Datta Et Al 2013

Download as pdf or txt
Download as pdf or txt
You are on page 1of 13

Journal of Banking & Finance xxx (2013) xxx–xxx

Contents lists available at SciVerse ScienceDirect

Journal of Banking & Finance


journal homepage: www.elsevier.com/locate/jbf

Product market power, industry structure, and corporate earnings


management
Sudip Datta a,⇑, Mai Iskandar-Datta a, Vivek Singh b
a
Department of Finance, School of Business Administration, Wayne State University, 5201 Cass Avenue, Detroit, MI 48202, United States
b
Department of Finance and Accounting, College of Business, University of Michigan-Dearborn, Dearborn, MI 48126, United States

a r t i c l e i n f o a b s t r a c t

Article history: This is the first study to establish a link between product market power of firms and the degree of earn-
Received 25 July 2012 ings management. We hypothesize and document a significant and robust association between (a) a
Accepted 22 March 2013 firm’s product market pricing power and its degree of earnings management, and (b) industry competi-
Available online xxxx
tiveness and the degree of earnings management in the industry. Our study reveals that firms with infe-
rior product market pricing power engage in greater discretionary earnings accruals, adding a new
JEL classification: dimension to our understanding of the transparency and informativeness of firms’ financial statements.
G30
These findings are mirrored at the industry level where we document that more competitive industries
L11
M4
are associated with greater earnings manipulation. The empirical evidence has direct implication on the
M41 informativeness and earnings quality of firms based on their product market power and competitiveness.
Ó 2013 Elsevier B.V. All rights reserved.
Keywords:
Product market power
Industry structure and competition
Earnings management
Discretionary accruals management

1. Introduction to the value of further research to explain how business factors


drive accruals.’’
The idea that competitive pressure is an important determinant In this study we argue that a firm’s pricing power has the poten-
of managerial decision-making has received empirical support in tial to influence the degree of earnings management for a number
the literature. A number of studies establish that a firm’s product of reasons. First, pricing power can serve as a cushioning mecha-
market environment influences its investments, financing, cash nism that affords the firm the ability to pass on any cost shocks
distributions, corporate governance, analysts’ earnings forecasts, to the customers, reducing cash flow fluctuations, and thereby
and hedging decisions (see e.g., Akdogu and MacKay, 2012; Datta diminishing the need to manipulate earnings. Second, in light of
et al., 2011; Haushalter et al., 2006; Grullon and Michaely, 2007; the fact that the market punishes firms when they fail to meet
Fama, 1980). Yet, how product market power impacts the strategic earnings expectations, one can argue that firms under weak pricing
decision to manage a firm’s reported earnings is an issue that has power are more likely to manipulate earnings to meet market
largely been overlooked. A central issue in earnings management expectation. Another motivation for earnings manipulation is to
research is to identify which firms have a propensity to engage strategically limit and obfuscate the information available to rivals
in earnings manipulation. Much of the literature on earnings man- in an attempt to maintain a competitive advantage. Therefore,
agement delves into the degree to which firms are able to ‘‘game’’ firms facing greater competitive pressures may be motivated to
the capital markets through earnings manipulation. However, we manage earnings to limit information available to their rivals. Till
know very little about firm attributes that drive earnings manage- date, the potentially important link between product market pric-
ment. In fact, Healy and Wahlen (1999) state ‘‘These studies point ing power of firms and earnings management remains unexplored.
The primary focus of this study is to address whether product mar-
⇑ Corresponding author. Address: Department of Finance – Prentis 216, School of ket power influences the degree of earnings manipulation by cor-
Business Administration, Wayne State University, 5201 Cass Avenue, Detroit, MI porate managers.
48202, United States. Tel.: +1 313 577 0408; fax: +1 313 577 0058. Specifically, we seek to answer the following questions: Is there
E-mail addresses: sdatta@wayne.edu (S. Datta), mdatta@wayne.edu (M. Iskan- a link between product market pricing power and the transparency
dar-Datta), vatsmala@umich.edu (V. Singh).

0378-4266/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jbankfin.2013.03.012

Please cite this article in press as: Datta, S., et al. Product market power, industry structure, and corporate earnings management. J. Bank Finance (2013),
http://dx.doi.org/10.1016/j.jbankfin.2013.03.012
2 S. Datta et al. / Journal of Banking & Finance xxx (2013) xxx–xxx

