Discussion of Earnings Based Bonus Plans and Earnings Management by Business Unit Managers by cpd60066

VIEWS: 53 PAGES: 56

More Info
									Earnings Management Risk Exposure and the
           Pricing of Earnings




              Emeka T. Nwaeze




           Working Paper Series WCRFS: 06-28
                                                                                                1


           Earnings Management Risk Exposure and the Pricing of Earnings



                                    Emeka T. Nwaeze
                                   School of Business
                               Rutgers University-Camden
                      (856) 225-6651  nwaeze@camden.rutgers.edu


                                          Abstract

In this study, I examine the effect of earnings management incentives on the pricing of
earnings. Drawing from normative arguments and anecdotal evidence about the earnings
information problems associated with agency conflicts, I predict and test that exposure to
earnings management incentives negatively affects the valuation weight of earnings. Using
a stylized exposure score in a returns-earnings model, I find that the weight of earnings is
reliably decreasing in a firm’s exposure to earnings management incentives. Additional
analyses show that the weights of both cash-flow (operating) and accrual components of
earnings are decreasing in the exposure measure. These results are not sensitive to the
inclusion of traditional earnings quality measures, and are robust across alternative price-
and returns-earnings specifications. In an expanded analysis, I test whether institutional
stock holding and number of analysts’ following—as proxies for external monitoring—
mediate the negative effect of the exposure measure on the weight of earnings. The results
of this latter analysis show that none of the proxies moderates the negative effect of the
exposure measure on the weights of earnings or earnings components; rather, the negative
effect of the incentive exposure on the weights of earnings, cash flows, and accruals appears
to intensify as institutional stock holding or number of analysts’ following increases.

Key words: earnings management incentives, earnings information risk, security prices

Data availability: The data are available from the sources identified in the text



                                        January 2006


I owe special thanks to Ray King and Jeff Boone for providing detailed comments and
suggestions. I also benefited from comments by Angela Gore, James Groff, Dave Guenther,
Zite Hutton, Julie Lockhart, Lisa Kutcher, Steve Matsunaga, Dale Morse, Austin Reitenga,
Ted Skekel, Michael Williams, Jennifer Yin, and workshop participants at Western
Washington University, University of Oregon, and University of Texas-San Antonio. I
thank the Whitcomb Center for Research in Financial Services for providing research
support through use of the WRDS system. Funding is also provided by the Rutgers
University School of Business research grant.
                                                                                                 2


           Earnings Management Risk Exposure and the Pricing of Earnings

1. Introduction

       Recent financial reporting failures have revived interest in the connection between

managerial reporting incentives and earnings information quality. In several public

speeches, the former SEC chairman, Arthur Levitt, discusses the erosion in the quality of

financial reporting due principally to the desire by managers to meet profit benchmarks

through aggressive accounting (Levitt 1998, 2000). His remarks tacitly suggest a connection

between factors that expose firms to earnings management risks and investors’ skepticism

about the quality of earnings reports (see, also, Apostolou and Crumbley 1990; Hunt 2000).

The business press similarly predicts increased investor skepticism about the quality of

accounting reports, especially among firms with elevated exposures to earnings

management risks (Stuart 2002; Browning and Weil 2002; Berenson 2003).

       Besides such anecdotes, several a priori arguments suggest a connection between

managers’ reporting incentives and earnings information problem (i.e., uncertainty about

the information in earnings). Ball, Robin, and Wu (2003), for example, argue and provide

evidence that the incentives that drive the applications of accounting standards exert greater

influence on earnings quality than the standards. Other studies provide an agency view of

the accounting quality effects of reporting incentives. Notably, Watts and Zimmerman

(1978) contend that managers will use their accounting discretion to make reporting choices

that relax restrictions on their behavior imposed by accounting-based contracts; the earnings

report that emerges in such contexts tend to serve private incentives but may not necessarily

convey the economic substance of managerial actions (see, also, Holthausen and Leftwich

1983). Earnings management (EM) studies similarly predict increased likelihood that
                                                                                                3


neutral operation of the financial reporting process is compromised in strong incentive

contexts. A corollary described in several models is that the discretion managers have over

decisions that affect reported profits exacerbates the earnings information problem, more so

in contexts that earnings are a device for performance evaluation and wealth expropriation

(Watts and Zimmerman 1990). The information problem occurs mainly because, as Hirst

and Hopkins (1998) note, outsiders do not observe insiders’ opportunistic actions.

       Of particular interest in this study is the impact of the exposure to EM incentives on

the pricing of earnings. Drawing from several anecdotes and normative arguments linking

EM incentives to investor skepticism about earnings information, I propose and test that the

valuation weights of earnings and earnings components vary systematically with exposure

to EM incentives. Specifically, the amount of earnings and earnings components that map

into stock prices are negatively associated with exposure to EM incentives. The premise is

that EM incentives create uncertainties or diffuse beliefs among investors regarding the

economic information in earnings which, in turn, lead to lower valuation weight of earnings.

As I argue later, the lower valuation weight reflects the perceived noise in accounting

system conditional on extant EM incentives cum investors’ reluctance to trade on the basis

of the earnings signal. Moreover, to the extent that exposure to EM incentives is related to

cross-sectional variation in the priced attributes of earnings, the specification of EM

incentives in returns-earnings models will improve explanatory power.

       In designing the empirical analysis, I consider criticisms of extant studies that

examine the effect of single incentives on accounting quality/choice. In particular, Watts

and Zimmerman (1990) and Fields, Lys, and Vincent (2001) observe that managers

typically face multiple incentives that interact in, perhaps, a non-additive fashion to
                                                                                                                       4


determine the intensity of EM pressures and/or accounting choices. As a result, in many

settings, insights into the behavioral and economic consequences of an incentive, for

instance, are confounded because the result ascribed to the particular incentive is often

consistent with several competing explanations or may be driven by factors not identified in

the analysis. Fields, et al. (2001, page 291) observe that,

    By focusing on one goal at a time, much of the literature misses the more interesting question of
    the interactions between and trade-offs among goals. Moreover, it is not clear whether the
    conclusions are attributable to the specific motivation being analyzed; generally, results
    consistent with one hypothesis are consistent with many.

They also discuss the problems with the control-variables approach to modeling the effect

of other mediators, arguing that the conventional proxies used in analyses are typically

coarse, correlated, and non-additive in how they affect reporting incentives and choices.

         To mitigate several limitations of the single-incentive analysis, I construct a simple

measure of EM risk exposure based on numerous contexts predicted to motivate EM, and

analyze the effect of the stylized incentive exposure on the pricing of earnings. The measure

is a factor score, hereafter i-score, extracted from a variety of incentive indicators. The

specific indicators are surrogates of contexts and factors that notably predispose firms to

manage earnings, including restrictive debt contract (Duke and Hunt 1990; Press and

Weintrop 1990), incentive compensation (Healy 1985), political/regulatory exposure (Watts

and Zimmerman 1978; Key 1997), managers’ job security (Fudenberg and Tirole 1995),

deteriorating financial performance (Rosner 2003), analysts’ stock report (Abarbanell and

Lehavy 2003), and risk (Lilien and Pastena 1982). A key premise is that the interactions

among the indicators characterize a common incentive to manage earnings.1 Moreover, the

factor model is expected to capture the communal EM incentive that describes a firm’s

1
 A corollary noted in the literature and press is that the reporting actions of firms are the result of interactions
among multiple motivations (Hunt, Moyer, and Shevlin 1996; Mulford and Comiskey 2002).
                                                                                                                  5


overall exposure to EM incentives. The advantage is that the construct permits an analysis

of the pricing effects of the underlying incentives, while mitigating the noted omitted-

variables and inference problems in a single incentive-variable analysis.2

         To validate i-score as a proxy for earnings information risk, I examine the

distribution of several indicators of earnings information uncertainty--dispersion in

analysts’ forecasts, price-earnings ratio, interest-to-debt ratio, and poor accrual quality--

across quintiles ranked by i-score. Consistent with the key argument, i-score closely tracks

dispersion in analysts’ forecasts, declines in price-earnings ratio, increases in interest-to-

debt ratio, and deterioration in accrual quality. In a further validation check, I compare i-

scores between firms that restated their income for improper revenue/expense recognition

from 1997 to 2002, and a control group that had no restatements during the period. (Income

restatements arguably reflect prior reporting failures and earnings information problems).

This latter analysis shows that the restating firms have higher i-scores for the restatement

year and for the three prior years. In addition, results of simple discriminant analyses show

that i-score is reasonably successful in classifying firms that restated earnings: It correctly

classifies 77% of the firms that restated earnings. The success rate jumps to 84% and 87%

when the restating sample is restricted to firms whose cumulative dollar amount of

restatements equals or exceeds 1% and 5%, respectively, of the market value of equity.


2
 The problem with the single incentive-variable analysis is that specifying only one incentive in a model (e.g., in
Warfield, Wild, and Wild 1995) ignores interactions among other factors that may intensify/modify the overall EM
incentive or outsiders’ beliefs about earnings quality. For example, an EM incentive that arises from the pressure
to relax debt covenant restrictions may be reinforced by an incentive plan that ties managers’ rewards to earnings
performance; a ―buy‖ stock recommendation may similarly intensify the EM incentive connected with stock
option grants. (The high correlations among the variables used by Warfield et al., 1995, 88, Panel B support this
view). The factor analysis used here exploits the correlations among multiple indicators to construct a composite
index of EM exposure. It is noteworthy that the factor model is particularly suited to modeling the joint effects of
coarse proxies. A limitation of the factor model is that the marginal effect of each variable is subsumed in the
common factor. Note, however, that single incentive-variable analyses do not necessarily overcome this limitation.
                                                                                                             6


        For the primary test, I examine the effect of i-score on the weight of earnings in a

return-earnings regression. As predicted, the impact of i-score on the weight of earnings is

reliably negative. This result supports the primary thesis that elevated exposure to EM

incentives reduces earnings information quality. Moreover, the main effect of i-score on

returns is negative, which indicates that the market effects of i-score extend well beyond

reducing the valuation weight of earnings. I also test for an improvement in R2 attributable

to the inclusion of i-score in the model. This test is motivated by the expectation that i-score

is related to cross-sectional variation in the information quality of earnings and will thus

improve empirical specification. The results are consistent with this expectation: The

inclusion of i-score in the model produces a remarkable improvement in R2.

        To test whether i-score has identical effect on the weights of cash flow from

operations (CFO) and total accruals (TAC) components of earnings, I interact each

component with i-score and re-estimate the model. The results of this analysis show that the

weight of each component is reliably decreasing in i-score. The specification also achieves

a significant improvement in R2 over the more restricted model that conditions the weight of

aggregate earnings on i-score. This finding is noteworthy: It implies that the weight of TAC

as well as that of CFO is decreasing in the exposure to EM incentives, and suggests further

that earnings uncertainty associated with managers’ reporting incentives is not confined to

the accrual portion of earnings.3 The increase in R2, however, implies a marked variation in

the effect of i-score on the weights of CFO and TAC. These results are unaltered after I

control for alternative indicators of earnings quality, including accrual quality derived from


3
 This finding suggests an investor concept of earnings quality that encompasses the information in TAC and CFO,
and questions the view that the capital market mechanism emphasizes the role of CFO as an alternative indicator
of value when investors’ doubt about overall earnings quality increases.
                                                                                                7


the modified Dechow and Dichev (2002) model and discretionary accrual size derived from

the modified Jones model (see, Dechow, Sloan, and Sweeny 1995).

       On the premise that accrual components are differently vulnerable to the information

uncertainty induced by exposure to the EM incentives, I extend the analysis to consider the

effect of i-score across accrual components--accounts receivable accrual, inventory accrual,

accounts payable accrual, deferred tax accrual, net other working capital accrual, and

depreciation. The results show that i-score has a negative effect on the weights of CFO and

all accruals components, except accounts-payable accrual, on which the effect is positive,

albeit insignificant. Accelerated payments to suppliers thus appear to be least vulnerable to

the effects of EM incentives. The negative effect on the weight of deferred tax accrual is

consistent with the argument in the literature that investors interpret increases in deferred

tax as a signal of opportunistic reporting (Phillips, Morton, and Rego 2003; Hanlon 2005).

The R2 from this latter model is significantly higher than the R2 from the model in which

only CFO and TAC are conditioned by i-score. This implies that the pricing effect of EM

incentives is also assessed differently for different accrual components. These results are

robust across various specifications of the price- and returns-earnings relations.

       I also consider the effect of external monitoring on the relation between i-score and

earnings weight. Specifically, I test whether institutional stock ownership or analysts’

following mediates the impact of i-score on the weight of earnings. This aspect of the

analysis relates to the premise that external monitoring alleviates the information problems

induced by exposure to EM incentives (Muller and Riedl 2002; Bushee 1998) and thus

portends a moderate effect of i-score on the valuation weight of earnings. Strikingly, the

negative effect of i-score on the weights of earnings and earnings components intensifies as
                                                                                                       8


the percentage of institutional stock holding increases and as the number of analysts’

following increases; neither proxy of external monitoring appears to alleviate the perceived

uncertainty of earnings measured by i-score. The result is, however, consistent with the

explanation that institutional stock holding and analysts’ following are proxies for investor

sophistication and market’s ability to more fully/promptly recognize and price elevated

exposure to EM risk (see, e.g., Balsam, Bartov, and Marquardt 2002).

        Altogether, this study contributes to the growing literature on the economic effects

of earnings uncertainty. In particular, the exposure construct represents an additional

dimension of earnings information risk that complements various measures of earnings

quality identified in the literature (e.g., persistence, predictability, variability). The construct

offers several advantages: It captures the interactions among a variety of observable factors

that notably motivate earnings management. In this context, the incentive model avoids the

omitted-variables problem that often characterizes analyses of the incentive effects of a

single factor. Moreover, the incentive score is measured for each firm-year and, in that

sense, facilitates empirical analysis of the contemporaneous connection between a firm’s

vulnerability to earnings management and investors’ perception of earnings quality.

        The rest of the study proceeds as follows: Section 2 discusses the effect of exposure

to EM incentives on the valuation role of earnings. Section 3 discusses the EM incentive

contexts and indicators. Section 4 describes the data and EM incentive measures. Section 5

presents the main results. Section 6 summarizes and concludes the study.


2. EM incentives, earnings uncertainty, and the pricing of earnings

      The arguments relating incentive exposure to investors’ beliefs about, and pricing of

earnings uncertainty are both anecdotal and normative. In particular, there are burgeoning
                                                                                                               9


press editorials of increased investor mistrust of accounting reports, especially for firms

perceived to be vulnerable to accounting distortions (Stuart 2002; Browning and Weil 2002;

Berenson 2003). In several public speeches, Levitt (1998, 2000) and Hunt (2000) draw a

connection between factors that affect managers’ reporting motives (e.g., incentive pay

plans and capital market’s expectations) and earnings quality, and tacitly suggest that

corporate exposure to such factors erode investor confidence in financial reports. The doubt

about the information in earnings reflects a common belief that managers exploit their

accounting and enterprise discretions in ways that distort the mapping between reported

earnings and price-relevant attributes of a firm.4 Particularly relevant for investor judgment

is the fact that managers have legal/economic impetus to conceal reporting/enterprise

actions motivated by opportunistic incentives (Palmrose and Scholz 2004). Moreover, such

EM actions typically involve complex accounting maneuvers that create difficult analysis

problems for outsiders (Lev 2003b). This view is similar to the financial reporting problem

described by Hirst and Hopkins (1998) in which the variety of ways firms can manage

earnings and the non-trivial effort required to distinguish EM activities from normal

enterprise activities create difficult expectation and valuation problems for analysts.

