contain little discretion. If so, managing earnings provides no bonus-related benefit.

Also, the fact that early evidence showed a link between bonuses and earnings management whereas later evidence did not suggests that incentives have changed. The incentive to manage earnings to maximize bonuses may have decreased, but it may have been replaced by incentives to manage earnings to influence the stock price and maximize stock-based compensa- tion. The conjecture that managers have incentives to manage earnings related to stock-based compensation is predicated on the assumption that managers believe the managed earnings will be mispriced and the manager will be able to take advantage of the mispricing before it is corrected—either by selling shares or by exercising options. The evidence in Richardson, Tuna, and Wu; Beneish (1999); Cheng and Warfield (2003); and Gao and Shrieves (2002) suggests that earnings management increases with the extent to which managers have stock-based compensation.

■ Political incentives. Various studies have argued that governmental and nongovernmental organizations, although not necessarily intended users of financial statements, create incentives for companies to manage reported earnings because they can impose costs on companies that are “too profitable.” High profits can bring attention to a company that would otherwise like to stay off the regulatory radar screen. More importantly, high profits may be used as evidence that the company is gouging its customers. Agencies that serve to protect consumer interests may attempt to regulate prices charged by companies that show high profits. Note that the somewhat unusual feature of this political cost argument to explain earnings management is that it predicts that companies will manage earnings down.

Many of the studies have looked for earnings management during specific periods when regulatory scrutiny of an industry was high and high profits would invite harmful regulation. For example, evidence indicates that compa- nies in the cable television industry recorded income-decreasing discretionary

Earnings Quality

accruals during periods of U.S. Congressional hearings aimed at regulating cable prices (Key 1997), that chemical companies recorded income-decreas- ing discretionary accruals when Congress was debating Superfund site legis- lation (Cahan, Chavis, and Elmendorf 1997), and that companies investigated for antitrust merger violations engaged in earnings management through discretionary accruals to lower “excess” profits because regulators might interpret excess profits as an indication of an anticompetitive environment (Makar, Alam, and Pearson 1998).

Other related studies include that of Jones (1991), who showed that companies record income-decreasing discretionary accruals during import- relief investigations, and Hall and Stammerjohan (1997), who showed that companies manage earnings downward when faced with outstanding litiga- tion with potentially large damage awards. Liberty and Zimmerman (1986) and DeAngelo and DeAngelo (1991), who suggested that managers have incentives to manage earnings downward before labor union negotiations to strengthen their bargaining positions, found mixed results.

An important question related to political cost incentives for earnings management concerns the assumed sophistication of the contracting parties: Is it reasonable to assume that regulatory agencies have access only to publicly available information? Could they not obtain more detailed cost records to determine appropriate regulatory pricing intervention or to assess “excess” profits available for unionized employees or plaintiffs? Assuming that managers believe they can fool a diffuse group of investors with little access to important internal financial information may be plausible. And assuming that Congress, which has access to detailed internal records, will not “undo” earnings manipulation when making decisions may also be plausible (because lower profits provide Congress with a way to justify their political decisions to constituents who are not so well informed). But assum- ing that managers believe they can fool import-relief investigators, labor union negotiators, and plaintiffs’ attorneys is not so plausible. These parties have access to private information and the incentive to carefully undo earn- ings manipulation. Thus, in these situations, one might question why a manager would believe that any benefit is to be gained by engaging in earnings management. Nonetheless, the evidence suggests that managers do manage earnings for “political” reasons.

■ Tax incentives. Studies of incentives for earnings management created by tax codes generally examine changes in accruals or activities around the time of a significant change in tax regulations. Although this approach is a statistically powerful way to test for tax incentives for earnings management, it means that the research findings may not be generalizable to periods that

Earnings Management

do not include any significant changes in tax regulations. Tax incentives for earnings management may dominate other incentives in periods of significant tax law changes, but they may be less important at other times.

Research studies have documented several significant tax-related events that induced earnings management. Scholes, Wilson, and Wolfson (1992) showed that companies timed revenue and expense recognition to defer income in anticipation of lower tax rates in the mid-1980s. Maydew (1997) found that companies shifted recognition of revenues and expenses to take advantage of significant tax code changes in the Tax Reform Act of 1986—in particular, changes related to the treatment of provisions for net operating losses. He also found that the degree of earnings management was associated with the magnitude of the potential benefits. Boynton, Dobbins, and Plesko (1992) found, by looking at earnings patterns around the initiation of the alternative minimum tax, that earnings management is associated with a company’s exposure to the AMT. A somewhat contradictory result is the finding by Scholes et al. (1992) that earnings management exists for large companies but not for small companies. They concluded that the larger companies are more efficient tax planners. Boynton et al., however, found that earnings management is more pronounced for small companies (where size was measured by total assets).

The studies cited in this last section all indicate that companies manage the timing of earnings recognition to decrease current-period taxes. But in many cases, earnings manipulation will not affect a company’s tax obligation. The company must manage real transactions—adjust its investment, financing, and operating decisions—to manage taxable income. Unlike simply managing the timing of earnings recognition associated with existing transactions, this form of earnings management has real cash flow implications, and the benefits of the earnings management must exceed the costs. Nonetheless, evidence indicates that such manipulations occur. For example, Scholes et al. (1990) showed that companies adjusted their security holdings in response to tax code changes specifically related to banks during the mid-1980s.