of firm’s reported earnings? Put differently, does pricing power The degree to which firms manipulate earnings has ramifications
that affords firms the ability to pass on cost shocks to customers for the informativeness of their reported financial statements.
lead to less earnings manipulation? Or, does lack of pricing pres- A survey of CFOs by Graham et al. (2005) documents that earn-
sure on firms that enjoy high product market power exacerbates ings management is pervasive. They report that a vast majority of
earnings management? How does corporate governance influence managers admit to smoothing earnings via manipulation of real or
the managerial decision to manage earnings in light of its product accrual activities to influence the stock price and firm’s risk pre-
market pricing power? How does industry competitiveness influ- mium. Skinner and Sloan (2002) show that managers manipulate
ence corporate earnings management? earnings to avoid revealing the true value of their firms because
This study contributes to the finance and accounting literatures reporting lower than expected earnings is severely penalized by
by documenting how a firm’s relative product market pricing financial markets.
power and industry competitiveness determine earnings manage- Researchers have addressed various aspects of discretionary
ment decisions. Based on a comprehensive sample of 43,628 firm- earnings management. A prominent focus of this body of work is
year observations during the period spanning 1987–2009, our on capital markets based incentives to manipulate earnings, such
study documents that product market power is an important as boosting stock prices (Collins and Hribar, 2000), and obtaining
determinant of corporate earnings transparency. Notably, our anal- lower financing costs (Dechow et al., 1996). Past studies have also
ysis shows that firms with inferior product market pricing power documented that managing reported earnings is intended to influ-
engage more in discretionary accruals management, which sug- ence the decisions of external capital providers. In particular, some
gests that such firms strategically act to limit transparency in their research reports that managers inflate earnings prior to seasoned
financial reporting. This validates the notion that the ability of high equity offerings, initial public offerings, and stock-financed acquisi-
market power firms to pass on cost shocks to customers reduces tions (Adams et al., 2009; Teoh et al., 1998; Rangan, 1998; Erickson
the need to use accruals manipulation. and Wang, 1999) and to meet regulatory requirements (Yu et al.,
We extend our analysis by also examining the link between dif- 2006) while earnings are managed downward prior to management
ferent measures of industry structure and earnings management. buyouts (Perry and Williams, 1994). In contrast to this strand of re-
Using three alternative proxies of industry competition, our results search that focuses on managerial discretion in reported earnings
indicate that the greater the competition in an industry, the greater around a certain event, the focal point in this study is distinctly dif-
the earnings management indicating that a lack of competitive ferent because it addresses whether the firm’s product market
environment diminishes the need to engage in earnings manipula- power is a major driving force behind earnings management.
tion. Moreover, all our findings are highly robust to controlling for
(a) internal and external disciplinary governance mechanisms, (b)
executive compensation, and (c) the firm’s information environ- 2.2. Hypotheses development
ment, indicating that governance factors cannot be considered sub-
stitutes for product market pricing power or competitiveness in an There are a number of arguments that suggest a potential link
industry. between product market power and earnings management. Intra-
Our findings contribute to the literature on financial disclosure industry pricing power (which we interchangeably refer to as
and the disciplinary effect of competition. Specifically, our results product market power) emanates from the firm’s ability to extract
provide empirical evidence in support of Verrecchia’s (1983) model abnormal rents (higher prices) from its customers with little im-
that competitive pressure reduces information disclosure. In line pact on demand, thus conferring a competitive pricing edge to
with predictions by Shleifer (2004) and Rotemberg and Scharfstein the firm. The preceding argument does not require us to make
(1990), the positive association between competition and earnings the assumption of a perfectly inelastic demand curve for it to hold.
management supports the notion that managerial career concerns Uniqueness and superiority of product lines or a strong brand
arising from operating in an intense competitive environment name are the hallmarks of strong pricing power and competitive
pressures managers to manage earnings. The results do not sup- advantage. While industry-wide elasticity of demand is deter-
port the view that competition serves as a disciplinary mechanism mined by the aggregate demand curve for the industry, intra-
by providing more information transparency. industry product differentiation (among firms within the industry)
To the extent that earnings management can distort the finan- can affect the price elasticity of demand faced by a specific firm,
cial picture of the firm, our analysis should help investors better regardless of the industry structure in which it operates.
understand the association between a firm’s product market power Pricing power confers a number of advantages on the firm. For
and the degree of earnings manipulation, thereby enhancing their example, firms with greater pricing power can better maintain
ability to gauge the real numbers behind the reported earnings. their profit margins when they are subject to exogenous productiv-
The remainder of the paper is structured as follows. Section 2 ity shocks because of the uniqueness of their products and/or
presents the background literature and formulates the hypotheses. strong brand name. Greater product differentiation (or lower prod-
Sample formation, measurement of product market power, indus- uct substitutability) can lead to more inelastic demand curve for a
try-level competition measures, and description of the sample are firm’s products, affording it the flexibility to pass on cost shocks to
presented in Section 3. Section 4 presents the empirical findings. its customers.1,2 In Gaspar and Massa (2006), product market power
Section 5 concludes.
1
Price competition increases with increased product substitutability (see Salop,
1979).
2
2. Background literature and hypotheses development To see how cost shocks impact firms within a specific industry, consider two
firms: Firm A has differentiated its products from the rest of the industry and is able
to command higher prices and Firm B has undifferentiated products (i.e. experiences
2.1. Background literature greater product substitutability) and hence a decreased ability to raise prices. Firm A
exhibits greater product market power than Firm B by virtue of its ability to extract a
Financial reporting is a key source of information to capital mar- greater price from its customers in the face of an idiosyncratic cost shock. In other
kets. Opportunistic earnings manipulation subverts the purpose of words, it enjoys less price competition. Thus, the inability of firms with weak pricing
power to pass on cost shocks and protect profit margins serves to exert more pressure
financial reporting by distorting firm’s true economic performance, on them to manage earnings. In contrast, greater product market power enables firms
and thus can act as a hindrance to the full flow of information to to reduce the uncertainty about their future cash flows without resorting to earnings
market participants leading to higher informational asymmetry. management.

Please cite this article in press as: Datta, S., et al. Product market power, industry structure, and corporate earnings management. J. Bank Finance (2013),
http://dx.doi.org/10.1016/j.jbankfin.2013.03.012
S. Datta et al. / Journal of Banking & Finance xxx (2013) xxx–xxx 3