       The impact of the perceived uncertainty on the valuation weight of earnings derives

from several sources. Holthausen and Verrecchia (1988), for example, derive an equilibrium

pricing model in which the price response per unit of an information signal is decreasing in

the level of noise associated with the signal. A pertinent argument for this result is that an


4
  In a report for the USA Today, Ahren (2002) observes that investors’ mistrust of corporate financial results
lingers, citing the USA Today/CNN/Gallup Poll of 650 respondents conducted from July 5 to July 8 of 2002 which
shows that 77% of those surveyed believe top executives take self-serving [accounting] actions at the expense of
the corporation. The earnings management literature also predicts increased probability that neutral operation of
the financial reporting process is compromised in strong incentive contexts (see, Healy and Wahlen 1999, for the
related literature).
                                                                                                  10


increase in the [perceived] noise in a signal such as earnings motivates investors to seek out,

and rely on alternative information sources, which, in turn, reduces the weight they apply to

the signal. Relying on this argument, Imhoff and Lobo (1992) predict and find that ex-ante

uncertainty about earnings, defined as the dispersion in analysts’ forecasts, leads to a lower

earnings response coefficient. Models of the pricing effects of uncertainty corroborate the

tenor of this result. Miller (1977), for example, argues that a potential consequence of

divergence of opinion among risk-averse investors is increased reluctance to hold, and

pressure to sell the security. The logic, noted in several extant models, is that uncertainty

about a signal causes unwillingness to trade on the basis of the signal as investors price-

protect themselves against potential losses from trading with more informed investors

(Diamond and Verrecchia 1991; Bhattacharya and Spiegel 1991; Chen and Jiambalvo

2004). Easley and O’Hara (2004) similarly argue that investors, in general, are averse to

securities about which they know little, and would prefer to hold stocks for which there is

more public than private information. These results suggest that exposure to EM incentives

will moderate the valuation role of earnings to the extent that it creates diffuse beliefs

among investors regarding the economic information in earnings.

      The incentive-driven uncertainty is further expected to affect the valuation weight of

earnings via its effect on the conditional expectation of cash flows. In particular, for firms

perceived to be highly vulnerable to accounting distortions, investors are likely to associate

reported earnings with greater noise and lower certainty-equivalent cash flows. The logic is

that investors do not or cannot observe extant opportunistic reporting actions (their scope

and multi-period effect) and must rely on a subjective evaluation of the accounting system

to form expectations about cash flows. For high incentive firms, often characterized by
                                                                                                                    11


more opaque financial reporting (Bhattacharya, Daouk, and Welker 2003), such evaluations

become even more subjective and susceptible to high estimation errors, which investors

anticipate and use in deriving [lower] uncertainty-adjusted cash flows.5 This view is similar

to the general filtering model described by Duffie and Lando (2001) in which investors

receive a noisy accounting signal and apply a filtering process to estimate the unobservable

system. Hirst and Hopkins (1998) point out that the menu of EM possibilities and difficulty

in distinguishing EM from normal business activities make it harder for investors to rely on

financial statements wholly for valuation judgments.


3. EM incentive contexts and indicators

         In this study, I focus on several factors that have popular support in the literature as

indicators of contexts that create strong incentives for EM. The variables used in analysis

are indicators of restrictive debt covenant, incentive contract, political exposure, threat of

executive dismissal, deteriorating financial performance, business risk, analysts’ stock

recommendation, regulatory exposure, and business growth.6



5
  For instance, assume investors assess a firm’s permanent earnings, X*, based on the earnings signal, X. Let the
assessment follow a simple adjustment process: X* = X – Xm (Xm = assessed earnings management). Outsiders do
not observe earnings management directly but make a probability assessment that earnings are managed based on
extant EM incentives. Let the assessed probability that X are managed be Pm and the assessed ratio of managed
earnings be m. m is unbounded and takes a positive or negative value depending on the direction of X and X m. If
X > 0 and Xm < 0 (Xm > 0) [earnings are positive and a portion is believed to have been deferred (embellished)],
then < 0 ( > 0). On the other hand, if X< 0 and X m < 0 (Xm >0), then  > 0 ( < 0). Using this framework, the
assessed permanent earnings, X* = Pm(X–Xm) + (1–Pm)X ≡ Pm(X –mX) + (1–Pm)X. This yields X* = X–PmmX.
The PmmX is the size of adjustment for earnings uncertainty connected with EM incentives. The pricing of X via
a returns model may be presented as: R = a + b(X–X ) + error ≡ a + b(1–)X + error, where  = (Pmm). In this
formulation, the slope effect of the uncertainty associated with extant EM incentives is measured by (1–).
6
  The initial variables include indicators of investment opportunity set, initial public offering, auditor size, and
insider ownership; based on initial data reduction results, these indicators have a negligible effect on i-score in the
presence of other indicators, and are dropped from analysis. I also exclude factors that are mere indicators of past
EM actions (e.g., earnings restatements, SEC enforcement actions) and de-emphasize weak/ambiguous proxies of
managers’ EM incentives or contemporaneous EM pressures (e.g., earnings persistence, variability, predictability).
                                                                                                     12


3.1. Debt contracts and EM exposure

        Debt covenants place a variety of restrictions on enterprise activities, including

restrictions on dividend payment, working capital balances, coverage ratios, and new debt.

At higher levels of debt, firms face tighter covenants and more intense pressures to avoid a

violation (Kalay 1979; Duke and Hunt 1990; Press and Weintrop 1990). Such pressures

create strong incentives for EM actions intended to relax the restrictions (Dhaliwal 1980;

Holthausen 1981; Leftwich 1981; Sweeny 1994; DeFond and Jiambalvo 1994).7 Healy and

Palepu (1990) further show that firms close to covenant violation often omit dividends.

These studies provide compelling evidence that high leverage and dividend omission are

common features of restrictive covenants (Kalay 1979; Holthausen 1981; Bowen, Noreen,

and Lacey 1981; Fazzari and Hubbard 1988; Healy and Palepu 1990) and suggest an

elevated exposure to EM risk.

3.2. Incentive compensation plans and EM exposure

        Incentive compensation plans have grown in popularity as firms seek ways to align

managers’ interest with shareholders’. The behavioral effects of such plans have received

attention. Healy (1985), for example, argues that bonuses tied to earnings performance

create an incentive for managers to take actions that maximize the bonus payout. Similarly,

there is a preponderance of evidence that ties market-based compensation plans, such as

stock option grants, to EM incentives (Healy and Wahlen 1999; Bartov and Monhanran

2004). This latter incentive arises because the gains managers obtain from exercising their

stock options are tied to earnings performance via the effects of the latter on stock prices

(Lev 2003a). Bonus and stock options plans are thus expected to increase EM risk exposure.

7
 Other studies that associate higher levels of debt with the propensity to manage earnings include
Zmijewski and Hagerman (1981), Daley and Vigeland (1983), and Ayres (1986).
                                                                                                  13


3.3. Threat of executive dismissal and EM exposure

       Threats to managers’ jobs are frequently associated with deteriorating financial

conditions and/or failure to meet certain performance thresholds (Coughlan and Schmidt

1985; Warner, Watts, and Wruck 1988; Weisbach 1988; Farrell and Whidbee 2003). A

number of studies have considered the earnings management incentive created by such

threats. To ease exposition, I discuss the incentives for EM created by threats of dismissal

and job security under the incentive effects of declining financial performance.


3.4. Deteriorating financial performance and EM exposure

       Deteriorating financial performance describes repeated episodes of poor financial

results and increased risk of business failure. A number of studies discuss the EM incentives

that arise when firms experience declining financial performance (e.g., Feroz, Park, and

Pastena 1991; Lev 2003a; Rosner 2003). To capture the incentive effects of a variety of

conditions that reflect poor financial performance and/or increased risk of business failure, I

focus on three indicators of poor financial condition, including (1) financial distress, (2)

losses and small gains, and (3) asset sales.

       Financial distress: There is considerable research on the incentive problems

associated with financial distress. Jensen (1986) argues that managers of poor performers

often engage in value diminishing activities intended to mitigate the cost of financial

distress. Poor performers have a higher probability of failure and an overriding incentive to

improve profits in the near term to avert a takeover and bolster their ability to bargain with

creditors (Koch 1989). Rosner (2003) argues that distressed firms manage earnings to

conceal their poor conditions in hopes of bouncing back soon (see, also, Lev 2003a).8
                                                                                                                 14


Managers of ailing firms face an elevated threat of job loss and are under pressure from

within to improve profits in the near term. The threat of job loss creates an incentive for

such managers to manage profits to avert dismissal (DeFond and Park 1997). These views

suggest a link between financial distress and EM exposure.

         Recent loss: Losses create certain economic and incentive problems. Miller and

Modigliani (1966) note that losses complicate the use of earnings-based models of value

since they have limited ability to signal earnings power. In lieu of earnings-based models, a

number of studies use real options theory to explain investors’ approach to valuing a loss

firm. Hayn (1995), Burgstahler and Dichev (1997), and Nwaeze (2005) explain that

investors assign a value to a loss firm on the expectation the loss is transitory; else they will

exercise their option to liquidate the firm.9 Such a market mechanism implies that a firm

with a loss episode faces an inherent expectation it will return to profitability in the near

term. The ability to show profits in the near term is important since further loss episodes are

likely to erode investors’ confidence in the firm’s capacity to ―bounce back,‖ and raise the

prospect that investors will exercise their liquidation option. The pressure to ―bounce back‖

and restore investors’ confidence is likely to elevate the EM risk exposure.

         Fudenberg and Tirole (1995) also analyze the incentive effects of poor financial

performance. They argue that, in bad times, concerns about job loss provide an incentive for

managers to embellish income. Managers are also concerned with loss of credibility and

8
  DeFond and Jiambalvo (1994) contend that firms with a going-concern qualification (an indicator of financial
distress) are likely to have negative accruals because of increased monitoring by auditors (see, also, DeAngelo,
DeAngelo, and Skinner 1994). This view conflicts with the predictions of the agency literature that see managers
of firms in distress acting to transfer future income into the current period to (1) show their firms in the most
favorable light, (2) save their jobs and (3) reduce agency costs that would accompany persistent bad performance
(see, also, DeFond and Jiambalvo 1993).
9
 Berger et al. (1996) note that when profits are consistently low or negative and bankruptcy is imminent, exit
values of assets are more important determinants of firm value (see, also, Joos and Plesko 2005).
                                                                                                                   15


potential wealth effects when they miss key thresholds or expectations (Merchant 1989).10

As a result, they have an incentive to embellish income immediately following a loss.

         Current small loss or gain: Extant evidence suggests that firms whose current pre-

managed earnings are negative resort to earnings management to conceal the severity of

their losses (Gaver et al. 1995; DeFond and Park 1997). A number of studies find patterns

of accruals among small profit firms that are consistent with earnings management to avoid

reporting a loss (Burgstahler and Dichev 1997; DeGeorge, Patel, and Zeckhauser 1999;

Burgstahler and Eames 2003). Recently, Dechow, Richardson, and Tuna (2003) show that

small-loss firms report abnormally high positive discretionary accruals, consistent with

earnings management to show small losses. They note that, ―since small loss firms have, on

average, positive discretionary accruals this suggests that they also face incentives to boost

earnings even though they reported a loss.‖11 These results raise the prospect that incidents

of small losses or gains are strong hints of deeper financial problems and EM incentives.

         Asset sales: The literature on asset sales suggests that firms sell assets largely in

response to poor performance (Alexander, Bensen, and Kampmeyer 1984; John, Lang, and

Netter 1992; Lang, Poulsen, and Stulz, 1995). Bartov (1993) and Herrmann, Inoue, and

Thomas (2003) find evidence that managers time such sales strategically to achieve results

that alleviates certain agency problems, including reducing the costs of contracting or re-



10
  Farrell and Whidbee (2003) and Puffer and Weintrop (1991) find an increased likelihood of CEO turnover when
realized earnings fall short of analysts expectations. (See, also, DeFond and Park 1999; Goyal and Park 2002).
11
  Despite the loss avoidance argument presented by Burgstahler and Dichev (1997), firms have compelling
incentives to report small rather than large losses. In particular, a small loss is easier to explain as a transitory
episode than a large loss, and may be used to avert a more severe market reaction associated with a large loss.
Moreover, firms that report small losses may be in a stronger position to negotiate favorable financing terms than
those that show large losses in that small losses may be perceived as having limited cash flow/economic
consequences, whereas large losses may be judged to have real cash-flow/economic consequences.
                                                                                                  16


contracting and the probability of dismissal. Such evidence raises the prospect that asset

sales are informative about extant pressures on managers to bolster near-term performance.

3.5. Political/regulatory exposure and EM exposure

       Watts and Zimmerman (1978, 1986) hypothesize that firms vulnerable to politically

motivated wealth transfers adopt accounting methods that reduce the probability of political

intervention. They contend that large firms are highly visible and more vulnerable to

politically motivated wealth transfers. The intense scrutiny such firms face over their

accounting choices reduces their incentive for opportunistic reporting (Chaney and Jeter

1992). By contrast, small firms are less subject to public scrutiny and/or political pressures.

Small firms often engage small external auditors that lack the resources to flag aggressive

reporting (DeAngelo 1981; Palmrose 1988). Thus, the opportunity and incentive to garble

earnings are greater for small firms. Explicit price or profit regulation has also been shown

to reduce the incentive for income-boosting earnings management (Key 1997). Rate of

return regulation, in particular, involves a regulated profit structure that penalizes firms

when they exceed a preset profit ceiling, and allows them to make up profit shortfalls. Such

a profit structure provides an incentive for firms to select accounting procedures that reduce

the risk of regulatory intervention and/or justify requests for favorable profit adjustments.

3.6. Analysts’ stock recommendation and EM exposure

       Extant analysis of the behavioral effects of analysts’ reports suggests a connection

between analysts’ stock recommendations and opportunistic financial reporting. Abarbanell

and Lehavy (2003), for example, examine the distribution of unexpected accruals following

analysts’ stock recommendations, and conclude that firms whose stocks are rated ―buy‖ are

predisposed to manage earnings upward; firms whose stocks are rated ―sell‖ are predisposed
                                                                                                              17


to defer current income in order to deliver positive earnings in the future. Such actions are

taken, perhaps, to elicit price behaviors that corroborate the stock recommendation. This

suggests a positive link between analysts’ stock recommendations and EM risk exposure.