<< /ASCII85EncodePages false /AllowTransparency false /AutoPositionEPSFiles true /AutoRotatePages /All /Binding /Left /CalGrayProfile (Dot Gain 20%) /CalRGBProfile (sRGB IEC61966-2.1) /CalCMYKProfile (U.S. Web Coated \050SWOP\051 v2) /sRGBProfile (sRGB IEC61966-2.1) /CannotEmbedFontPolicy /Warning /CompatibilityLevel 1.4 /CompressObjects /Tags /CompressPages true /ConvertImagesToIndexed true /PassThroughJPEGImages true /CreateJDFFile false /CreateJobTicket false /DefaultRenderingIntent /Default /DetectBlends true /ColorConversionStrategy /LeaveColorUnchanged /DoThumbnails false /EmbedAllFonts true /EmbedJobOptions true /DSCReportingLevel 0 /EmitDSCWarnings false /EndPage -1 /ImageMemory 1048576 /LockDistillerParams false /MaxSubsetPct 100 /Optimize true /OPM 1 /ParseDSCComments true /ParseDSCCommentsForDocInfo true /PreserveCopyPage true /PreserveEPSInfo true /PreserveHalftoneInfo false /PreserveOPIComments false /PreserveOverprintSettings true /StartPage 1 /SubsetFonts true /TransferFunctionInfo /Apply /UCRandBGInfo /Preserve /UsePrologue false /ColorSettingsFile () /AlwaysEmbed [ true ] /NeverEmbed [ true ] /AntiAliasColorImages false /DownsampleColorImages true /ColorImageDownsampleType /Bicubic /ColorImageResolution 300 /ColorImageDepth -1 /ColorImageDownsampleThreshold 1.50000 /EncodeColorImages true /ColorImageFilter /DCTEncode /AutoFilterColorImages true /ColorImageAutoFilterStrategy /JPEG /ColorACSImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /ColorImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /JPEG2000ColorACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000ColorImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasGrayImages false /DownsampleGrayImages true /GrayImageDownsampleType /Bicubic /GrayImageResolution 300 /GrayImageDepth -1 /GrayImageDownsampleThreshold 1.50000 /EncodeGrayImages true /GrayImageFilter /DCTEncode /AutoFilterGrayImages true /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /GrayImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /JPEG2000GrayACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000GrayImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasMonoImages false /DownsampleMonoImages true /MonoImageDownsampleType /Bicubic /MonoImageResolution 1200 /MonoImageDepth -1 /MonoImageDownsampleThreshold 1.50000 /EncodeMonoImages true /MonoImageFilter /CCITTFaxEncode /MonoImageDict << /K -1 >> /AllowPSXObjects false /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError true /PDFXTrimBoxToMediaBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXSetBleedBoxToMediaBox true /PDFXBleedBoxToTrimBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXOutputIntentProfile () /PDFXOutputCondition () /PDFXRegistryName (http://www.color.org) /PDFXTrapped /Unknown /Description << /FRA <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> /ENU (Use these settings to create PDF documents with higher image resolution for improved printing quality. The PDF documents can be opened with Acrobat and Reader 5.0 and later.) /JPN <FEFF3053306e8a2d5b9a306f30019ad889e350cf5ea6753b50cf3092542b308000200050004400460020658766f830924f5c62103059308b3068304d306b4f7f75283057307e30593002537052376642306e753b8cea3092670059279650306b4fdd306430533068304c3067304d307e305930023053306e8a2d5b9a30674f5c62103057305f00200050004400460020658766f8306f0020004100630072006f0062006100740020304a30883073002000520065006100640065007200200035002e003000204ee5964d30678868793a3067304d307e30593002> /DEU <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> /PTB <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> /DAN <FEFF004200720075006700200064006900730073006500200069006e0064007300740069006c006c0069006e006700650072002000740069006c0020006100740020006f0070007200650074007400650020005000440046002d0064006f006b0075006d0065006e0074006500720020006d006500640020006800f8006a006500720065002000620069006c006c00650064006f0070006c00f80073006e0069006e006700200066006f00720020006100740020006600e50020006200650064007200650020007500640073006b00720069006600740073006b00760061006c0069007400650074002e0020005000440046002d0064006f006b0075006d0065006e0074006500720020006b0061006e002000e50062006e006500730020006d006500640020004100630072006f0062006100740020006f0067002000520065006100640065007200200035002e00300020006f00670020006e0079006500720065002e> /NLD <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> /ESP <FEFF0055007300650020006500730074006100730020006f007000630069006f006e006500730020007000610072006100200063007200650061007200200064006f00630075006d0065006e0074006f0073002000500044004600200063006f006e0020006d00610079006f00720020007200650073006f006c00750063006900f3006e00200064006500200069006d006100670065006e00200070006100720061002000610075006d0065006e0074006100720020006c0061002000630061006c006900640061006400200061006c00200069006d007000720069006d00690072002e0020004c006f007300200064006f00630075006d0065006e0074006f00730020005000440046002000730065002000700075006500640065006e00200061006200720069007200200063006f006e0020004100630072006f00620061007400200079002000520065006100640065007200200035002e003000200079002000760065007200730069006f006e0065007300200070006f00730074006500720069006f007200650073002e> /SUO <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> /ITA <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> /NOR <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> /SVE <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> >> >> setdistillerparams << /HWResolution [2400 2400] /PageSize [612.000 792.000] >> setpagedevice