acts as a natural hedge that firms employ to smooth out firm-specific cause the cushion of supranormal profits in the form of Ricardian
fluctuations. They predict a negative correlation between market rents for typical firms in this industry is expected to be corre-
power and firm specific volatility. Gaspar and Massa (2006) also pro- spondingly smaller (see Peteraf, 1993). Therefore, along the lines
vide a mathematical solution proving that the change in profit of our preceding argument at the firm level, we correspondingly
caused by a cost shock negatively impacts lower market power.3 argue that, at the industry level, firms in more competitive indus-
It is important to note that financial analysts consider product tries are also expected to resort to greater degree of earnings man-
market power a critical factor in their evaluation of a firm’s pros- agement because they are less able to pass along an adverse cost
pects. Warren Buffet notably stated, ‘‘[T]he single most important shock to the consumers in that industry.
decision in evaluating a business is pricing power. If you’ve got There are two other important channels through which compet-
the power to raise prices without losing business to a competitor, itive pressure can influence the transparency of reported earn-
you’ve got a very good business. And if you have to have a prayer ings—namely, (a) information disclosure channel, and (b) the
session before raising the price by 10%, then you’ve got a terrible disciplinary effect channel. The disclosure decision by the firm de-
business.’’4 There is also empirical evidence in support of the view pends on the costs and benefits of revealing underlying informa-
that financial analysts earnings forecasts are more accurate (Datta tion about the firm’s prospects. A number of theoretical models
et al., 2011) and exhibit less dispersion (Gaspar and Massa, 2006) predict that firms in industries characterized by intense competi-
for firms with greater product market power (see also Hoberg and tive will opt to report less useful information (Gertner et al.,
Phillips, 2010). These findings are attributed to the price setting abil- 1988; Verrecchia, 1983). Verrecchia (1983) proposes that due to
ity of such firms and firms’ lower variability in demand, revenues the adverse impact from disclosure, firms in industries character-
and cash flows. ized by intense product market competition prefer less informative
A strong product position also endows the firm with superior disclosure policies to reduce predatory threats from rivals. Further,
staying power—in other words, the firm has greater flexibility Fan and Wong (2002) show that limiting the information flow al-
when responding to unexpected changes in consumer product lows insiders to sidestep the competition, while Verrecchia and
needs. Because of the ability to set prices, strong market power is Weber (2006) find empirically that disclosure of proprietary infor-
associated with more stable cash flows and lower stock return vol- mation is less in competitive industries in support of the view that
atility (Peress, 2010). Pricing power gives these firms deeper pock- product market competition and disclosure are negatively related.5
ets, allowing them to maintain their superior positions. The Thus, it can be argued that competition can induce more opaqueness
enhanced immunity of these firms against cash shortfalls, in- in earnings in order to restrict the information revealed to rivals. For
creases their capability to face deteriorating economic conditions instance, firms with higher demand may strategically determine that
and result in lower likelihood of distress vis-à-vis firms with weak it is best to withhold such information through earnings manage-
pricing power. Schmidt (1997) contends that enhanced competi- ment because by doing so they hold back the signal of good future
tion increases the threat of liquidation of the firm, thus providing prospects from rival firms.
strong incentives for managers to work harder to retain their jobs. However, theoretical models and empirical evidence on infor-
Thus, intense competition may induce the managers to manipulate mation disclosure by competing firms offer divergent views. In
earnings more aggressively to ward off the threat of liquidation. In contrast to Verrecchia (1983), Gal-Or’s (1985) model predicts the
contrast, the higher financial flexibility of firms with strong prod- opposite – that firms in concentrated industries will disclose less
uct market positions suggests that the pressures on managers to information. Stivers (2004) advances a similar argument that there
engage in earnings management would be less pronounced. is a greater likelihood of unraveling of proprietary information in
Studies have also documented other benefits of strong market competitive settings. He postulates that when an industry contains
power, such as the greater informativeness of stock prices (Peress, a large number of rivals, it is more likely to have at least one high
2010) and higher stock liquidity because investors channel more quality firm willing to reveal its information. This unraveling of
capital when they are better informed (Kale and Loon, 2011). The information ensures full disclosure. There is also empirical evi-
greater informativeness of stock prices combined with higher dence documenting that more concentrated sectors are less likely
liquidity imply that the earnings of firms with stronger market po- to report segment information than highly competitive industries
sition exhibit more stability, hence, a lesser need to manipulate (Harris, 1998; Hayes and Lundholm, 1996).
earnings. The literature also argues that firms compete for limited funds
Based on the above arguments that indicate multiple benefits from the public capital markets. Diamond and Verrecchia (1991)
enjoyed by firms with strong product market position, we reason contend that increased disclosure will reduce information asym-
that firms with weak product market power are more likely to re- metry, and as a result, lower the firm’s cost of capital. This argu-
sort to earnings manipulations. Hence, we contend that product ment suggests that as the number of firms in an industry
market power and earnings management can be viewed as substi- competing for limited capital increases, the informational trans-
tutes. This leads us to our first hypothesis. parency increases. Hoberg and Phillips (2010) offer another ratio-
nale as to why firms in more competitive environments may
Hypothesis 1. Firms with greater product market pricing power choose to disclose more proprietary information—namely, the
relative to other firms in a specific industry will be associated with need to reduce informational asymmetry in order to obtain financ-
lower discretionary accruals management. ing at more favorable rates. They document evidence in support of
the view that gathering firm-specific information in less competi-
tive setting is less costly. The mixed empirical evidence and the
We also examine structural industry competition and its influ-
contrasting predictions in the information disclosure literature
ence on discretionary accruals. It can be argued that greater prod-
raise interesting empirical question on how competition may im-
uct market competition in an industry, as measured by the
pact the transparency of financial statements.
industry structure, will lead to more earnings management be-
The disciplinary effect of competition is another channel
through which earnings management can be impacted. Product
3
See footnotes #2 and #3 in Gaspar and Massa (2006).
4
See ‘‘Buffett Says Pricing Power More Important Than Good Management.’’
5
by Andrew Frye and Dakin Campbell (Bloomberg.com February 18, 2011). See http:// Other studies provide models showing that under certain conditions, firms will
www.bloomberg.com/news/2011-02-18/buffett-says-pricing-power-more-impor- bias their disclosure of financial information because of competitive motives (Bagnoli
tant-than-good-management.html. and Watts, 2007).

Please cite this article in press as: Datta, S., et al. Product market power, industry structure, and corporate earnings management. J. Bank Finance (2013),
http://dx.doi.org/10.1016/j.jbankfin.2013.03.012
4 S. Datta et al. / Journal of Banking & Finance xxx (2013) xxx–xxx