3.7. Business risk and EM exposure

        Lilien and Pastena (1982) and Malmquist (1990) examine the determinants of

accounting policies in samples of oil and gas firms and conclude that the risk associated

with enterprise activities affect accounting choice. They find that high risk firms (i.e., firms

that operate in conditions in which payoffs on investments are risky) tend to use accounting

methods that reduce the profit effects of their risky activities. Christie (1990) reviews extant

studies on the contracting theories of accounting methods choice and concludes that risk,

among other factors, significantly explains firms’ choice of accounting procedures (see,

also, Hagerman and Zmijewski 1979; Lys 1984). These studies motivate the hypothesis that

high risk firms possess a greater predisposition to earnings management than low risk firms.

3.8. Growth expenditure and price-earnings multiples

        Growth investments enter the analysis as a variable that augments the weight of

earnings in returns-earnings regressions. The premise is that the profit effects of growth

investments lag the economic activity by several years, and are only partially reflect on

current earnings. In other words, reported earnings of expanding firms are an imperfect

indicator of the prospects implied by the efforts. On the other hand, price responses to such

activities reflect more fully the expected profit effects. As a result, the stock price changes

that accompany such investments are typically large relative to the current earnings effect.12


12
  Recognizing the non-responsiveness of current profits to contemporaneous investment activities, the contracting
theory emphasizes pay-performance contracts that use non-profit measures jointly with, or in lieu of earnings for
high growth firms, noting that earnings are an imperfect measure of managers’ effort if a firm is investing and
                                                                                                                18


Such a phenomenon is akin to the effect of accounting conservatism on price-earnings

relation noted by Beaver and Ryan (1993), and portends a higher coefficient on earnings in

returns-earnings regressions.


4. Data selection and EM incentive measure

4.1. Data selection

        Data for total earnings and earnings components are obtained from the Compustat

from 1992 to 2002.13 The data items and their respective Compustat item numbers include:

EARN = Income before extraordinary items and discontinued operations (#18)
AR = Change in accounts receivables (#302)
INV = Change in inventories (#303)
AP = Change in accounts payables (#304)
ATX = Change in accrued income taxes (#305)
OT = Net change in other current assets and current liabilities (#307)
CFO = Net cash flow from operating activities (#308)
DPR = Depreciation plus amortization charge (#14)
TAC = EARN – CFO (#18 – #308)

        Stock returns from CRSP are compounded monthly for a 12-month period ending

three months after the fiscal year end. Indicators of EM incentives from which i-score is

derived are leverage and dividend omission (debt constraints), bonus and stock option

values (incentive contracts), Altman’s Z-score, recent loss, current small gain/loss, and

asset sales (threat of dismissal and deteriorating financial performance), firm size (political

exposure), membership of the Drug or Electric & Gas Utility or Banking & Insurance

industry (regulatory exposure), analysts’ stock recommendation (equity-market incentive),

book-to-market ratio (business risk), and capital expenditure growth (business growth).

growing rapidly (Smith and Watts 1992; Gaver and Gaver 1993; Skinner 1993; Baber et al. 1996). Moreover,
growth investments reduce current earnings via higher depreciation, and thus increase the price-earnings multiple.
13
  The sample period begins in 1992, which is the first year detailed CEO compensation data are available on the
ExecuComp database.
                                                                                               19


       The data are obtained from a variety of sources identified as follows: Institutional

stock holding and analysts’ following data are from the Thompson Financial Securities Data

and IBES, respectively. Bonus and stock option data are from the ExecuComp. Dividend

omissions are from the merged CRSP-Compustat PDE file, corroborated/augmented by a

search of Lexis/Nexis, using key phrases, omit dividend, dividend omission, miss dividend.

The remaining indicators are collected or computed from data on the Compustat. A

complete definition of the incentive variables and description of the sampling procedure are

shown in Table 1. The final sample contains 1,137 firms and 8,628 firm-year observations.

                                     [Table 1 about here]

4.2. Measurement and validation of i-score

4.2.1. Factor analysis results

       A firm-year EM incentive measure is the factor score extracted from the incentive

variables defined in Table 1 (based on year-specific factor analysis). Table 2 shows simple

statistics for the variables and some results from the factor analysis, including the factor

loadings (i.e., the correlation of each factor with the common factor), communalities (i.e.,

the variation in each indicator common to other indicators), and scoring coefficients.

                                         [Table 2 about here]

       In Column 1 of Panel B, each indicator, except regulation and business growth, has

a positive factor loading. Book-to-market ratio has the highest loading of 0.745. Other

proxies that load highly on the common factor include firm size (loading = 0.68), stock

options (loading = 0.498) and a recent loss (loading = 0.44). Dividend omission, financial

distress, and leverage also have appreciable loadings on the common factor. Regulation and

capital expenditure growth load negatively on the common factor. Both variables appear to
                                                                                                   20


mitigate the positive influence of other variables on the underlying common factor,

consisted with their expected conflict with indicators of aggressive accounting.

4.2.2. i-score validity

        The empirical prediction relating i-score to the price informativeness of earnings

relies on the argument that exposure to factors that generate i-score is related to investors’

perceived uncertainty about earnings. The specific argument is that higher (lower) i-score

portends greater (weaker) investors’ perceived ex-ante uncertainty about the information in

earnings. Such uncertainty is often defined with reference to investors’ divergence of

opinions (Miller 1977). Extant literature uses the dispersion in analysts’ forecasts as an

operational proxy of such perceived uncertainty (e.g., Daley, Senkow, and Vigeland 1988;

Morse, Stephan, and Stice 1991; Imhoff and Lobo 1992; Barron and Stuerke 1998; Kinney,

Burgstahler, and Martin 2002). Thus, i-score is expected to track forecast dispersion, Disp,

to the extent that it reflects the information in analysts’ divergence of opinions. To validate

the prediction, I present in Panel A of Table 3, the mean and median Disp1 and Disp2

(defined as forecast dispersion scaled by absolute mean EPS forecast and by fiscal year-end

share price, respectively) across quintiles ranked by i-score. The mean and median

dispersions are monotonically increasing in the level of i-score. The differences in both the

mean and median dispersions between the top and bottom quintiles are highly significant (P

< 0.0001). These results portray a reasonable correspondence between i-score and

dispersion in analysts’ forecasts.

                                      [Table 3 about here]

        Additional validation tests focus on the correspondence between i-score and other

indicators of earnings information risk identified in the literature. Francis et al. (2005), for
                                                                                                    21


example, use price-earnings (P/E) ratio and ratio of interest expense to interest bearing debt

as indicators capital costs reflected on earnings information risk. In the context of this study,

both indicators are expected to be increasing in i-score to the extent they share a common

information (regarding earnings) with i-score. Moreover, recent studies have used the

standard deviation of residuals from working-capital accrual models as a source of earnings

information risk (e.g., Dechow and Dichev 2002; Francis et al. 2005). This latter measure,

which reflects departures of the realized accruals from the specified accrual process, will be

related to i-score to the extent that it has information that overlaps with what is in i-score.

       In Panel B of Table 3, I present the mean and median P/E (ratio of market equity to

earnings before extraordinary items), Debtcost1 (ratio of interest expense to total interest

bearing debt), Debtcost2 (ratio of the sum of interest expense and capitalized interest to

total interest bearing debt), and AQ (poor accrual quality derived from the modified Dechow

and Dichev model in Francis et al. 2005) across quintiles ranked by i-score. The mean and

median P/E decrease steeply and monotonically from the bottom to the top quintile. The

differences in mean and median ratios between the top and bottom quintiles are negative

and highly significant (P< 0.0001). The means and medians of both debt-cost proxies, and

the mean and median AQ increase monotonically in the level of i-score. The differences in

mean and median values of these latter indicators between top and bottom quintiles are

positive and highly significant (P< 0.01). These results are strongly confirmatory of the

positive correspondence between exposure to EM incentives, indexed by i-score, and

indicators of capital costs and earnings uncertainty.

       In view of the extensive use of earnings restatements to corroborate contexts in

which EM incentive was strong, I compare the distribution of i-score between firms that
                                                                                                                 22


restated their earnings for improper revenue/expense recognition from 1997 to 2002, and a

control group from the initial sample that had no restatements during the period.14 Earnings

restatements are used widely as an indicator of prior reporting failures and elevated EM

risks, and thus provides a reasonable benchmark (Srinivasan 2005; Kinney, Palmrose, and

Scholz 2004; Myers et al. 2003; Richardson et al. 2003; Palmrose and Scholz 2004). The

test period covers four years, starting three years prior to the restatement year (Year –3) and

ending in the restatement year (Year 0). The results are shown in Panel A of Table 4.

                                           [Table 4 about here]

         The two samples differ markedly in the mean and median i-scores across all four

years, with the restating firms having significantly higher mean and median i-scores each

year. This result is consistent with the prediction that restating firms have higher EM risk

exposure than the general population. Note also that the mean i-score is positive and highly

significant in year t–3 through year t for the restating firms. By contrast, the statistic is

insignificantly different from zero across all four years for the non-restating firms. Next, I

test the ability of i-score to correctly classify the restating firms from the combined sample

of restating and non-restating firms. In a discriminant analysis (Proc discrim in SAS), i-

score correctly classifies 77% of the restating firms and 53% of the non-restating firms. The

scores for the remaining 47% of non-restating firms are either tied or classified wrongly.15




14
  The restatement data are firms that announced restatements from 1997 to 2002 as reported by the General
Accounting Office (2002). The report lists 919 firms, from which I exclude firms that eventually did not restate,
restatements for technical reasons, multiple restatements by the same firms, foreign firms, firms for which I cannot
determine the size of restatement, and firms that do not have valid Compustat and/or CRSP data during the sample
period. The data requirement results in 423 usable restatement firms.
15
  Non-restating firms may be harder to classify correctly since non restatement does not necessarily imply an
aversion to earnings manipulation. It is quite likely that many firms that would have been subject to restatements
were successful in avoiding detection of reporting violations for which restatements would have been required.
                                                                                                                  23


         To improve efficiency of tests, I restrict the restating sample to firms for which the

size of restatement equals or exceeds one percent of the book value of equity, and further

select an equal but random sample of firms that have no records of restatements during the

sample period. Panel B compares the mean and median i-scores between the two groups.

The mean and median i-scores are positive and highly significant for the restating firms

across all four years (P < 0.0001). For the non-restating firms, both statistics are negative

from t–3 to t–1. The restating firms have by far larger positive mean and median i-scores.

The difference in mean or median i-score between the two groups is highly significant for

each test year (P < 0.0001). Next, I re-estimate the discriminant model for the combined

samples. The results (not tabulated) show that i-score correctly classifies 84% of the

restating firms and 61% of the non-restating firms.16 These results suggest that i-score is

reasonably informative about firms most likely to restate earnings.

         The size of earnings restatement is often used as a measure of the severity of past

financial reporting failures (Palmrose, Richarson, and Scholz 2004; Srivinasan 2005).

Arguably, the size of restatement is also indicative of the scale of reporting failures and

earnings distortion. Accordingly, I expect a positive relation between the size of restatement

and i-score. The results of regressing absolute size of restatement on i-score for each of the

four test years are presented in Panel C. The coefficient on i-score is positive and highly

significant (P < 0.01) for each test year. Notably, the R2 is largest for the year of restatement

and the year immediately preceding the restatement year.




16
  Additional results not reported in the Table show that the percentage of restating firms classified correctly
increases from 84% to 87% when I restrict the restating sample to firms for which the absolute size of earnings
restatement relative to the book value of equity equals or exceeds 5%.
                                                                                                                    24


5. Main results

5.1. Benchmark returns-earnings relation

             Initially, I regress returns on earnings and earnings components to obtain the

weights and explanatory power of the earnings variables, exclusive of the i-score effects.

The following three models are estimated (firm and time subscripts are omitted for brevity):

R = + EEARN + 
     B                            B                                                                          (1a)
                                                                                                             B




R =  + 1CFO + ATAC + 
                      B                   B                                                                  (2a)
                                                                                                             B




R =  + 1CFO +  2 AR +  3INV +  4 AP +  5DTX +  6OT +  7 DPR + 
                          B   B       B             B   B      B             B          B   B       B        (3a)
                                                                                                             B




All right-hand side variables (defined in section 4) are scaled by the beginning-period

market value of equity to align the metrics with the returns metric and reduce scale effects.

In (1a), E measures the unconditional weight of earnings; (2a) allows for differences in the
               B




weight of CFO and TAC; (3a) disaggregates TAC into the various working-capital and

long-term components. The results are reported in Table 5.

                                                    [Table 5 about here]
             The earnings-only model achieves a modest adjusted R-square (hereafter, R2) of

3.5%. The R2 increases to 5.8% when earnings are decomposed into CFO and TAC. The

weights of CFO and TAC are positive and significant (P<0.0001), consistent with prior

results. (3a) that includes CFO and components of TAC achieves even a higher and

significant increase in R2 over (2a).17 The increase in R 2 corroborates prior evidence that
                                                                     P




17
     The test of significant improvement in R2 between (2a) and (1a), for example, relies on a Z-statistic computed as
                       
 R2a  R1a =  R2a   R1a
   2     2       2       2
                                                       
                                       2 R2a   2 R12a , where E(•) and  2(•) are the expected value and variance
                                           2

         2
of R , respectively, based on the Cramer’s method (Cramer 1987). Z ~ N(0, 1) under the null hypothesis that the
increase in R2 is statistically insignificant. The advantage of the approach is that it can be applied to nested and
non-nested models, and is robust across different linear restrictions and sample sizes.
                                                                                                       25


disaggregating earnings into CFO and components of accruals provides a better explanation

of stock returns than aggregate earnings (e.g., Bath, Cram, and Nelson 2001).


5.2. The effect of i-score on return-earnings sensitivity

          Initially, I test the effect of each incentive indicator on the weight of earnings to

corroborate predictions about their individual (albeit correlated) effect on earnings quality.

For each indicator, I regress returns on earnings augmented with an interaction between the

indicator and earnings. For example, I assess the leverage affect via the model: R =  +

EEARN + E(EARN*LEV) + error. The model is estimated annually. The mean and median
          B   B




coefficients describing the effect of each indicator on the weight of earnings are reported in

Table 6. Regulation and business growth have positive loadings on the weight of earnings,

whereas the remaining indicators have negative loadings on the weight of earnings. These

results are consistent with the predicted effects of the indicators on earnings information.