market competition has been shown to serve as an external disci- Hypothesis 2B. Firms facing greater competitive pressure from
plinary corporate governance mechanism aligning the interests of product markets will be associated with lesser degree of discre-
the managers with shareholders and enhancing efficiency (Hart, tionary accruals management.
1983; Grullon and Michaely, 2007; Shleifer and Vishny, 1997). The-
oretical studies have shown that competition can directly influence
managerial behavior, thereby mitigating agency problems. In 3. Sample and measurement of variables
Hart’s (1983) model, higher competition leaves managers with less
slack, motivating them to work harder. Similarly, Holmström 3.1. Sample formation
(1982) and Nalebuff and Stiglitz (1983) postulate that increased
competition provides owners with more information that can mit- Our sample selection process starts by including all firms in the
igate moral hazard problem.6 COMPUSTAT database during the period 1987–2009. The begin-
Recent empirical evidence supports the idea that product mar- ning of our sample period is determined by the availability of the
ket competition provides incentives for managers to be more clo- key variable, cash flow from operations, needed to estimate accru-
sely aligned with shareholders’ interests (e.g. Giroud and Mueller, als. We also require that the sample firms be covered in the Center
2011). Guadalupe and Pérez-González (2010) show the greater for Research in Securities Prices (CRSP) monthly files, trade on the
the intensity of product market competition, the less the private NYSE, AMEX, or NASDAQ exchanges, and whose securities corre-
benefits of managerial control.7 Allen and Gale (2000) conclude spond to common equity (CRSP share code between 10 and 19).
that competition between firms is a more effective disciplinary We drop firms that changed their fiscal year-end during the sample
mechanism than either internal governance mechanisms or exter- period and confine our analysis to firms based in the US. To remove
nal monitoring mechanisms such as the market for corporate the effect of small firms, we only include firms that have at least $1
control. million in sales and assets. We define each firm’s industry based on
A number of studies examine the linkage between the disciplin- Fama–French 49 industry classification. Financial firms and utili-
ary environment and earnings management. Specifically, these ties are eliminated from the sample. Finally, we delete all firm-
studies show that a weak disciplinary environment allows manag- years with inasufficient data to calculate discretionary accruals
ers to engage in more earnings manipulation (see Becker et al., (as defined below) or any of the variables needed to estimate the
1998; Bowen et al., 2008; Guidry et al., 1999). Klein (2002) finds cross-sectional modified Jones model with Kothari et al.’s (2005)
that strong internal governance is associated with less pronounced adjustment for firm performance. The above selection criteria yield
abnormal accruals, while Becker et al. (1998) document the mon- a maximum sample of 43,628 firm-year observations representing
itoring effect of auditors by showing that clients of non-Big 5 audi- 6019 unique firms.
tors have a greater propensity to engage in discretionary accruals,
especially income increasing accruals, relative to clients of Big 5 3.2. Measuring product market power and industry-level competition
auditors.
However, this view of competition as a disciplining mechanism 3.2.1. Intra-industry measure of market power
is not unanimous. An alternative view is that product market com- 3.2.1.1. Firm-specific product market pricing power. Following much
petition could lead to pressures to boost short-term performance. of the industrial organization literature (e.g. Lindenberg and Ross,
For example, Shleifer (2004) posits that the pressure from intense 1981 and Domowitz et al., 1986), we construct our product market
competitive environment may provide managers with greater pricing power measure based on the Lerner Index (LI) (see Lerner,
incentives to manipulate earnings to influence the stock price. 1934) which is also referred to as the price–cost margin scaled by
According to this argument the more competitive the product mar- sales. It is calculated as follows:
ket, the greater the manager’s career concerns, thereby resulting in Sales  COGS  SG & A
opportunistic and myopic managerial behavior (e.g. Fama, 1980; PCM ¼ LI ¼ ð1Þ
Sales
Narayanan, 1985; Rotemberg and Scharfstein, 1990; Hermalin
and Weisbach, 2012). In a related study, Karuna (2007) shows that where Sales is COMPUSTAT variable SALE, cost of goods sold, COGS, is
firms in more competitive industries monitor their CEOs more clo- COMPUSTAT variable COGS, and sales, general and administrative ex-
sely than in less competitive industries which exacerbates the ca- penses, SG&A, is COMPUSTAT variable XSGA. This measure excludes
reer concern problem. depreciation, interest, special items and taxes. We use operating in-
Other studies argue that competitive pressure may aggravate come before depreciation (COMPUSTAT variable OIBDP) to calculate
the moral hazard problem (Milgrom and Roberts, 1992), increase price–cost margin when there is missing data for the above items.
managerial shirking (Scharfstein, 1988), and raise the likelihood Although the price–cost margin has been used to capture a
of misreporting of accounting information (Rotemberg and Scharf- firm’s product market power, this measure does not, however, iso-
stein, 1990). Thus, both the information disclosure effect and the late the firm-specific factors that influence product market pricing
external disciplinary effect of competition provide conflicting power from industry-wide factors. This metric can fluctuate due to
views on how competition will impact managerial behavior. Based industry-specific attributes that are unrelated to a firm’s market
on the competing arguments discussed above, we propose the fol- pricing power. Given that we are interested in examining the link
lowing alternative hypotheses: between earnings manipulation and a firm’s product market power
within an industry, we use an industry-adjusted Lerner Index to
Hypothesis 2A. Firms in more competitive industries will be capture firm-specific product market power. To do so, we compute
associated with greater discretionary accruals management. the value-weighted industry-adjusted Lerner Index (Market Power),
which is the difference between the firm’s price–cost margin and
the sales-weighted price–cost margin of the all firms within an
industry and is described by the following equation.
6
For other work that examines the potential channels through which competition X
N
can have an effect managerial, see for example Aghion et al., 1999; Jagannathan and MarketPower ¼ LIi  xi LIi ð2Þ
Srinivasan, 1999; Raith, 2003). i¼1
7
Fama (1980), among others, posits that product market competition can be
instrumental in enhancing corporate governance since competition encourages where LIi is the Lerner Index (defined in Eq. (1) above) for firm i, xi
managers to avoid wasteful spending and misallocation of resources. is the proportion of sales of firm i to total industry sales where

Please cite this article in press as: Datta, S., et al. Product market power, industry structure, and corporate earnings management. J. Bank Finance (2013),
http://dx.doi.org/10.1016/j.jbankfin.2013.03.012
S. Datta et al. / Journal of Banking & Finance xxx (2013) xxx–xxx 5

industry is defined as the firm’s industry as per Fama–French 49 firms in the industry as an alternative proxy for industry competi-
Industry Classifications, and N is the total number of firms in the tion (see Cremers et al., 2008). Specifically we measure concentra-
industry (using the entire universe of firms in that industry avail- tion as the fraction of entire industry sales that is accounted for by
able in the COMPUSTAT database). This modified Lerner Index mea- the aggregate sales of the four largest firms in the industry. This
sure captures purely the intra-industry market power of a firm, measure, common in empirical industrial organization literature,
therefore purging the effects of industry-wide factors common to is routinely applied by government agencies. Because we are using
all firms in a specific industry. Further, this adjustment addresses three different proxies for industry competition that are unrelated
the fact that different industries have structurally different profit to HHI, our conclusions from the industry level analysis should not
margins due to factors unrelated to intra-industry differences in be affected by the issues surrounding HHI as a measure of
market power of the firms. We also emphasize that all our conclu- competition.
sions remain robust to the use of a normalized market power vari-
able, measured as the average of Market Power (industry-adjusted 3.2.3. Measurement of earnings management
Lerner Index, LIIA) over the preceding 3 years. It is also important To estimate accruals management, we have to distinguish be-
to note that because we are using Kothari et al.’s (2005) methodol- tween two types of accruals: non-discretionary accruals that are
ogy of calculating discretionary accruals that controls for firm per- indispensible accounting adjustments and discretionary accruals
formance (detailed below in the Section 3.2.3), the relationship made at the discretion of managers to manipulate earnings. Be-
between accruals and Market Power is unlikely to be mechanical. cause not all accruals are the result of opportunistic manipulation
by managers, we first have to estimate non-discretionary accruals
3.2.2. Industry-level measures of competition and extract it from total accruals to derive the discretionary
3.2.2.1. Industry lerner index. Following Cremers et al. (2008), we component.
use the industry median price–cost margin to capture industry Following previous studies we use the modified version of the
competitiveness. They argue that higher profit margins in an Jones model (see e.g., Jones, 1991; Dechow et al., 1995) to cap-
industry reflect less intense competitive environment and that thin ture discretionary accruals (DA). Hribar and Collins (2002) point
margins are associated with greater competitive pressures since out the concerns associated with estimating accruals using the
adverse input price shocks could not be passed through to custom- balance sheet approach. Therefore to estimate discretionary
ers through output price hikes. accruals more accurately, we use the cash flow data from the
statement of cash flow in COMPUSTAT. This is currently the ac-
3.2.2.2. Number of firms in the industry. One of the main factors that cepted and popular methodology for capturing accruals manage-
shape the intensity of rivalry in an industry is the number of firms ment in the literature (see e.g. Bartov et al., 2001; Kothari et al.,
in a sector (Porter, 1980) where larger number of firms in the 2006).
industry magnifies competition. Balakrishnan and Cohen (2011) This methodology derives discretionary accruals in two stages.
argue that since firms in an industry compete not only for eco- First, total accruals variable (defined as the difference between
nomic profits but also for funds from capital markets, the number net income and cash flows from operations) is regressed on key
of firms in an industry reflects competition for limited funds. In the variables that are expected to influence it. Specifically, we esti-
presence of greater competition, they posit that firms in highly mate nondiscretionary accruals from cross-sectional regressions
populated industries will provide higher quality of information, of total accruals (TACC) on changes in sales minus change in
and hence earnings management will be lower. An opposing argu- receivables, property, plant, and equipment (PPE), and lagged re-
ment based on evidence that managers inflate earnings prior to turn on assets (ROA) for each of the Fama–French industry classi-
seasoned equity offerings, initial public offerings and stock- fication in every fiscal year. We include lagged return on assets
financed acquisitions (Teoh et al., 1998; Rangan, 1998; Erickson (ROA) as an additional regressor to control for the effect of perfor-
and Wang, 1999) implies that the greater the competition for mance on a firm’s accruals (Kothari et al., 2005; Ronen and Yaari,
external funding the greater the earnings management. Hence, 2008). We run the following cross-sectional OLS regression using
the impact of number of firms in an industry on earnings manage- Fama–French industry code to estimate the coefficients a1, a2,
ment is an empirical issue. In line with Cremers et al. (2008), we and a3. These cross-sectional regressions require a minimum of
employ 1/n, where n is the number of firms in an industry, as an 15 observations for each year and Fama–French industry
alternative proxy for competition. combination.
 
3.2.2.3. Industry concentration. Industry concentration is typically TAit 1 DREV it DARit PPEit NetIncomeit1
¼ a1 þ  þ a3 þ a4 þ eit ð3Þ
used to measure competition for industry-level analysis (as op- Ait1 Ait1 Ait1 Ait1 Ait1 Ait1
posed to firm-level product pricing power). Industry concentration
where i indexes firms, t indexes time, TAit equals Net Income (COM-
is usually measured by Herfindahl–Hirschman Index (HHI).
PUSTAT variable NI) minus cash flow from operations (COMPUSTAT
Although HHI, as a measure of concentration, is fairly well rooted
variable OANCF), DREVit is the changes in sales (COMPUSTAT vari-
in industrial organization theory (Curry and George, 1983; Tirole,
able SALE), DARit is the change in Receivables (COMPUSTAT variable
1998, pp. 221–223), doubts exist that it could imply both high
RECT) and PPE is the total property, plant, and equipment (COMPU-
and low competition. Recent research suggests that when market
STAT variable PPEGT). All these variables are scaled by lagged value
structure is assumed to be endogenous, it is unclear whether low
of assets (COMPUSTAT variable AT). We use the estimated coeffi-
values of concentration capture low or high competition, especially
cients a
^1 ; a
^2 ; a
^ 3 , and a
^ 4 to compute discretionary accrual as follows:
in cross-industry analyses (e.g., Demsetz, 1973; Symeonidis, 2002;
  
Raith, 2003; Aghion et al., 2005). Therefore, a tension exists on the TAit 1 DREV it DARit
DAit  eit ¼  a ^1 þa^2 
topic of whether industry concentration (competition) is associ- Ait1 Ait1 Ait1 Ait1
ated with low or high degree of industry competition and how this 
PPEit NetIncomeit1
competition acts as an external governance mechanism in influ- þa ^3 þa ^4 ð4Þ
Ait1 Ait1
encing corporate earnings management.
Because of the recognized shortcomings associated with the Large values of discretionary accruals are generally construed to
traditional Herfindahl–Hirschman Index, we measure industry indicate earnings management. Because discretionary accruals
concentration considering the aggregate sales of the four largest could be positive (when firms inflate earnings) or negative (when

Please cite this article in press as: Datta, S., et al. Product market power, industry structure, and corporate earnings management. J. Bank Finance (2013),
http://dx.doi.org/10.1016/j.jbankfin.2013.03.012
6 S. Datta et al. / Journal of Banking & Finance xxx (2013) xxx–xxx

in good years managers conceal earnings for future use), both posi- capture distinct aspects of a firm’s competitive environment. In
tive and negative discretionary accruals capture earnings manage- line with prior research, we find that absolute discretionary accrual
ment. We winsorize the variables at 1% and 99% levels to reduce is positively correlated with Growth, Leverage, and Volatility of
the influence of outliers. It is important to note that because we sales.
are using Kothari et al. (2005) methodology of calculating discre-
tionary accruals, which controls for firm performance, the relation-
ship between accruals and Market Power is unlikely to be 4. Empirical findings
mechanical. Furthermore, we all our conclusions remain robust to
the use of a normalized market power variable, measured as the 4.1. Univariate analysis
average of Market Power (industry-adjusted Lerner Index) over the
preceding 3 years. In Table 3, we sort firms, in each fiscal year in quintiles on the
basis of market power and compute their average level of absolute
3.2.4. Sample description discretionary accrual. For each market power quintile, we then cal-
Table 1 presents several focus-relevant salient summary statis- culate the time-series mean and median absolute level of discre-
tics for our sample. We classify the descriptive statistics into four tionary accruals for the whole sample and then separately for
categories based on firm characteristics, product market power positive and negative absolute accruals. We also test for the mean
measures, industry competition measures and accruals manage- difference in accruals between the extreme quintiles.
ment metric. All the variables are defined in Appendix A. The sam- In support of our first hypothesis, these univariate results show
ple firms have a mean (median) market capitalization of $2,038 that the absolute level of discretionary accruals declines signifi-
million ($193 million). The median sample firm has a sales volatil- cantly as market power increases. Specifically, the median absolute
ity of 13.6%. level of discretionary accruals of 13.3 is the highest for the lowest
Next we present in Table 1 summary statistics for our measure market power group (quintile 1), while the accrual measure for
of product market power. The median LIIA for our sample is -3.44 firms with the highest market power (quintile 5) is almost half that
which is comparable to that reported by Gaspar and Massa’s amount at 7.3 of lagged assets. The difference between the mean
(2006) (5.8) for a different sample period. Regarding industry absolute levels of accruals for these two quintiles is highly statisti-
competition measures, the median of median industry Lerner In- cally significant (<1% level). The progression of accruals indicates
dex is 9.44% while the median number of firms in an industry is that it is monotonically decreasing in firm’s market power. The re-
about 79. The median market share of the four largest firms in sult is robust to using the median values or a market power metric
an industry is 5.17%. Finally, the median absolute level of discre- that is not industry adjusted. All our conclusions remain robust to
tionary accruals for our sample is 5.5% of lagged assets, which is the use of a normalized market power variable, measured as the
similar to prior studies. average of Market Power (industry-adjusted Lerner Index) over the
Table 2 presents the Pearson correlations between the variables preceding 3 years. Overall, the findings are consistent with the no-
used in the analysis. The correlations between firm characteristic tion that lack of market power makes it more likely that a firm’s
control variables (Leverage, Size, Book-to-market, Growth and Vola- management will engage in earnings management, whereas firms
tility of sales), although significant, tend to be small in magnitude. that enjoy a more powerful pricing power in their product market
One notable observation regarding the relationship between the are less likely to do so.
various test variables on product market pricing power and indus- In the remaining columns of Table 3 we provide the absolute
try competition is the lack of significant correlation between Mar- positive and negative accruals separately. The results confirm the
ket Power and Concentration indicating that they capture two findings from the total sample indicating that firms with greater
different aspects of competition – one at the firm-level and the market power engage in less earnings manipulation whether it is
other at the industry level. This validates our premise that intra- upward or downward management and the differences between
firm product market pricing power and industry concentration the two extreme quintiles are statistically significant.