                                                      [Table 6 about here]

          Next, I include an interaction between i-score and EARN in (1a) to assess the effect

of i-score on the weight of earnings. On the premise that CFO and accruals are differently

vulnerable to EM (Wayman 2003) and to perceived uncertainty due to EM incentives, I

augment (2a) and (3a) with interactions between i-score and CFO, and between i-score and

various accruals. In addition, i-score is specified as a separate variable in the models to test

for a systematic effect of the latent factor on returns. The models are given as:

R =  + i-score + EEARN + E(EARN*i-score) + 
 B    B                               B   B                                B                       (1b)
                                                                                                   B




R =  + i-score + 1CFO + ATAC + 1(CFO*i-score) + A(TAC*i-score) + 
 B    B                          BB                                                    B           (2b)
                                                                                                   B




R =  + i-score + 1CFO +  2 AR +  3INV +  4 AP +  5DTX +  6OT +  7 DPR
 B                           B                B   B      B         B   B       B   B       B   B




      + 1(CFO*i-score) + 2(AR*i-score) + 3(INV*i-score) + 4(AP*i-score)
      + 5(DTX*i-score) + 6(OT*i-score) + 7(DPR*i-score) +                                    (3b)
                                                                                                   B
                                                                                                   26



 measures the main effect of i-score on returns; E, 1, and A reflect the impact of i-score

on the weights of earnings, CFO, and TAC, respectively; 1, ..., 7 evaluate the impact of i-

score on the weights of the various earnings components. Table 7 presents the results.

                                      [Table 7 about here]

       In Column 1, E which assesses the impact of i-score on the weight of earnings is

negative (E = -0.1388) and highly significant (P < 0.0001), consistent with the prediction

that i-score is associated with lower quality of earnings. The R2 is 10.3% and about three

times the size of the R2 for (1a). The increase in R2 is highly significant (P < 0.0001), in

support of the expectation that i-score enhances model explanatory power. Moreover, the

effect of i-score on returns is reliably negative, which indicates that the market effects of i-

score extend beyond reducing the weight of earnings. A plausible intuition for these results

is that firms that have elevated exposure to EM incentives are also likely to be experiencing

declining performance and/or capital market problems.

       The results in Column 2 show that the coefficient on the interaction between CFO

and i-score and between TAC and i-score are negative and highly significant (P < 0.0001).

The negative effect on the weight CFO is striking: It indicates that exposure to EM

incentives, indexed by i-score, affects investors’ perception of CFO quality as well. That is,

the information uncertainty associated with management reporting incentives is not

confined to the accrual component of earnings. Rather, CFO and TAC both appear to suffer

significant declines in valuation weights as the incentive for earnings management

intensifies. The model achieves a significant improvement in R2 over (2a), and also over

(1b) that includes an interaction between i-score and aggregate earnings (P < 0.0001). This

latter result implies that i-score affects the weights of CFO and total accruals differently.
                                                                                                                  27


         Estimates of the impact of i-score on finer partitions of earnings--CFO and accrual

components--are shown in the third column. The results show a marked variation in the

effect of i-score across the components. The weight on the interaction of each current-asset

accruals with i-score is negative and significant (P < 0.01). The impact of i-score on the

weight of accounts-payable accrual is positive but insignificant. This latter result suggests

that exposure to EM incentives does not affect how investors interpret accelerated payment

to vendors. On the other hand, the effect of the exposure measure on the weight deferred tax

accrual is negative and highly significant. This result suggests that investors assess deferred

tax accrual to be vulnerable to managerial opportunism, consistent with the evidence

reported by Hanlon (2005) and Phillips, Pincus, and Rego (2003).18


5.3. Expanded tests

         Recent studies have relied on the behavior of accrual earnings for inferences

regarding earnings quality. Dechow and Dichev (2002) and Francis et al. (2005), for

example, test and validate the role of accrual variability measure (derived from the standard

deviation of the residuals from an accrual model) as an indicator of accrual quality (AQ).

The AQ model uses departures from the underlying relation between accruals and cash flow

process (augmented with revenues and fixed operating assets) to index earnings quality. To

examine whether and the extent the presence of AQ modifies the effect of i-score on the

weights of earnings and earnings components, I re-estimate (1b) and (2b), augmented with

an interaction of AQ (estimated as in Francis et al. 2005) with earnings in (1b), and with

interactions of AQ with CFO and with TAC in (2b). The models are given as:

18
  To reduce the bias in parameters caused by pooling data across time, I estimate each model annually (1993 to
2002) and then average the annual coefficients for each variable. The mean coefficients and their corresponding t-
values (not reported for brevity) lead to the same inferences as those based on cross-sectional time-series results.
                                                                                                     28


    R =  + i-score +AQ + EEARN + E(EARN*i-score) + E(EARN*AQ) +                     B(1c)

    R =  + i-score +AQ + 1CFO + ATAC + 1(CFO*i-score) + 1(CFO*AQ)
           + A(TAC*i-score) + A(TAC*AQ) +                                                B(2c)


The coefficients, E in (1c), and 1 and A in (2c) assess the additional effect of the accrual

quality measure on the weights of earnings, CFO and TAC, respectively, in the presence of

i-score. The coefficients are expected to be significant to the extent that AQ has information

about (poor) earnings quality that is not reflected on i-score. In both models,  evaluates the

main effect of AQ on returns. The results are reported in Table 8.

                                      [Table 8 about here]

       In (1c), E is negative and significant (P<0.01), consistent with the theory that poor

accrual quality reduces the informativeness of earnings for security returns. Despite this

result, the effects of i-score on returns and on the weight of earnings remain reliably

negative; no apparent erosion in the size of the coefficient on i-score or on EARN*i-score.

Nonetheless, the model achieves a significant increase in R2 over (1b), indicating that AQ

has incremental information about earnings quality beyond the information in i-score. In

(2c), the effect of AQ on the slope coefficients is negative and pronounced for TAC, but

positive, albeit insignificant for CFO. Based on this latter result, one infers that deterioration

in accrual quality does not affect market’s interpretation of the information in CFO: The

pricing effect of AQ is largely restricted to the information in accruals. The model achieves

an appreciable increase in R2 over (2b), indicating incremental contribution of AQ in

explaining cross-sectional variation in the pricing of earnings components, even with i-

score in the model. It is noteworthy, however, that the effect of i-score on returns and on the

weights of CFO and TAC remain reliably negative and unaffected by the inclusion of AQ.
                                                                                                 29


These results suggest strongly that, compared to AQ, i-score contains more robust and

distinct information used by investors to assess the valuation relevance of reported earnings.


5.4. Other robustness checks

       For additional robustness check, I replace returns by the fiscal year-end price per

share, and define earnings, CFO, TAC, and TAC components on per share basis, and re-

estimate (1b), (2b), and (3b). Theory suggests that per share price-earnings models are

better specified than returns-earnings models in that the slope coefficients are less biased,

although the former are more vulnerable to econometric problems (Kothari and Zimmerman

1995). The results of these revised models (not reported for brevity) show that the effect of

i-score on the weights of earnings, CFO, TAC, and TAC components remain as shown for

the returns models. Next, I define returns as the fiscal year-end price scaled by the

beginning-period price (adjusted for stock splits). Earnings, CFO, TAC, and TAC

components are similarly scaled by the beginning-period price. The results based on these

alternative variable definitions (not shown) are virtually the same as those based on monthly

compounded returns. Moreover, the inclusion of either a market return index or firm-year

market beta (based on the two-factor market model) does not alter the effect of i-score on

the weights of the earnings variables or the significance of the effect.


5.5. External monitoring and the effect of i-score on earnings weight

       Arguably, exposure to EM incentive reduces investors’ confidence in the reliability

of financial reports. Howbeit, external monitoring has been shown to mediate investors’

pessimism about the reliability of accounting reports (Dietrich, Harris, and Muller 2000;

Muller and Riedl 2002). External monitoring raises the risk to firms that reporting gimmicks
                                                                                                                30


and improprieties will be detected. Detection of accounting improprieties is often followed

by severe reprisals to the firm and its managers in the form of large stock price declines, job

loss, and bankruptcy (Palmrose and Scholz 2004). Given the scale of such costs, managers

will be less inclined to manipulate earnings when the risk of detection is high. Accordingly,

to the extent that investors view external monitoring as an effective check on managers’

predisposition to reporting improprieties/accounting distortions, the effect of EM incentives

on the valuation weight of earnings will decline as external monitoring increases.

         To test the preceding prediction, I use the percentage of a firm’s stocks held by

institutions and the number of analysts’ following as proxies for external monitoring.

Institutional investors actively monitor firms in which they have interests, and are predicted

to bolster investors’ confidence in the reliability of financial reports (Lev 1988, 1992;

O’Brien and Bhushan 1990, Porter 1992; Bushee 1998; Del-Guercio and Hawkins 1999;

Gillan and Starks 2000).19 Analysts’ information collection role parallels the institutional

investors’ (O’Brien and Bhushan 1990). Several studies associate the richness and quality

of accounting reports with analyst’s following (Lang and Lundholm 1996). A corollary that

emerges from the results is that the number of analysts’ that follow a firm bolsters

investors’ confidence in the firm and its financial reports.

         For the test, I divide the sample into three equal portfolios--low, medium, and high

external monitoring--ranked, first, by the proportion of stocks held by institutional investors

and, second, by the number of analysts’ following. Next, I re-estimate (1b) and (2b) across

the three groups for each external monitoring proxy. Table 9 presents the results.


19
  In fact, DeFond and Jiambalvo (1991) show that overstatement of earnings occur more often among firms that
have diffuse ownerships, giving credence to the view that institutional investors provide effective monitoring and
deterrence on managers’ predisposition to reporting improprieties.
                                                                                                                    31


                                            [Table 9 about here]

         In Panels A and B, the effects of i-score on returns and on the weights of earnings,

CFO, and TAC remain reliably negative across the ranks of institutional holding and ranks

of analysts following. Strikingly, however, the negative effect of i-score on the weights of

earnings, CFO, and TAC in both panels is most pronounced for the high external

monitoring group. This result suggests that both proxies of external monitoring do not

alleviate the earnings quality problems connected with EM incentives. Rather, the results

imply that each proxy is positively associated with investor sensitivity to the earnings

information problems in i-score. Balsam et al. (2002) explain that institutional stock holding

(and analysts’ following) is a proxy for sophisticated investors who recognize elevated

exposures to earnings management more fully and readily, and price reported earning

accordingly (see, also, Bartov et al. 2000; Jiambalvo et al. 2002). Note, however, that the

main effects of earnings, CFO, and TAC on returns increase with the ranks of institutional

holding (Panel A) and analysts’ following (Panel B). A plausible view of these findings is

that sophisticated investors and rich information environment improve earnings valuation

multiples, but also facilitate more prompt and complete assessment of the uncertainties

associated with EM incentives.20



20
   For further insight, I use the percentage of outside board members and big 4(5) versus non-big 4(5) audit firms
as alternative proxies of monitoring. The unreported results show no discernible variation in the slope effect of i-
score across high and low ranks of these latter proxies. This finding is, perhaps, unsurprising in that prior studies
find only a modest relation between governance factors and managers’ decisions (Dechow, Sloan, and Sweeny
1996; Larcker, Richardson, and Tuna 2004; Agrawal and Chadha 2005). More important, investors may place a
limited confidence in the ability of board governance to moderate EM incentives, given the economic interest
many board members have in the firms. Similarly, investors may be less willing to rely on auditors’ opinion to
assess EM risk given the preponderance of clean audit opinions for firms that were later confirmed to have
managed earnings. Browning and Weil (2002), for example, observe that ―An opinion letter from a Big Five
accounting firm, once viewed as a trusted seal of approval, now may not carry the imprimatur of authority.‖
Feltham, Hughes, and Simunic (1991) and Muller and Riedl (2002) also show that auditors’ attestation does not
appear to reduce the information asymmetry between the firm and outsiders.
                                                                                                    32


5.6. Additional considerations

       The principle that underlies the i-score construct is that total variation in each

indicator comprises a component that is common to other indicators and a component

unique to the indicator. The unique component is akin to a latent signal unrelated to the EM

incentives. Such a feature raises the prospect that the unique components have pricing

information unrelated to the EM incentive effect. This is especially tenable for indicators

that are primitive value drivers or signals of long-run economic activity. For example, the

agency cost of debt modeled by Myers and Majluf (1984) and Jensen (1986) portends a

valuation consequence for leverage, notwithstanding any effect debt may have on EM

incentives. Similarly, a loss episode or dividend omission may signal future cash flow

problems despite any reporting incentives they may create. To capture such effects and

simultaneously test the robustness of i-score effect in the presence of the indicators, I

include the indicators as separate variables in the model. The model is given as:

R=+    ARN +  i-score * ARN  +  LEV +  MSS +  BNS +  STO +  SZE
        i 1
               i
                   i

                       i 1
                              i
                                        i
                                             1         2         3        4         5


      + 6FDS + 7LSS + 8SGL + 9ASL + 10RSK + 11REC + 12REG + 13GRI +                  (4)

EARNi = 1, …, 7  {CFO, AR, INV, AP, DTX, OT, DPR}; 1…13 reflect the impact

of the corresponding indicators on returns. Table 10 presents the results.

                                     [Table 10 about here]

       The coefficients on the interactions between i-score and earnings components

remain reliably negative in the presence of the indicators. However, the results for the

indicators provide new insights: Adding the indicators in the model produces a significant

increase in R2 (about 17% increase). The weights of book-to-market equity, leverage, stock

options, and dividend omission are reliably negative (P < 0.01), whereas the weights of
                                                                                                               33


capital-expenditure growth and bonus are positive (P < 0.01). The results imply that these

indicators have pricing effects beyond their common effect on earnings uncertainty.21

        To test whether book-to-market ratio that has the highest factor loading captures the

information in i-score parsimoniously, I replace i-score in (3b) with book-to-market ratio

indicator and re-estimate the model. The results based on this latter specification (not shown

for brevity) are generally consistent with the predicted effect of incentive exposure on the

weights of earnings components. However, the R2 is significantly smaller than the R2

provided by i-score. Next, I replace i-score with alternative indictors of earnings quality

identified in the literature-- accrual quality measure derived from the modified Dechow and

Dichev (2002) model and absolute size of discretionary accruals derived from the modified

Jones model (see, Dechow, et al. 1995). The effects of both indicators on the weights of

earnings components (not reported) are generally in the expected direction, but the R2 based

on each alternative proxy is significantly lower than the R2 from the i-score specification (P

< 0.0001). Altogether, the results suggest that i-score models the common information in

the indicators more efficiently than each alternative proxy used here.


6. Conclusion

        In this study, I construct a score of EM incentives based on several indicators of

contracting and economic factors that predispose firms to manage earnings and examine the

effect of the stylized score on the valuation weight of earnings and earnings components.


21
  The effects of the indicators are consistent with evidence in the literature. For example, book-to-market equity
and leverage are traditional indicators of imperiled long-term growth and have been shown to exert downward
pressures on stock prices (Collins and Kothari 1989; Dhaliwal, Lee, and Fargher 1991). The negative weight on
stock option is perhaps related to the dilution effect hypothesis advanced by Aboody (1996) or dividend-reduction
effect reported by Lambert, Lanen, and Larcker (1989). Chamberlain and Hsieh (1999) also note that stock options
have the features of a corporate expense and can be expected to be priced negatively. Dividend omission has also
been shown to impact prices negatively. The positive effect of the investment growth is also suggested by theory.
                                                                                               34


The major hypothesis is that the amount of earnings and earnings components that map into

security returns are negatively associated with EM incentive exposure. The prediction rests

on the premise that such exposure raises the information risk of earnings, where information

risk of earnings is the risk that reported earnings are weakly informative or misleading

about the firm attributes priced by investors.