Table 1
Descriptive statistics. This table reports summary statistics for some salient characteristics of our sample. We have categorized the descriptive statistics into four groups: firm
characteristics, market power measure, industry completion measures, and accruals management metrics. The statistics are based on a maximum of 43,628firm-year
observations drawn from the intersection of the COMPUSTAT and CRSP databases spanning the period 1987–2009 for firms meeting our data requirements. These represent 6,019
unique firms spanning 36 industries based on Fama–French 49 Industry classification. All other variables are as defined in Appendix A.

Variable Obs. Mean Median Std. dev.


Firm characteristics
Market capitalization (in $ millions) 43,628 2037.70 192.77 8510.46
Asset growth rate 43,628 17.20 6.43 57.39
Market-to-book ratio 43,628 3.19 1.98 11.95
Volatility of sales 39,675 0.211 0.136 0.244
Leverage 43,508 15.81 10.74 17.00
Institutional Holdings (%) 41,176 43.65 41.53 29.79
Number of Analysts 43,628 5.52 3.00 6.89
GIM Index 13,109 9.03 9.00 2.76
Average Bid-ask Spread 40,287 0.030 0.016 0.040
CEO Equity-based compensation (EBC) 9,535 41.62 43.10 29.71
Market power measure
Market Power (%) 40,258 12.20 3.44 37.65
Industry competition measure
Median Industry Lerner Index (%) 36 8.39 9.44 14.17
Concentration of 4 largest firms in industry sales (%) 36 7.05 5.17 5.86
Number of firms in an industry 36 100.79 78.75 73.78
Accruals management
Discretionary accruals/Assetst1 43,628 9.18 5.50 12.49

Please cite this article in press as: Datta, S., et al. Product market power, industry structure, and corporate earnings management. J. Bank Finance (2013),
http://dx.doi.org/10.1016/j.jbankfin.2013.03.012
S. Datta et al. / Journal of Banking & Finance xxx (2013) xxx–xxx 7

Table 2
Correlation matrix of salient variables. This table reports Pearson correlations of some salient focus-relevant characteristics of our sample. The statistics are based on maximum of
43,628 firm-year observations drawn from the intersection of the COMPUSTAT and CRSP databases spanning the period 1987–2009 for firms meeting our data requirements. See
Appendix A for variable definitions. Bolded correlations denote significance at the 1% level.

Variables Market Power Firms in industry Industry LI Concentration Volatility of Sales Growth Book-to-market Size Leverage
Market Power 1.000 0.170 0.445 0.010 0.028 0.030 0.050 0.094 0.159
Num Firms in industry 1.000 0.276 0.290 0.008 0.068 0.041 0.043 0.075
Industry LI 1.000 0.022 0.018 0.028 0.055 0.031 0.155
Concentration 1.000 0.031 0.020 0.013 0.013 0.023
Volatility of Sales 1.000 0.344 0.034 0.063 0.102
Growth 1.000 0.048 0.017 0.030
Book-to-market 1.000 0.034 0.003
Size 1.000 0.024
Leverage 1.000

Table 3
Market power and earnings management: univariate analysis. This table reports univariate analysis of the relationship between market power quintiles and absolute
discretionary accruals. The statistics are based on maximum of 43,628 firm-year spanning 1987–2009 for firms meeting our data requirements. The absolute level of discretionary
accruals is computed using modified Jones model with Kothari et al. (2005) adjustment for firm performance. We sort firms each year on the basis of Market Power variable and
compute the mean for each market power quintile. Then we take the time-series mean (median) for each market power quintile. The first (last) quintile represents firms with the
least (most) market power. Last row provides p-values for difference of mean between first and fifth quintiles.

Quintiles based on market power Total sample Positive absolute discretionary accruals Negative absolute discretionary accruals
Mean t-Stat Mean t-Stat Mean t-Stat
(Median) (Median) (Median)
First Quintile 0.133 18.66 0.114 15.90 0.146 18.22
(0.119) (0.098) (0.133)
Second Quintile 0.087 21.72 0.090 14.79 0.094 18.79
(0.079) (0.078) (0.080)
Third Quintile 0.078 30.07 0.080 23.82 0.072 20.36
(0.075) (0.073) (0.063)
Fourth Quintile 0.072 26.60 0.075 19.16 0.072 18.40
(0.067) (0.072) (0.063)
Fifth Quintile 0.073 25.95 0.070 18.45 0.075 23.97
(0.065) (0.066) (0.069)
p-Value for difference in Q1 and Q5 means <0.000 <0.000 <0.000

4.2. Multivariate analysis ment professionals as well as more political costs are less likely
to engage in accruals management. Size is the natural logarithmic
In this section we examine the relationship between absolute transformation of market capitalization. To calculate the Volatility
discretionary accruals scaled by lagged assets (Abs Disc Accruals) of sales measure, we obtain the standard deviation of sales from
and product market pricing power (Market Power) in a multivariate the preceding 3-year period scaled by 1 year lagged assets. For
setting, while controlling for the standard salient determinants of robustness, we also utilize two additional measures of volatility,
discretionary accruals using firm level characteristics previously namely, volatility of cashflows, and volatility of return on assets
identified in the literature, such as the growth rate in assets (ROA). The subscripts i and t refer to firm and year respectively.
(Growth), market-to-book ratio (Market-to-book) and volatility of Using a large sample of 43,628 firm-year observations we estab-
sales (Volatility), firm size (Size), and Leverage. We estimate various lish for the first time a strong link between product market power
configurations of the following model: of firms and their degree of discretionary accruals management.
This finding in the baseline regression, Model 1, is robust to alter-
Abs Disc Accrualsjt ¼ b0 þ b1 Market Power jt þ b2 Growthjt native specification that includes additional volatility measures
þ b3 Markettobookjt such as volatility of cash flows and volatility of return on assets.
For instance, the coefficient on this variable in Model 1 is
þ b4 Volatilityjt þ b5 Sizejt
0.451, which is significant at better than the 1% level. To gauge
þ b6 Lev eragejt þ ej ð5Þ the economic significance of this finding, we measure the impact
of one-standard deviation movement in the Market Power metric
We include year and industry dummies to control for business cycle from its median value and find that this increases discretionary
effects and differences across industries, respectively. All the stan- accruals by almost 14% of lagged assets. In unreported results,
dard errors in the regressions are clustered at the firm level. The we also find a strong association when the Market Power variable
regression estimates are presented in Table 4. is not industry-adjusted. Our results, which provide compelling
empirical evidence in support of Hypothesis 1, imply that firms
4.2.1. Control variables with low product market power place importance on the predict-
We control for standard firm characteristics. We include two ability of their earnings. This empirical finding also implies that
proxies for firm growth because such firms may face greater capital lack of external disciplinary forces (for firms with high market
market pressure to manipulate their earnings (Lee et al., 2006), power) does not encourage such firms to engage in greater manip-
namely, growth and market-to-book ratio. Growth is the change ulation of earnings; hence, these firms’ financial statements are
in assets scaled by 1 year lagged assets. Market-to-book is the ratio more transparent and stand to be more informative than their
of market capitalization to the book value of the firm. We control industry counterparts that have not been able to harness the pric-
for firm size since larger firms facing more scrutiny from invest- ing power.