       As predicted, the impact of the exposure measure on the valuation weight of

earnings is reliably negative. Moreover, the exposure measure has a negative impact on

returns, indicating that the market effects of EM incentives extend beyond the discount on

the weight of earnings. The results further show that the effect of the exposure measure on

CFO, total accruals, and accrual components is reliably negative, which suggests that the

valuation role of accruals as well as that of CFO is decreasing in a firm’s exposure to

earnings management pressures. For the accrual components, the negative effect is most

pronounced for the inventory accrual and least for the accounts-payable accrual.

Furthermore, the inclusion of the exposure variable in the models produces marked

improvement in R2 over the more restricted models. These results are robust across

alternative price- and returns-earnings models.

       I also consider the effect of external monitoring on the relation between the

exposure construct and valuation weight of earnings. Specifically, I test whether

institutional stock ownership and analysts’ following mediate the impact of EM exposure on

the valuation weight of earnings and earnings components; the general premise is that

external monitoring discourages opportunistic reporting actions and thus reduces the

information problems associated with EM incentives. The results show that the negative

effects of the exposure measure on the weight of earnings and earnings components
                                                                                                       35


increase with the rank of institutional stock holding and rank of analysts’ following

increases. These results suggest that both indicators of external monitoring do not alleviate

the information problems associated with extant earnings management pressures, but rather

appear to proxy for investor sophistication and market’s ability to more fully recognize and

price exposure to opportunistic reporting.

        The analysis presented here contributes to the literature on the economic effects of

earnings quality problems. In particular, the i-score construct represents an additional

dimension of earnings information risk that complements various measures of earnings

quality identified in the literature (e.g., persistence, predictability, variability). The construct

offers several advantages: It models the earnings management incentive from a variety of

observable factors that notably motivate reporting opportunism. In this context, the

construct avoids the omitted-variables problem that often characterizes analyses of the

incentive effects of a single incentive factor. Moreover, the exposure score is estimated for

each firm-year and, in that sense, facilitates empirical analysis of the contemporaneous

connection between a firm’s vulnerability to earnings management and investors’

perception of earnings quality.
                                                                                                                     36


                                                        Table 1

Definition of the incentive variables and Selection of the primary sample

LEV = Firm-year ratio of long-term debt to total assets adjusted by the industry median ratio (#9  #6
         – industry median LEV), where industry grouping is based on 2-digit SIC code. The
         adjustment is needed to control for industry-related differences in financial leverage.
MSS = Firm-year indicator of dividend action set equal to 1 if a firm missed dividend in the previous
         or current year and 0 otherwise; data on dividend omission are obtained from the merged
         CRSP-Compustat PDE file and corroborated/augmented by a search of Lexis-Nexis using
         the key terms ―omit dividend,‖ ―dividend omission‖, ―missed dividend.‖
BNS = Firm-year bonus payout divided by the beginning-period book value of equity (Bonus  #60).
STO = Black-Scholes value of stock options granted divided by the beginning-period market value of
         equity (Black-Scholes option value  #25  #199).
SZE = Firm-year indicator of political exposure set equal to the inverse of the log of total assets
         adjusted by industry-median total assets to reduce industry-related scale effects; the inverse
         transformation aligns the indicator with the predicted effect on EM incentive.
FDS = Firm-year indicator of financial distress set equal to the inverse of the Altman’s Z-score of
         financial distress. The Z-score is based on the Altman’s (1968) model; the inverse
         transformation is made to align the indicator with the predicted incentive effect.
LSS = Firm-year indicator of earnings distress set equal to 1 if earnings before extraordinary items
         were negative in the previous year and 0 otherwise.
SGL = Firm-year indicator of small gain or loss set equal to 1 if the ratio of reported earnings to
         market value of equity lies between -0.005 and 0.005, and 0 otherwise.
ASL = Firm-year asset sales derived as the sum of asset and investment sales divided by the
         beginning-period net property, plant, & equipment ((#107 + #109)  #8).
RSK = Firm-year ratio of the beginning-period book value of equity to the beginning-period market
         value of equity (#60  #25  #199).
REC = Firm-year indicator set equal to 1 if the mean of IBES analysts’ recommendation for the fiscal
         year is less than 2; the indicator is set equal to 0 if the mean of IBES analysts’
         recommendation is between 2 and 4 inclusive; the indicator is set equal to -1 if the mean of
         IBES analysts’ recommendation is greater than 4. The IBES consensus recommendation ranges
              from 1 to 5, where 1 corresponds to ―Strong buy,‖ 2 corresponds to ―Buy,‖ 3 corresponds to ―Hold,‖
              4 corresponds to ―Underperforming,‖ and 5 corresponds ―Sell.‖
REG = Firm-year indicator of regulatory exposure set equal to 1 if a firm is in the drug industry (SIC
        2833-2836), electric & gas utility industry (SIC 4911 & 4931), or in banking-deposit
        institutions (SIC 6011-6411), and 0 otherwise.
GRI= Firm-year growth in capital expenditures measured as the difference between current and prior
        year capital expenditures divided by beginning period net property, plant, & equipment
        (#128  #8), adjusted by the industry median growth in capital expenditures. The
        adjustment is made to control for industry-related differences in capital spending patterns.
                                                                                               Firms        Firm-years
Initial Compustat data from 1992 to 2002                                                        4,197          45,970
Remaining sample with valid total assets, book value of equity, and income                      3,545          25,532
Remaining sample with valid security returns from CRSP                                          3,260          21,578
Remaining sample with complete firm-year record for incentive variables†                        1,137           8,628
†
    A set of all incentive variables is required for every firm-year to obtain a valid firm-year factor score.
                                                                                                                    37


                                                    Table 2

Panel A: Simple statistics for the variables used to construct incentive index.

                                                           Mean                 Median               Stdev
Leverage (LEV)                                            0.20440               0.18987             0.15897
Dividend omission (MSS): YES/NO                             _na_                  _na_                _na_
Bonus/Book equity (BNS)‡                               0.000999955           0.000388697         0.003376340
Stock option value/market equity (STO)‡                0.001327490           0.000419556         0.005215632
1/Log_Total assets (SZE)                                  0.13390               0.12958            0.030087
1/Z-score (FDS)                                           0.36718               0.30002             1.80142
Prior period loss (LSS): YES/NO                             _na_                  _na_                _na_
Small gain/loss (SGL)                                  0.000167886           0.000651043         0.003144817
Asset sales /PPE (ASL)                                   0.021307            0.000835359            0.14102
Book equity/Market equity (RSK)                           0.51605             0.4489400             0.36554
Analysts’ stock recommendation (REC)                      2.23674                  2.2              0.49514
Regulation exposure (REG): YES/NO                           _na_                  _na_                _na_
Capital expenditure growth (GRI)                       0.003249859           0.005803411            0.23680


Panel B: Summary results of the factor analysis
            Mean factor loading         Mean communalities                        Mean scoring coefficient
LEV               0.30641                   0.09388487                                  0.15777
MSS               0.39307                   0.15450214                                  0.20240
BNS               0.17536                   0.03075101                                  0.09030
STO               0.49772                   0.24772497                                  0.25628
SZE               0.68012                   0.46255750                                  0.35020
FDS               0.33088                   0.10948063                                  0.17037
LSS               0.44148                   0.19490063                                  0.22732
SGL               0.08406                   0.00706561                                  0.04328
ASL               0.06508                   0.00423580                                  0.03351
RSK               0.74505                   0.55509786                                  0.38364
REC               0.14800                   0.02190537                                  0.07621
REG              -0.14857                   0.02207389                                 -0.07650
GRI              -0.19466                   0.03789349                                 -0.10023

Mean factor score                               1.5034E-16
Median factor score                            -0.18354
Minimum factor score                           -2.08994
Maximum factor score                           11.4313
‡
 As an alternative to using actual bonus, I define a firm-year bonus indicator as 1 if a firm has a bonus plan in
place during the year, and 0 otherwise. I also define the firm-year option indicator similarly. When defined in
such manners, both variables load weakly on i-score. A plausible reason for such a result is the presence of
firm-years for which no incentive awards are granted by firms that have incentive plans. For such firms, non-
grant of awards may reflect peculiar contexts/strategy that is in conflict with EM incentives (see, e.g., Healy
1985, for a discussion of negative EM incentives in certain bonus contexts).
                                                                                                               38


                                                 Table 3

Panel A: Mean and median dispersion in analysts’ earnings forecasts across quintiles ranked
         by i-score

A2: Mean dispersion in analysts earnings forecasts
             Low i-score       2         3            3    High i-score            Diff.      (t_diff.)
Mean Disp1    0.04887      0.08082 0.08767         0.09543   0.18243              0.1336      (5.13)***
Mean Disp2    0.00757      0.02197 0.02412         0.02564   0.03261              0.0251      (3.64)***

A2: Median dispersion in analysts’ earnings forecasts
             Low i-score       2         3            3   High i-score             Diff.     (Z-value)a
Med. Disp1     0.01351     0.01642 0.01744 0.021834 0.04274                       0.0292     (23.20)***
Med. Disp2     0.00039     0.00063 0.00076        0.00101   0.00211               0.0017     (30.61)***

Disp1 = Standard deviation of analysts’ forecast from the IBES forecast data scaled by the absolute
          mean forecast
Disp2 = Standard deviation of analysts’ forecast from the IBES forecast data scaled by the fiscal
          year-end closing price per share




Panel B: Mean and median price-earnings ratio (P/E), debt costs (Debtcost1, Debtcost2),
         and poor accrual quality (AQ), across quintiles ranked by i-score
B1: Mean P/E, Debtcost1, Debtcost2, and AQ
            Low i-score      2            3             4    High i-score Diff.             (t_diff.)
P/E          28.1195      20.1551      17.6665       16.1915    9.9056    -18.214          (-15.81)***
Debtcost1    0.06750      0.06918      0.07047       0.07288   0.07539     0.0079          (11.52)***
Debtcost2    0.07050      0.07233      0.07384       0.07655   0.07884     0.0083          (11.25)***
AQ           0.08067      0.08288      0.08376       0.08506   0.08917     0.0085           (4.60)***

B2: Median P/E, Debtcost1, Debtcost2, and AQ
         Low i-score         2            3             4    High i-score   Diff            (Z-value)a
P/E       24.3835         17.7735      15.6897       14.9442   11.1596    -13.224          (-23.07)***
Debtcost1 0.06604         0.06772      0.06779       0.07221   0.07417     0.0081           (9.85)***
Debtcost2 0.06863         0.07002      0.07050       0.07437   0.07609     0.0075           (7.39)***
AQ        0.06084         0.06579      0.06680       0.06888   0.07063     0.0098            (2.07)**
P/E = income before extraordinary items divide by fiscal year-end market equity [#18(#25#199)];
Debtcost1 = year-end interest expense divide by average interest-bearing debt [#15((#9+#34)t–1
+(#9+#34)t)/2]; Debtcost2 = sum of year-end interest expense and capitalized interest divide by average
interest bearing debt [(#15+#147)((#9+#34)t–1+ (#9+#34)t)/2]; AQ = firm-year standard error of the residual
from the regression of working-capital accruals on lagged, current, and future values of CFO, augmented with
change in revenues and fixed operating assets (see, e.g., Francis et al. 2005, Model 1, page 302).
a
  Z-value is based on a non-parametric test of difference in the location of the median between the two
samples.
*Significant at 0.05 probability; **Significance at 0.01 probability; ***Significant at 0.0001 probability
                                                                                                     39


                                              Table 4

 Panel A: Comparison of the mean and median i-scores between the restating and non-restating firms
                     Restating firms          Non-restating firms
 A1:                  Mean i-score            Mean i-score                Difference
                        (t-value)              (t-value)                   (t-value)
 Year –3                0.41203                -0.0205                     0.43253
                        (4.22)***              (-0.94)                     (4.32)***
Year –2                  0.37844                 -0.0224                   0.40084
                         (3.88)***               (-1.05)                   (4.01)***
Year –1                  0.42689                 0.00285                   0.42404
                         (5.03)***               (1.27)                    (4.99)***
Restatement Year         0.5942                  0.1019                    0.4923
                         (7.97)***               (1.81)                    (5.30)***

 A2:                 Median i-score          Median i-score               Difference (Z-value)a
 Year –3               0.11308                 -0.22389                    0.33697 (3.21)***
Year –2                  0.14885                 -0.21164                  0.36049 (3.95)***
Year –1                  0.17537                 -0.21101                  0.38638 (4.08)***
Restatement Year         0.25035                 -0.18353                  0.4338 8(5.10)***


Panel B: Comparison of the mean and median i-scores between firms whose earnings restatements
         equal or exceed 5% of their book value of equity and an equal number of firms that have
         no record of earnings restatement during the sample period.