Please cite this article in press as: Datta, S., et al. Product market power, industry structure, and corporate earnings management. J. Bank Finance (2013),
http://dx.doi.org/10.1016/j.jbankfin.2013.03.012
8 S. Datta et al. / Journal of Banking & Finance xxx (2013) xxx–xxx

Table 4
Market power and earnings management: absolute discretionary accruals. This table reports the results of OLS regressions examining the impact of market power on
discretionary accruals for a sample of firms spanning 1987–2009 that meet our data requirements. The dependent variable is the absolute level of discretionary accrual using
modified Jones model with Kothari et al. (2005) adjustment for firm performance. All models include year and industry dummies. All the variables are as defined in earlier tables.
p-Values reported in the parentheses are computed with standard errors adjusted for firm-level clustering.

Independent variables Model 1 Model 2 Model 3 Model 4 Model 5 Model 6


Market Power 0.451 0.043 0.045 0.040 0.339 0.449
(<0.000) (<0.000) (<0.000) (<0.000) (0.000) (<0.000)
Institutional Holdings 0.023
(<0.000)
Log (1 + Num of analyst) 0.109
(0.110)
GIM Index 0.023
(0.470)
CEO EBC 0.555
(0.166)
Spread 0.0532
(0.015)
Growth 4.225 4.290 4.219 3.287 2.966 4.342
(<0.000) (<0.000) (<0.000) (0.001) (0.000) (<0.000)
Log (Market-to-Book) 1.461 1.410 1.452 1.064 1.130 1.553
(<0.000) (<0.000) (<0.000) (<0.000) (<0.000) (<0.000)
Size 9.813 7.648 9.352 7.803 8.818 9.687
(<0.000) (<0.000) (<0.000) (<0.000) (<0.000) (<0.000)
Leverage 2.538 2.235 2.513 1.900 2.732 2.578
(<0.000) (<0.000) (<0.000) (0.004) (0.003) (<0.000)
Volatility of Sales 3.505 3.625 3.519 3.015 2.276 3.554
(<0.000) (<0.000) (<0.000) (0.000) (0.005) (<0.000)
Constant 0.071 0.089 0.073 0.079 0.048 0.066
(<0.000) (<0.000) (<0.000) (<0.000) (<0.000) (<0.000)
Year Dummy Yes Yes Yes Yes Yes Yes
Industry Dummy Yes Yes Yes Yes Yes Yes
Number of Obs. 39,563 37,388 39,563 13,013 9,282 39,408

4.2.1. Accounting for the influence of governance mechanisms firms. We follow the literature by assuming the last available value
In Models 2 through 6, we introduce different variables to con- when the index is between two updates. A high GIM Index repre-
trol for the possibility that various governance mechanisms may be sents lower shareholder rights. The coefficient for this variable is
behind the change in discretionary accruals. In Models 2 and 3, we insignificant while the result for our test variable, Market Power, re-
control for two external forms of monitoring such as institutional mains robust to the inclusion of this metric.8
holdings and analysts coverage, respectively. The private informa-
tion search activities and scrutiny by investment professionals, 4.2.2. Accounting for the influence of executive compensation
such as institutional investors and analysts serve to diminish infor- Some recent work has demonstrated empirically that high equi-
mational asymmetry as well as reduce the propensity of the firm to ty-based compensation can aggravate managerial incentives by
engage in earnings management. In support of that argument, Bal- encouraging earnings manipulation. For instance, studies have
akrishnan and Cohen (2011) find a negative relation between insti- shown that managers are found to manipulate earnings to influ-
tutional holdings and the frequency of earnings restatements, ence their bonuses (Guidry et al., 1999) and to gain from insider
while Yu (2008) documents a similar relationship between analyst sales of shares (Beneish and Vargus, 2002) while Bergstresser
coverage and accruals. and Philippon (2006) report that earnings management is more
Prior research has also established that institutional holdings pronounced in firms where a large fraction of CEO compensation
and the intensity of analysts’ following plays a key role in moni- is tied to stock and option holdings. We test the degree to which
toring the firm as well as in disseminating information (Lang and internal incentives affect the influence of competitive advantage
Lundholm, 1996). To the extent that the monitoring of invest- in the product market, by including the equity-based compensa-
ment professionals substitutes for competitive pressures, the rel- tion as a fraction of total CEO compensation package in Model 6.
evance of our Market Power variable may diminish in the Equity-based compensation (CEO EBC) is calculated as the sum of
presence of external monitoring from investment professionals. the value of stock options granted and restricted stockholdings of
The results in Models 2 and 3 show that the monitoring from lar- the CEO divided by total compensation. The coefficient estimate
ger institutional holdings significantly diminishes discretionary on CEO EBC although positive is insignificant. The findings indicate
accruals. Although greater analysts coverage of the firm also re- that our focus variable, Market Power, is robust to the inclusion of
duces the amount of earnings management, this reduction is this variable.
not statistically significant at conventional levels (p-value 0.11).
Importantly, in both cases, the significance of our Market Power 4.2.3. Accounting for the influence of the firm’s information
variable is maintained indicating that the effect of Market Power environment
is in addition to that from monitoring by institutional investors Previous work has established a negative link between informa-
and analysts. tion asymmetries and the firm’s bid-ask spread (Leuz and Verrec-
In Model 4 we control for internal governance by utilizing the chia, 2000). Firms with greater information asymmetries (higher
Gompers et al. (2003) GIM Index, which measures shareholder
rights based on 24 provisions. Because this index is updated infre- 8
In an unreported regression specification, we employ the E-Index proposed by
quently and it is available only for later years for part of our sample Bebchuk et al. (2004) that is based on 6 of the 24 provisions in the GIM index. Again,
period, our tests using this measure apply to a subset of our sample our results are robust to the inclusion of this external governance variable.