                     Restating sample     Non-restating sample
 B1:                  Mean i-score           Mean i-score                 Difference
                         (t-value)             (t-value)                   (t-value)
 Year –3                 0.6214582              -0.0732                    0.69466
                         (5.97)***              (-4.80)***                 (6.61)***
Year –2                  0.5727142               -0.0363                   0.608995
                         (5.35)***               (-2.97)**                 (5.65)***
Year –1                  1.0823912               -0.0206                   1.102997
                         (9.33)***               (-1.60)                   (9.45)***
Restatement Year 0       1.7780623               0.02149                   1.756568
                         (12.54)***              (1.09)                    (12.27)***

 B2:                 Median i-score          Median i-score               Difference (Z-value)a
 Year –3               0.28887                 -0.2279                     0.51677(4.17)***
Year –2                  0.30956                 -0.2177                   0.52728 (4.82)***
Year –1                  0.48634                 -0.2256                   0.71194 (6.95)***
Restatement Year 0       0.85936                 -0.1818                   1.04119 (7.88)***
                                                                                                                  40



                                               Table 4 (continued)

Panel C: The relation between the absolute size of earnings restatement and EM score

Model: Restate-sizet = b0 + b1i-scoret –  + errort
                                           b0                        b1           F-stat.         R2
                                     (P-value)                  (P-value)       (P-value)
Year of restatement,  = 0              0.043                      0.117           67.38        41.11%
                                      (0.0005)                   (0.0001)        (0.0001)
Year 1 before restatement,  = 1                0.088             0.178            71.27        43.53%
                                              (0.0002)          (0.0001)         (0.0001)
Year 2 before restatement,  = 2                0.099             0.080            32.38        19.50%
                                              (0.0001)          (0.0001)         (0.0001)
Year 3 before restatement,  = 3                0.113             0.037            18.14         8.01%
                                              (0.0001)          (0.0195)         (0.0195)
a
    Z-value is based on a non-parametric test of difference in location of the medians between the two samples.
*Significant at 0.05 probability; **Significant at 0.01 probability; ***Significant at 0.0001 probability.
                                                                                                              41

                                                    Table 5

Results of the benchmark association between returns and components of earnings
RB =  + EEARN + B                                                                                  B(1a)

R =  + 1CFOB + ATAC + B                                                                           B(2a)

R =  + 1CFO + B2BAR + B3INV + B4BAP + B5DTX + B6OT + B7BDPR + B B(3a)
                                             (1a)                      (2a)                     (3a)
                                           Coeff.                    Coeff.                   Coeff.
                                          (t-value)                 (t-value)                (t-value)
Intercept                                 0.0619                    0.0789                   0.0637
                                          (19.84)***                (25.79)***               (12.32)***
EARN                                      0.5000
                                          (19.02)***
CFO                                                                 0.6312                   0.6790
                                                                    (19.69)***               (12.59)***
TAC                                                                 0.4459
                                                                    (16.79)***



AR                                                                                          0.7899
                                                                                             (9.45)***
INV                                                                                         0.4723
                                                                                             (5.64)***
AP                                                                                          -0.5556
                                                                                             (-6.25)***
DTX                                                                                         -0.4263
                                                                                             (-1.62)
OT                                                                                          0.5646
                                                                                             (5.77)***
DPR                                                                                          -0.3002
                                                                                             (-3.97)***
       R2                                   0.035                     0.058                   0.091

  
 R 2a  R 1a
    2
           2
                                                                     3.62
  (P-value)                                                          (0.0001)

  
 R 3a  R 2
    2
           2a                                                                                 3.71
  (P-value)                                                                                  (0.0001)
R = Returns from CRSP compounded monthly for 12 months ending three months after the fiscal year end;
EARN = Income before extraordinary items and discontinued operations (#18); AR = Change in accounts
receivables (#302); INV = Change in inventories (#303); AP = Change in accounts payables (#304); ATX
= Change in deferred income taxes (#305); OT = Net change in other current assets and liabilities (#307);
CFO = Net cash flow from operating activities (#308); DPR = Depreciation plus amortization charge (#14).
***Significant at 0.0001 probability.
                                                                                                                   42

                                                     Table 6
Mean coefficient on the interaction between each incentive indicator and earnings
Model: RB =  + EEARN + x(Xi*EARN) +  ;
Where Xi = 1, …, 13 {LEV, MSS, BNS, STO, SZE, FDS, LSS, SGL, ASL, RSK, REC, REG, GRI}

Incentive indicator                  Mean x [# negative out of 10]                          Median x
LEV                                             -0.077 [ 7 ]                                  -0.054
MSS                                             -0.393 [ 7 ]                                  -0.229
BNS                                             -0.033 [ 8 ]                                  -0.064
STO                                             -0.255 [ 8 ]                                  -0.129
SZE                                             -0.233 [ 10 ]                                 -0.067
FDS                                             -0.136 [ 7 ]                                  -0.134
LSS                                             -0.814 [ 9 ]                                  -0.679
SGL                                             -0.506 [ 8 ]                                  -0.148
ASL                                             -0.067 [ 7 ]                                  -0.165
RSK                                             -0.002 [ 7 ]                                  -0.001
REC                                             -0.127 [ 7 ]                                  -0.066
REG                                              0.080 [ 4 ]                                   0.075
GRI                                              0.524 [ 3 ]                                   0.479
R = Returns from CRSP compounded monthly for 12 months ending three months after the fiscal year end;
EARN = Income before extraordinary items and discontinued operations (#18);
LEV = Firm-year ratio of long-term debt to total assets adjusted by the industry median leverage ratio (#9  #6
   – industry median LEV), where industry grouping is based on 2-digit SIC.
MSS = Firm-year indicator of dividend action set equal to 1 if a firm missed dividend in the previous or current
   year and 0 otherwise; dividend omission are obtained from the merged CRSP-Compustat PDE file and
   corroborated/augmented by a search of Lexis-Nexis.
BNS = Firm-year bonus payout divided by the beginning-period book value of equity (Bonus  #60).
STO = Black-Scholes value of stock options grants divided by the beginning-period market value of equity
   (Black-Scholes option value  #25  #199)
SZE = Firm-year indicator of political exposure set equal to the inverse of the log of total assets adjusted by
   industry-median total assets to reduce industry-related scale effects.
FDS = Firm-year indicator of financial distress set equal to the inverse of the Altman’s Z-score of financial
   distress. The Z-score is based on the Altman’s (1968) model.
LSS = Firm-year indicator of earnings distress set equal to 1 if earnings before extraordinary items were
   negative in the previous year and 0 otherwise.
SGL = Firm-year indicator of small gains or losses set equal to 1 if the ratio of reported earnings to market
   value of equity lies between -0.005 and 0.005, and 0 otherwise.
ASL = Firm-year asset sales derived as the sum of asset and investment sales divided by the beginning-period
   net property, plant, & equipment ((#107 + #109)  #8))
RSK = Firm-year ratio of the beginning-period book value of equity to the beginning-period market value of
   equity (#60  #25  #199).
REC = Firm-year indicator set equal to 1 if the mean of IBES analysts’ recommendation code over the fiscal
   year is less than 2; the indicator is set equal to 0 if the mean of the IBES analysts’ stock recommendation
   code is between 2 and 4 inclusive; the indicator is set equal to -1 if the mean of the IBES analysts’
   recommendation code is greater than 4.
REG = Firm-year indicator of regulatory exposure set equal to 1 if a firm is in the drug industry (SIC 2833-
   2836), electric & gas utility industry (SIC 4911 & 4931), or in banking-deposit institutions (SIC 6011 -
   6411), and 0 otherwise.
GRI= Firm-year growth in capital expenditures measured as the difference between the current and prior year
   capital expenditures divided by beginning period net property, plant, & equipment (#128  #8), adjusted by
   the industry median growth in capital expenditures.
                                                                                                             43

                                                   Table 7
Analysis of the effect of EM exposure on the association between security returns and weight of
earnings and earnings components in re

R =  + i-score +EEARNB +E(EARN*i-score) + B                                                    B(1b)

R =  + i-score +1CFO+ATAC + 1(CFO*i-score) + A(TAC*i-score) + B         B(2b)
R =  + i-score +1CFO+B2BAR + B3INV + B4BAP + B5DTX + B6OT + B7BDPR
      + 1(CFO*i-score) + 2(AR*i-score) + 3(INV*i-score) + 4(AP*i-score)
      + 5(DTX*i-score) + 6(OT*i-score) + 7(DPR*i-score) + B              B(3b)

                              (1b)                    (2b)                        (3b)
                              Coeff. (t-value)       Coeff. (t-value)            Coeff.   (t-vaue)
Intercept                     0.0516 (19.84)***      0.0749 (25.79)***           0.0637 (12.32)***
i-score                      -0.0707 (-23.99)***     -0.0818 (-26.43)***        -0.1114 (-22.45)***
EARN                          0.6288 (17.82)***
EARN*i-score                 -0.1388 (-13.16)***
CFO                                                  0.7974 (20.19)***           0.9991 (14.74)***
TAC                                                  0.5009 (13.89)***
CFO*i-score                                          -0.1439 (-10.70)***        -0.2577 (-4.77)***
TAC*i-score                                          -0.1177 (-10.94)***
AR                                                                              1.1285   (11.80)***
INV                                                                             1.0776   (7.99)***
AP                                                                             -0.9100   (-9.08)***
DTX                                                                            -0.2082   (-2.44)*
OT                                                                              0.9344   (8.61)***
DPR                                                                             -0.2337   (-2.13)*
AR*i-score                                                                     -0.1311   (-2.87)**
INV*i-score                                                                    -0.5354   (-6.92)***
AP*i-score                                                                      0.0456   (0.66)
DTX*i-score                                                                    -0.4175   (-5.21)***
OT*i-score                                                                     -0.3687   (-4.60)***
DPR*i-score                                                                     -0.1875   (-8.69)***
          R2                      0.103                   0.158                       0.237
  R2
     mb    R2a
             m    ; m =1, 2, 3    9.01                      9.27                     11.29
    (p-value)                     (0.0001)               (0.0001)                   (0.0001)

  
 R 2b  R 1b
    2
           2
                                                            3.07
  (p-value)                                              (0.0011)

  
 R 3b  R 2
    2
           2b                                                                        5.65
  (p-value)                                                                         (0.0001)
R = Returns from CRSP compounded monthly for 12 months ending three months after the fiscal year end;
EARN = Income before extraordinary items and discontinued operations (#18); AR = Change in accounts
receivables (#302); INV = Change in inventories (#303); AP = Change in accounts payables (#304); ATX
= Change in deferred income taxes (#305); OT = Net change in other current assets and liabilities (#307);
CFO = Net cash flow from operating activities (#308); DPR = Depreciation plus amortization charge (#14).
*Significant at 0.05 probability; **Significant at 0.01 probability; ***Significant at 0.0001 probability.
                                                                                                             44

                                                     Table 8
Analysis of the effects of EM incentives exposure (i-score) and accrual quality (AQ) on the
association between security returns and earnings and earnings components.

R =  + i-score +AQ + EEARN + E(EARN*i-score) + E(EARN*AQ) +                                 B(1c)

R =  + i-score +AQ + 1CFO + ATAC + 1(CFO*i-score) + 1(CFO*AQ)
         + A(TAC*i-score) + A(TAC*AQ) +                                                          B(2c)

                                 (1c)                                       (2c)
                                Coeff     (t-value)                        Coeff.     (t-value)
Intercept                       0.0416    (19.84)***                       0.0311     (21.22)***
i-score                        -0.0770     (-18.87)***                    -0.0778     (-23.35)***
AQ                             -0.1267     (-1.42)                        -0.2641     (-2.38)*
EARN                            1.038      (14.14)***
EARN*i-score                    -0.224     (-11.84)***
EARN*AQ                         -1.180     (-3.37)**
CFO                                                                        1.2228     (17.02)***
TAC                                                                        1.1721     (13.89)***
CFO*i-score                                                               -0.2881     (-12.47)***
TAC*i-score                                                               -0.1899     (-9.95)***
CFO*AQ                                                                     0.2917     (0.97)
TAC*AQ                                                                    -1.6577     (-3.92)***
     2
 R                              0.124                                       0.179


                  
 R 2  R 2 b ; m =1, 2
    mc     m                     2.38                                       2.11
  (p-value)                    (0.0087)                                   (0.01743)

 
 R 2  R 1c
    2c
          2
                                                                           5.74
  (p-value)                                                               (0.0001)
R = Returns from CRSP compounded monthly for 12 months ending three months after the fiscal year end;
EARN = Income before extraordinary items and discontinued operations (#18); AR = Change in accounts
receivables (#302); INV = Change in inventories (#303); AP = Change in accounts payables (#304); ATX
= Change in deferred income taxes (#305); OT = Net change in other current assets and liabilities (#307);
CFO = Net cash flow from operating activities (#308); DPR = Depreciation plus amortization charge (#14).
*Significant at 0.05 probability; **Significant at 0.01 probability; ***Significant at 0.0001 probability.
                                                                                                             45

                                                    Table 9

Model (1b): RB = B+ i-score + EEARNB B+ E(EARN*i-score) + 
Model (2b): RB = B+ i-score + 1CFO + ATAC + 1(CFO*i-score) + A(TAC*i-score) + 
Panel A: The effect of EM exposure on the weight of earnings and earnings components across three
        partitions of firms based on the percentage of institutional holding of stocks
            Low external monitoring           Medium external monitoring         High external monitoring
Model (1b)     Coeff     t-value                  Coeff.    t-value                   Coeff.    t-value
Intercept      0.03925 5.34***                     0.02437 2.77***                    0.02726 2.82**
i-score       -0.06557 -10.53***                  -0.04402 -6.25***                  -0.02250 -2.70*
EARN           0.28062 3.69**                      0.86030 8.27***                    1.22817 9.02***
EARN*i-score -0.04325 -1.70                       -0.30566 -7.41***                  -0.47123 -7.32***
 R2            0.0871                              0.1081                             0.1207

Model (2b)           Coeff    t-value               Coeff.       t-value             Coeff.       t-value
Intercept            0.05100 8.16***                 0.05640      8.41***             0.08257     12.56***
i-score             -0.07396 -11.35***              -0.07299     -9.78***            -0.05328     -6.56***
CFO                  0.50209 5.80***                 1.14323     10.17***             1.47462     10.19***
TAC                  0.11042 1.38                    0.63490      5.74***             1.18064      7.99***
CFO*i-score         -0.07577 -2.19*                 -0.32392     -6.25***            -0.36940     -7.53***
TAC*i-score         -0.01007 -0.39                  -0.27996     -6.54***            -0.44182     -6.43***
R2                   0.1076                          0.1145                           0.1483


Panel A: The effect of EM exposure on the weight of earnings and earnings components across three
        partitions of firms based on the number of analysts’ following
                 Low external monitoring        Medium external monitoring        High external monitoring
Model (1b)           Coeff t-value                 Coeff.     t-value               Coeff.      t-value
Intercept           0.07105 8.77***                 0.01910    2.14*                -0.00100 -2.01*
i-score            -0.05429 -8.61***               -0.05784 -9.01***                -0.05486 -8.93***
EARN                0.48774 5.95***                 1.13968 10.29***                 1.31844 14.10***
EARN*i-score       -0.08533 -3.61**                -0.3442    -8.44***              -0.45485 -11.62***
R2                  0.0949                          0.1188                           0.1375

Model (2b)            Coeff    t-value              Coeff.    t-value                Coeff.    t-value
Intercept            0.08381   12.05***              0.07222 11.84***                 0.06239 12.50***
i-score             -0.05854   -8.92***             -0.07754 -11.53***               -0.08501 -13.41***
CFO                  0.70508    7.81***              1.23633 10.61***                 1.39339 14.28***
TAC                  0.27172    3.17**               0.96155   8.39***                1.22250 12.82***
CFO*i-score         -0.16383   -5.16***             -0.35742 -7.24***                -0.34814 -6.83***
TAC*i-score         -0.03874   -1.57                -0.30288 -7.25***                -0.42147 -10.65***
R2                   0.1231                          0.1605                           0.1677
R = Returns from CRSP compounded monthly for 12 months ending three months after the fiscal year end;
EARN = Income before extraordinary items and discontinued operations (#18); CFO = Net cash flow from
operating activities (#308);TAC = EARN – CFO.
*Significant at 0.05 probability; **Significant at 0.01 probability; ***Significant at 0.0001 probability.
                                                                                                      46

                                               Table 10
Regression of returns on earnings components conditional on earnings management risk exposure
and controlling for the unique information in the incentive indicators

R =  + i-score +    ARN +  i-score* ARN  +  LEV +  MSS +  BNS +  STO
                     i 1
                            i
                                i

                                    i 1
                                           i
                                                    i
                                                            1     2         3        4


 + 5SZE + 6FDS + 7LSS + 8SGL + 9ASL + 10RSK + 11REC + 12REG + 13GRI                    (4)

Where EARNi = 1, 2, 3, …, 7  {CFO, AR, INV, AP, DTX, OT, DPR}

Variable                        Coefficient       (t-value)
Intercept                         0.05242           (7.33***
CFO                               1.50480         (16.70)***
AR                               1.65837         (15.07)***
INV                              1.17344           (9.77)***
AP                              -1.44585        (-12.55)***
DTX                             -0.76956          (-2.25)*
OT                               1.24506           (9.63)***
DPR                              -0.47083          (-3.17)**
CFO*i-score                      -0.28683          (-7.70)***
AR*i-score                      -0.17442          (-2.96)**
INV*i-score                     -0.30289          (-5.29)***
AP*i-score                        0.21697           (2.94)**
DTX*i-score                     -0.64377          (-2.63)**
OT*i-score                      -0.29822          (-3.23)**
DPR*i-score                       0.02218           (0.36)
LEV                              -0.02536          (-4.28)***
MSS                              -0.03784          (-2.57)**
BNS                               0.01888           (2.17)*
STO                              -0.01645          (-2.40)*
SZE                              -0.00999          (-1.06)
FDS                              -0.00695          (-1.04)
LSS                              -0.02711          (-0.46)
SGL                              -0.00384          (-0.75)
ASL                              -0.07982          (-0.68)
RSK                              -0.13103        (-13.55)***
REC                               0.00269           (1.85)
REG                               0.02329           (1.06)
GRI                               0.01860           (3.70)**
R2                                                0.274
 
 R 2  R 3b
    4
          2
                                                  2.72
  (P-value)                                      (0.0032)
                                                                                                                  47


                                           Table 8 (continued)
R = Returns from CRSP compounded monthly for 12 months ending three months after the fiscal year end;
EARN = Income before extraordinary items and discontinued operations (#18); AR = Change in accounts
receivables (#302); INV = Change in inventories (#303); AP = Change in accounts payables (#304); ATX
= Change in deferred income taxes (#305); OT = Net change in other current assets and liabilities (#307);
CFO = Net cash flow from operating activities (#308); DPR = Depreciation plus amortization charge (#14).
LEV = Firm-year ratio of long-term debt to total assets adjusted by the industry median leverage ratio (#9  #6
   – industry median LEV), where industry grouping is based on 2-digit SIC.
MSS = Firm-year indicator of dividend action set equal to 1 if a firm missed dividend in the previous or
   current year and 0 otherwise; dividend omission are obtained from the merged CRSP-Compustat PDE file
   and corroborated/augmented by a search of Lexis-Nexis.
BNS = Firm-year bonus payout divided by the beginning-period book value of equity (Bonus  #60).
STO = Black-Scholes value of stock options grants divided by the beginning-period market value of equity
   (Black-Scholes option value  #25  #199)
SZE = Firm-year indicator of political exposure set equal to the inverse of the log of total assets adjusted by
   industry-median total assets to reduce industry-related scale effects.
FDS = Firm-year indicator of financial distress set equal to the inverse of the Altman’s Z-score of financial
   distress. The Z-score is based on the Altman’s (1968) model.
LSS = Firm-year indicator of earnings distress set equal to 1 if earnings before extraordinary items were
   negative in the previous year and 0 otherwise.
SGL = Firm-year indicator of small gains or losses set equal to 1 if the ratio of reported earnings to market
   value of equity lies between -0.005 and 0.005, and 0 otherwise.
ASL = Firm-year asset sales derived as the sum of asset and investment sales divided by the beginning-period
   net property, plant, & equipment ((#107 + #109)  #8)).
RSK = Firm-year ratio of the beginning-period book value of equity to the beginning-period market value of
   equity (#60  #25  #199).
REC = Firm-year indicator set equal to 1 if the mean of IBES analysts’ recommendation code over the fiscal
   year is less than 2; the indicator is set equal to 0 if the mean of the IBES analysts’ stock recommendation
   code is between 2 and 4 inclusive; the indicator is set equal to -1 if the mean of the IBES analysts’
   recommendation code is greater than 4.
REG = Firm-year indicator of regulatory exposure set equal to 1 if a firm is in the drug industry (SIC 2833-
   2836), electric & gas utility industry (SIC 4911 & 4931), or in banking-deposit institutions (SIC 6011 -
   6411), and 0 otherwise.
GRI= Firm-year growth in capital expenditures measured as the difference between the current and prior year
   capital expenditures divided by beginning period net property, plant, & equipment (#128  #8), adjusted
   by the industry median growth in capital expenditures.
                                                                                                       48


                                              References

Abarbanell, J. and R. Lehavy. 2003. Can Stock Recommendations Predict Earnings Management and
   Analysts’ Earnings Forecast Errors. Journal of Accounting Research 41 (March): 1 - 31.
Aboody, D. 1996. Market valuation of Employee Stock Options. Journal of Accounting and
   Economics 22 (August-December): 357 - 391.
Agrawal, A., and S. Chadha. 2005. Corporate Governance and Accounting Scandals. Journal of Law
   and Economics October (2005): Forthcoming.

Ahrens, Robert W. 2002. Investors Mistrust Lingers: USA Today (July 10): Money Section.
Alexander, G.J., P. George Benson, Joan M. Kampmeyer. 1984. Investigating the Valuation Effects
   of Announcements of Voluntary Corporate Selloffs. Journal of Finance 39 (June): 503 - 517.
Apostolou, Nicholas G., and Larry D. Crumbley. 1990. Forensic Investing: Red Flags. At
    http://www.bus.lsu.edu/accounting/faculty/napostolou/forensic.html#Red%20Flags.

Ayers, Frances L. 1986. Characteristics of Firms Electing Early Adoption of SFAS 52. Journal of
   Accounting and Economics 8 (June) 143- 158.
Altman, E. 1968. Financial ratios, discriminant analysis and the prediction of corporate bankruptcy.
   Journal of Finance 23 (September): 589 - 609.
Baber, W.R., S.U. Janakiraman, and S. Kang. 1996. Investment opportunities and the structure of
   executive compensation. Journal of Accounting and Economics 21 (June): 297-318.
Ball, R., Robin, A., Wu, J., 2003. Incentives versus Standards: Properties of Accounting Income in Four
    East Asian countries. Journal of Accounting and Economics 36, 1-3, (December): 235 - 270.
Balsam, Steven, Eli Bartov, and Carol Marquardt. 2002. Accruals Management, Investor
    Sophistication, and Equity Valuation: Evidence from 10-Q Filings. Journal of Accounting
    Research 40 (September): 987 - 1012.
Barron, Orie E., Pamela S. Stuerke. 1998. Dispersion in Analysts’ Earnings Forecasts as a Measure of
    Uncertainty. Journal of Accounting, Auditing, and Finance 13, 3 (Summer): 243-268.
Barth, M.E., D.P. Cram, and K.K. Nelson. 2001. Accruals and the Prediction of Future Cash Flows.
    The Accounting Review 76 (January): 27 - 58.
Bartov, Eli. 1993. The Timing of Asset Sales and Earnings Manipulation. The Accounting Review 68
    (October): 840-855.
Bartov, Eli, Suresh Radhakrishnan, and Itzhak Krinsky. 2000. Investor Sophistication and Patterns in
    Stock Returns After Earnings Announcements The Accounting Review 75 (January): 43-63.
Bartov, Eli, and Mohanram, Partha. 2004. Private Information, Earnings Manipulation, and Executive
    Stock-Option Exercises. The Accounting Review 79 (Oct): 889 - 920.
Bhattacharya, Utpal, Hazem Daouk, and Michael Welker. 2003. The World Price of Earnings
   Opacity. The Accounting Review 78, 3 (July): 641 - 678
Bhattacharya, Utpal, and Matthew Spiegel. 1991. Insiders, Outsiders, and Market Breakdowns. The
   Review of Financial Studies 4, 2: 255 - 282.
                                                                                                    49

Beaver, W.H., and S.G. Ryan. 1983. Accounting Fundamentals of the Book-to-Market Ratio.
   Financial Analysts’ Journal 49 (November-December): 50 - 56.
Berenson, A. 2003. The Number: How the Drive for Quarterly Earnings Corrupted Wall Street and
   Corporate America. Random House, New York.
Berger, P.E., Ofek, E., Swary, I., 1996. Investor valuation of abandonment option. Journal of
   Financial Economics 42 (October): 257 - 287.
Bowen, R.M., E.W. Noreen, and J.M. Lacey. 1981. Determinants of the Corporate Decision to
   Capitalize Interest. Journal of Accounting and Economics 3 (August): 151 - 179.

Browning, E. S. and Jonathan Weil, 2002, Burden of doubt: Stocks Take a Beating as
   Accounting Worries Spread beyond Enron, Wall Street Journal, January 30, A1.
Burgstahler, D., and I. Dichev. 1997. Earnings Management to Avoid Earnings Decreases and Losses.
   Journal of Accounting and Economics 24 (December): 99 - 126.
Burgstahler, D.C., and M.J. Eames. 2003. Earnings Management to Avoid Losses and Earnings
   Decreases: Are Analysts Fooled? Contemporary Accounting Research 20 (Summer): 253 - 294.
Bushee, Brian. 1998. The Influence of Institutional Investors on Myopic R&D Investment Behavior.
   The Accounting Review 73 (July): 305 - 333.
Chamberlain, S. L., and S. J. Hsieh. 1999. The Effectiveness of SFAS 123 Employee Stock Option
   Cost Disclosures. Working paper, Santa Clara University.
Chaney, Paul K., and Debra C. Jeter. 1992. The Effect of Size on the Magnitude of Long-Window
   Earnings Response Coefficients. Contemporary Accounting Research 8 (Spring): 540-560.

Chen, Shuping, and James Jiambalvo. 2004. The Relation between Dispersion in Analysts’
   Forecasts and Stock Returns: Optimism Versus Drift. Working Paper: University of
   Washington.
Christie, A.A. 1990. Aggregation of Test Statistics: An Evaluation of the Evidence on Contracting
   and Size Hypothesis. Journal of Accounting and Economics 12 (January): 15 - 36.
Collins, Dan W., and S.P. Kothari. 1989. An Analysis of Intertemporal and Cross-sectional
    Determinants of Earnings Response Coefficients. Journal of Accounting and Economics 11
    (July): 143 - 181.
Coughlan, A.T., and R.M. Schmidt. 1985. Executive Compensation, Management Turnover, and
   Firm Performance: An Empirical Investigation. Journal of Accounting and Economics 7 (April):
   43 - 66.
Cramer, J.S. 1987. Mean and Variance of R2 in Small and Moderate Samples. Journal of
   Econometrics 35 (July): 253 - 266.
Daley, L.A., D.W. Senkow, and R.L. Vigeland. 1988. Analysts’ Forecasts, Earnings Variability, and
   Option Pricing: Empirical Evidence. The Accounting Review 63, 4 (October): 563-585.
Daley, L.A., and R.L. Vigeland. 1983. The Effects of Debt Covenants and Political Costs on the
   Choice of Accounting Methods: The Case of Accounting for R&D Costs. Journal of Accounting
   and Economics 5 (December): 195 - 211.
                                                                                                       50

DeAngelo, L. 1981. Auditor Size and Audit Quality. Journal of Accounting and Economics 3
   (December): 183 - 199.
DeAngelo, H., L. DeAngelo, and D.J. Skinner. 1994. Accounting Choice in Troubled Companies.
   Journal of Accounting and Economics 17 (January): 113 - 144.
Dechow, P.M., R.G. Sloan, and A. Sweeny. 1995. Detecting Earnings Management. The Accounting
   Review 70 (April): 193 - 225.
Dechow, P.M., R.G. Sloan, and A. Sweeny. 1996. Causes and Consequences of Earnings
   Manipulation: An Analysis of Firms Subject to Enforcement Actions by the SEC. Contemporary
   Accounting Research 13 (Spring): 1-36.
Dechow, P.M., and I.D. Dichev. 2002. The quality of accruals and earnings: The role of accrual
   estimation errors. The Accounting Review 77 (Supplement): 35 - 59.
Dechow, P.M., S.A. Richardson, and Irem Tuna. 2003. Why Are Earnings Kinky? An Examination of
   the Earnings Management Explanation. Review of Accounting Studies 8 (June): 35t5 - 384.
DeFond, Mark L. and James Jiambalvo. 1991. Incidence and Circumstances of Accounting Errors.
  The Accounting Review 66 (July): 643-655.
DeFond, M.L. and J. Jiambalvo. 1993. Factors Related to Auditor-Client Disagreements over Income-
   Increasing Accounting Methods. Contemporary Accounting Research 9 (Spring): 415 - 431.
DeFond, M.L. and J. Jiambalvo. 1994. Debt covenant violation and manipulation of accruals. Journal
   of Accounting and Economics 17 (January): 145-176.
DeFond, M. L., and C. W. Park. 1997. Smoothing Income in Anticipation of Future Earnings. Journal
   of Accounting and Economics 14 (July): 347 - 374.
DeFond, Mark L., and Chul W. Park. 1999. The Effect of Competition on CEO Turnover. Journal of
   Accounting and Economics 27 (February): 35-56.
Degeorge, F., J. Patel, and R. Zeckhauser. 1999. Earnings Management to Exceed Thresholds.
   Journal of Business 72 (January): 1 - 33.
Del Guercio, Diane, and Jennifer Hawkins. 1999. The Motivation and Impact of Pension Fund
    Activism. Journal of Financial Economics 52 (June): 293 - 340.
Dhaliwal, D.S. 1980. The Effect of the Firm’s Capital Structure on the Choice of Accounting
   Methods. The Accounting Review (January): 78 - 84.
Dhaliwal, Dan S., Kyung J. Lee, and Neil L. Fargher. 1991. The Association between Unexpected
   Earnings and Abnormal Security Returns in the Presence of Financial Leverage. Contemporary
   Accounting Research 8 (Fall): 20-41.

Diamond, Douglas W., and Robert E. Verrecchia. 1991. Disclosure, Liquidity, and the Cost of
   Capital. The Journal of Finance 46, 4 (September):1325-1360.
Dietrich. J.R., M.S. Harris, and K.A. Muller, III. 2000. The Reliability of Investment Property Fair
    Value Estimates. Journal of Accounting and Economics 30 (October): 125-158.
Duffie, Darrell, and David Lando. 2001. Term Structures of Credit Spreads with Incomplete
   Accounting Information. Econometric 69, 3 (May): 633- 664.
                                                                                                     51


Duke, J.C., and H.G. Hunt. 1990. An Empirical Examination of Debt Covenant Restrictions and
   Accounting-Related Debt Proxies. Journal of Accounting and Economics 12 (January): 45 - 63.
Easley, David, and Maureen O’Hara. 2004. Information and the Cost of Capital. The Journal of
    Finance 59, 4 (August): 1553-1584
Farell, K.A., and D.A. Whidbee. 2003. The Impact of Firm Performance Expectations on CEO
    Turnover and Replacement Decisions. Journal of Accounting and Economics 36 (December): 165
    - 196.
Fazzari, Steven M., R. Glenn Hubbard, G., Bruce C. Petersen. 1988. Financial constraints and
   corporate investments. Brookings Papers on Economic Activity 1, 141 - 195.
Feltham, G.A., J.S. Hughes, and D.A. Simunic. 1991. Empirical Assessment of the Impact of Auditor
    Quality on the Valuation of New Issues. Journal of Accounting and Economics 14 (December):
    375-99,
Feroz, E.H., K. Park, V.S. Pastena. 1991. The Financial and Market Effects of the SEC’s Accounting
    and Auditing Enforcement Releases. Journal of Accounting Research 29 (Supplement): 107-142.
Fields, Thomas D., Thomas Z. Lys, and Linda Vincent. 2001. Empirical Research on Accounting
    Choice. Journal of Accounting and Economics 31 (September): 255 - 307.
Francis, Jennifer, Ryan LaFond, Per Olsson, and Katherine Schipper. 2005. The Market Pricing of
    Accruals Quality. Journal of Accounting and Economics 39, 2 (June): 295 - 327.
Fudenberg, Drew and Jean Tirole. 1995. A Theory of Income and Dividend Smoothing Based on
   Incumbency Rents. The Journal of Political Economy 103 (February): 75 - 93.
Gaver, J. and K. Gaver. 1993. Additional evidence on the association between the investment
   opportunity set and corporate financing, dividend, and compensation policies. Journal of
   Accounting and Economics 16 (January/April/July): 125-160.
Graver, J.J., K.M. Graver, and J.R. Austin. 1995. Additional Evidence on Bonus Plans and Income
   Management. Journal of Accounting and Economics 19 (February): 3 - 28.
General Accounting Office. 2002. Financial Statement Restatements: Trends, Market Impacts,
   Regulatory Responses, and Remaining Challenges. General Accounting Office, GAO-03-138. At:
    http://www.gao.gov/new.items/d03138.pdf

Gillan, S. L., and L.T. Starks. 2000. Corporate Governance Proposals and Shareholder Activism: The
    Role of Institutional Investors. Journal of Financial Economics 57 (August): 275- 305.
Goyal, Vidhan K., and Chul W. Park. 2002. Board Leadership Structure and CEO Turnover. Journal
   of Corporate Finance 8 (January): 49 - 66.
Hagerman, R.L., and M.E. Zmijewski. 1979. Some Economic Determinants of Accounting Policy
   Choice. Journal of Accounting and Economics 1 (August): 141 - 161.
Hanlon, Michelle. 2005. The Persistence and Pricing of Earnings, Accruals, Cash Flows When Firms
   Have Large Book-Tax Differences. The Accounting Review 80 (January): 137- 166.
Hayn, C. 1995. The Information Content of Losses. Journal of Accounting and Economics 20
   (September): 125-153.
                                                                                                      52

Healy, P.M. 1985. The effect of bonus schemes on accounting decisions. Journal of Accounting and
   Economics 7 (April): 85 - 107.
Healy, P.M., and K.G. Palepu. 1990. Effectiveness of Accounting-Based Dividend Covenants.
   Journal of Accounting and Economics 12 (January): 97 - 123.
Healy, P.M. and J.M. Wahlen. 1999. A review of the earnings management literature and its
   implications for standard setting. Accounting Horizons 13 (December): 365 - 383.

Herrmann, Don, Tatsuo Inoue, and Wayne B. Thomas. 2003. The Sale of Assets to Manage Earnings
   in Japan. Journal of Accounting Research 41 (March): 89 - 109.

Hirst, Eric D., and Patrick E. Hopkins. 1998. Comprehensive Income Reporting and Analysts'
    Valuation Judgments. Journal of Accounting Research 36 (Supplement): 47 - 75.
Holthausen, R.W. 1981. Evidence on the Effect of Bond Covenants and Management Compensation
   Contracts on the Choice of Accounting Techniques: The Case of Depreciation Switch-back.
   Journal of Accounting and Economics 3 (March): 73 - 109.
Holthausen, R.W., and R.W. Leftwich. 1983. The Economic Consequences of Accounting Choice:
   Implications of Costly Contracting and Monitoring. Journal of Accounting and Economics 5
   (August): 77 - 117.
Holthausen, R.W., and Robert E. Verrecchia. 1988. The Effect of Sequential Information Releases on
   the Variance of Price Changes in an Intertemporal Multi-Asset Market. Journal of Accounting
   Research 26, 1 (Spring): 82 - 106.
Hunt, A., Moyer, S., Shevlin, T., 1996. Managing Interacting Accounting Measures to Meet multiple
   objectives: A Study of LIFO Firms. Journal of Accounting and Economics 21 (June): 339 - 374.
Hunt, Isaac C. 2000. Current SEC Financial Fraud Developments. Washington, DC: Securities and
   Exchange Commission (March 3): At www.sec.gov/news/speeches/spch351.htm.

Imhoff, E., Lobo, G., 1992. The effect of ex ante earnings uncertainty on earnings response
   coefficients. The Accounting Review 67, 2 (April): 427- 439.
Jensen, M.C. 1986. Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers. American
    Economic Review 76 (May): 323 - 329.
Jiambalvo, J., S. Rajgopal, and M. Venkatachalam, and. 2002. Institutional Ownership and the Extent
    to which Stock Prices Reflect Future Earnings. Contemporary Accounting Research 19 (Spring):
    117 - 145.
John, Kose, Larry H.P. Lang, and Jeffry Netter. 1992. The Voluntary Restructuring of Large Firms in
   Response to Performance Decline. The Journal of Finance 47 (July): 891 - 917.
Joos, Peter, and Geroge A. Plesko. 2005. Valuing Loss Firms. The Accounting Review 80 (July): 847-
    870.
Jones, J. 1991. Earnings management during import relief Investigations. Journal of Accounting
    Research 29 (Autumn): 193 - 228.
Kalay, A. 1979. Towards a Theory of Corporate Dividend Policy. Unpublished Ph.D. Thesis,
   University of Rochester, Rochester, NY.
                                                                                                      53

Key, K.G. 1997. Political cost incentives for earnings management in the cable television industry.
   Journal of Accounting and Economics 23 (November): 309 - 337.
Kinney, William, David Burgstahler, and Roger Martin. 2002. Earnings Surprise ―Materiality‖ as
   Measured by Stock Returns. Journal of Accounting Research 40, 5 (December): 1297-1329.
Kinney, William R., JR., Zoe-Vonna Palmrose, and Susan W. Scholz. 2004. Auditor Independence,
   Non-Audit Services, and Restatements: Was the U.S. Government Right? Journal of Accounting
   Research 42 (June): 561 - 588.
Koch, P. 1989. Accounting Changes and Firms Approaching Financial Distress. Working Paper.
   University of Illinois, Chicago.
Kothari, S.P., and Jerold L. Zimmerman. 1995. Price and Return Models. Journal of Accounting and
   Economics 20 (September): 155 - 192.
Lambert, R.A., W. N. Lanen, and D. F. Larcker. 1989. Executive Stock Option and Corporate
   Dividend Policy. Journal of Financial and Quantitative Analysis 24 (December): 409- 425.

Lang, L., A. Poulsen, and R. Stulz. 1995. Asset Sales, Firm Performance, and the Agency
   Cost of Managerial Discretion. Journal of Financial Economics (January): 3 - 37.
Lang, Mark H., and Russell J. Lundholm. 1996. Corporate Disclosure Policy and Analyst Behavior.
   The Accounting Review 71 (October): 467 - 492
Larcker, D.F., S. Richardson, and I. Tuna. 2004. Does Corporate Governance Really Matter?
    Working paper, University of Pennsylvania.
Leftwich, R. 1981. Evidence of the Impact of Mandatory Changes in Accounting Principles on
    Corporate Loan Agreements. Journal of Accounting and Economics 3 (March): 3 - 36.
Lev, B. 1988. Toward a Theory of Equitable and Efficient Accounting Policy. The Accounting
   Review 63 (January): 1 - 21.
Lev, B. 1992. Information Disclosure Strategy. California Management Review 34 (Summer): 9 - 32.
Lev, B. 2003a. Corporate Earnings: Facts and Fiction. Journal of Economic Perspectives 17 (Spring):
   27 - 50.
Lev, B. 2003b. Testimony before the Committee on Energy and Commerce, U.S. House of
   Representatives, Washington D.C. (September 25): At
    http://pages.stern.nyu.edu/~blev/docs/Baruch%20Lev%20Testimony-9-25-2003.pdf

Levitt, A. 1998. The Numbers Game. Remarks Delivered at the NYU Center for Law and Business,
   New York, NY (September 28): At http://www.sec.gov/news/speech/speecharchive/1998/spch220.txt.
Levitt, A. 2000. Renewing the Covenant with Investors. Remarks Delivered at the NYU Center for
   Law and Business, New York, NY (May 10).
Lilien, Steven, and Victor Pastena. 1982. Determinants of Intra-Method Choice in the Oil and Gas
    Industry. Journal of Accounting and Economics 4 (1982): 145 -170.
Lys, T. 1984. Mandated Accounting Changes and Debt Covenants: The Case of Oil and Gas
    Accounting. Journal of Accounting and Economics 6 (April): 39 - 65.
                                                                                                       54

Malmquist, D.H. 1990. Efficient Contracting and the Choice of Accounting Method in the Oil and
   Gas Industry. Journal of Accounting and Economics 12 (January); 173 - 205.
Merchant, K.A. 1989. Rewarding Results: Motivating Profit Center Managers. Harvard Business Press,
   Boston, Massachusetts.

Miller, Edward M. 1977. Risk, Uncertainty, and Divergence of Opinion. The Journal of
   Finance 32, 4 (September): 1151-1168.
Miller, Merton H. and Franco Modigliani. 1966. Some Estimates of the Cost of Capital to the Electric
    Utility Industry, 1954-1957. The American Economic Review LVI (June): 333 - 391.
Morse, D., Jens Stephan, and Earl K. Stice. 1991. Earnings Announcements and the Convergence (or
   Divergence) of Beliefs. The Accounting Review 66, 2 (April) 376-388.
Mulford, C.W., and E.E. Comiskey. 2002. The Financial Numbers Game: Detecting Creative
   Accounting Practices. John Wiley & Sons, New York.
Muller III, Karl A., and Edward J. Riedl. 2002. External Monitoring of Property Appraisal Estimates
   and Information Asymmetry. Journal of Accounting Research 40 (June): 865 - 881.
Myers, S.C. and N.S. Majluf. 1984. Corporate Financing and Investment Decisions when Firms Have
   Information that Investors Do Not Have. Journal of Financial Economics 13 (June): 187-221.
Myers, J. N., L. A. Myers, Zoe-Vonna Palmrose, and Susan Scholz. 2003. Mandatory Auditor
   Rotation: Evidence from Restatements. Working paper, University of Illinois at Urbana-
   Champaign.
Nwaeze, E.T. 2005. Replacement versus Adaptation Investments and Equity Value. Journal of
  Corporate Finance 11 (June): 523 - 549.
O’Brien, P., and R. Bhushan, , 1990. Analyst Following and Institutional Ownership. Journal of
   Accounting Research 28 (Supplement): 55 - 82.
Palmrose, Zoe-Vonna. 1988. An Analysis of Auditor Litigation and Audit Service Quality. The
   Accounting Review 63 (January): 55 - 73.
Palmrose, Zoe-Vonna, Scott Richardson, and Susan W. Scholz. 2004. Determinants of Market
   Reactions to Restatement Announcements. Journal of Accounting and Economics 37 (February):
   59 - 89.
Palmrose, Zoe-Vonna, and Susan W. Scholz. 2004. The Circumstances and Legal Consequences of
   Non- GAAP Reporting: Evidence from Restatements. Contemporary Accounting Research 21
   (Spring): 139 - 190.
Phillips, John, Morton Pincus, and Sonja Olhoft Rego. 2003. Earnings Management: New Evidence
    Based on Deferred Tax Expense. The Accounting Review 78 (April): 491 - 521.
Potter, G., 1992. Accounting Earnings Announcements, Institutional Investor Concentration, and
    Common Stock Returns. Journal of Accounting Research 30 (Spring): 146 - 155.
Press, E.G., and J. Weintrop. 1990. Accounting-Based Constraints in Public and Private Debt
    Agreements: Their Association with Leverage and Importance in Accounting Choice. Journal of
    Accounting and Economics 12 (January): 65 - 96.
                                                                                                      55

Puffer, Sheila M., and Joseph B. Weintrop. 1991. Corporate Performance and CEO Turnover: The
    Role of Performance Expectations. Administrative Science Quarterly 36 (March): 1 - 19.
Richardson, S., I. Tuna, and M. Wu. 2003. Capital Market Pressures and Earnings Management: The
   Case of Earnings Restatements. Working paper, University of Pennsylvania, 2003.
Rosner, R.L. 2003. Earnings Manipulation in Failing Firms. Contemporary Accounting Research 20
   (Summer): 361 - 408.
Skinner, D. 1993. The investment opportunity set and accounting procedure choice. Journal of
    Accounting and Economics 16 (October): 407- 46.
Smith, C. and R. Watts. 1992. The investment opportunity set and corporate financing, dividend, and
   compensation policies. Journal of Financial Economics 33 (December): 263 - 292.
Srinivasan, Suraj. 2005. Consequences of Financial Reporting Failure for Outside Directors:
    Evidence from Accounting Restatements and Audit Committee Members. Journal of Accounting
    Research 43 (May): 291 - 334.
Stuart, James B. 2002. Common Sense: Has Enron Fever Got You Down? Learn to Stay Cool --- The
    Market Hubbub Offers Opportunity. The Wall Street Journal, Europe, February 8, 27.
Sweeny, A.P. 1994. Debt-Covenant Violations and Managers’ Accounting Responses. Journal of
   Accounting and Economics (May): 281 - 308.
Warfield, T. D., J.J. Wild, and K.L. Wild. 1995. Managerial Ownership, Accounting Choices, and
   Informativeness of Earnings. Journal of Accounting and Economics 20 (July): 61 - 91.
Warner, J.B., R.L. Watts, and K.H. Wruck. 1988. Stock Prices and Top Management Changes.
   Journal of Accounting and Economics 20 (January/March): 461 - 492.
Watts, R.L., and J.L. Zimmerman. 1978. Towards a Positive Theory of the Determination of
   Accounting Standards. The Accounting Review 53 (January): 112 - 134.
Watts, R.L., and J.L. Zimmerman. 1986. Positive Accounting Theory. Prentice-Hall, Englewood
   Cliffs, New Jersey.
Watts, R.L., and J.L. Zimmerman. 1990. Positive Accounting Theory: A Ten Year Perspective. The
   Accounting Review 53 (January): 131 - 156 .
Wayman, Rick. 2003. Operating Cash Flow: Better than Net Income? Investopedia.Com (December
  22): At http://www.investopedia.com/articles/analyst/03/122203.asp.
Weisbach, M. 1988. Outside Directors and CEO Turnover. Journal of Financial Economics 20
   (January/March): 431 - 460.
Zmijewski, M.E., and R.L. Hagerman. 1981. An Income Strategy Approach to the Positive Theory of
   Accounting Standard Setting/Choice. Journal of Accounting and Economics 3 (August): 129-149.

								
To top