Please cite this article in press as: Datta, S., et al. Product market power, industry structure, and corporate earnings management. J. Bank Finance (2013),
http://dx.doi.org/10.1016/j.jbankfin.2013.03.012
S. Datta et al. / Journal of Banking & Finance xxx (2013) xxx–xxx 9

bid-ask spread) have greater incentives to manage earnings to re- imply the opposite. For example, Lee et al. (2010) find that large
duce the adverse impact of information asymmetries. In Model 7 negative DA and low profitability lead to more frequent CFO dis-
we control for firm’s information environment by incorporating missals than large negative DA and high profitability.
the bid-ask spread of the firm’s stock. Following the literature on Overall, the results presented in this section are consistent with
bid-ask spread, we compute Spread as the average of the daily clos- the notion that lack of product market pricing power makes it
ing (ask price-bid price)/closing price for each firm in each fiscal more likely that a firm’s managers will engage in earnings manage-
year. The positive Spread coefficient is significant with p-value of ment, whereas firms that enjoy a more powerful pricing power in
0.02. Again, the inclusion of this variable in Model 7 does not influ- their product market as a manifestation of their competitive
ence the relevance of our test variable, Market Power. advantage are less likely to do so. Our findings also support the
Overall, our firm characteristic control variables in the seven view that firms with strong product market positions are less likely
regression models presented in Table 4 consistently indicate that to manipulate reported earnings because of better abilities to bear
firms that utilize accruals management tend to be smaller firms, idiosyncratic cost shocks.
with less leverage and higher growth and volatility. Minton and
Schrand (1999) find that firms with higher cash flow volatility face 4.2.4. Multivariate results for positive and negative discretionary
higher costs of external financing. Thus, our finding of a positive accruals
relation between cash flow volatility and accrual implies that firms In Panels A and B of Table 5, we replicate the analysis of Table 4
with more costly external financing are more likely to manipulate after partitioning the firms that engage in positive and negative
reported earnings. Additionally, Maksimovic and Pichler (2001) ar- discretionary accruals separately to verify whether the impact of
gue that firms that utilize innovative new technologies incur larger product market power on accruals is symmetrical. The results in
disclosure costs. This reasoning implies that firms with high mar- Table 5 utilize the same firm characteristic control variables; how-
ket-to-book ratio are more concerned about divulging private ever, the coefficients on these control variables (which are similar
information through their financial statements and thus strategi- to those obtained earlier) are omitted from the table for parsimony.
cally manage earnings more. Our finding of positive and significant The coefficient estimate on Market Power in the baseline model,
coefficients on Market-to-Book confirms this conjecture. Model 1, is positive and significant whether the absolute discre-
Our method of calculating accruals, which adjusts for firm per- tionary accruals are positive (Panel A) or negative (Panel B) indicat-
formance, makes it unlikely that the relationship that we observe ing that firms with weaker product market positions are more
between accruals and the industry-adjusted Lerner Index (Market likely to engage in both types of earnings management.
Power) is mechanical. Further, a negative accrual does not neces- The coefficients in Model 2 from both panels reveal that institu-
sarily imply a low price–cost margin nor does a positive accrual tional holdings have a dampening effect on inflationary and defla-

Table 5
Market power and earnings management: Signed absolute discretionary accruals. This table reports the results of OLS regressions examining the impact of market power on
discretionary accruals for a sample of firms meeting our data requirements spanning 1987–2009. In Panels A and B, the dependent variables are respectively the positive and
negative absolute discretionary accrual using modified Jones model with Kothari et al. (2005) adjustment for firm performance. All models include year and industry dummies.
Firm characteristic control variables are suppressed for brevity. All the variables are as defined in Appendix A. p-Values reported in the parentheses are computed with standard
errors adjusted for firm-level clustering.

Independent Variables Model 1 Model 2 Model 3 Model 4 Model 5 Model 6


Panel A: Positive absolute discretionary accruals
Market Power 0.024 0.023 0.025 0.026 0.014 0.024
(<0.000) (<0.000) (<0.000) (0.002) (0.050) (<0.000)
Institutional Holdings 0.001
(<0.000)
Log (1 + Num. of analyst) 0.007
(<0.000)
GIM Index 0.001
(0.079)
CEO EBC 0.0005
(0.460)
Spread 0.082
(0.005)
Firm characteristic controls Yes Yes Yes Yes Yes Yes
Year and Industry dummies Yes Yes Yes Yes Yes Yes
Number of Obs. 19,154 18,058 19,154 5,820 4,209 17,648

Panel B: Negative absolute discretionary accruals


Market Power 0.067 0.064 0.067 0.057 0.055 0.066
(<0.000) (<0.000) (<0.000) (0.002) (0.000) (<0.000)
Institutional Holdings 0.001
(0.000)
Log (1 + Num. of analyst) 0.004
(0.000)
GIM Index 0.001
(0.059)
CEO EBC 0.007
(0.179)
Spread 0.033
(0.253)
Firm characteristic controls Yes Yes Yes Yes Yes Yes
Year and Industry dummies Yes Yes Yes Yes Yes Yes
Number of Obs. 20,409 19,330 20,409 7,193 5,073 21,760

Please cite this article in press as: Datta, S., et al. Product market power, industry structure, and corporate earnings management. J. Bank Finance (2013),
http://dx.doi.org/10.1016/j.jbankfin.2013.03.012
10 S. Datta et al. / Journal of Banking & Finance xxx (2013) xxx–xxx

tionary earnings management. Interestingly, in Model 3, the influ- and conclusions remain robust to the use of the normalized vari-
ence of analysts’ coverage on managerial behavior is asymmetrical able. Further, it is worth noting that when we employ alternative
where higher analyst following diminishes the engagement in po- measures to capture competitiveness at the industry level (in the
sitive accruals while increasing negative accruals. Akin to findings next section), our results mirror what we obtain at the firm level.
from Table 3, here also the significance of the Market Power vari- This confirms that our results are not driven by possible mechani-
able remains robust to the inclusion of institutional holdings and cal connection between our measure of pricing power and discre-
analysts’ coverage. tionary accruals. To account for conglomerates, we repeat all our
The results for Model 4, which controls for managerial analyses with and without multi-segment firms and find that all
entrenchment using GIM Index, also indicate asymmetry in mana- our results are robust to this distinction. In another specification,
gerial behavior regarding earnings management. The greater man- when we include a dummy variable to control for multi-segment
agerial entrenchment (higher GIM Index), the more likely the firm firms, the results are qualitatively the same. Finally, as an addi-
to inflate earnings (as shown in Panel A) and less likely to deflate tional (untabulated) robustness check, for Tables 4 and 5, we rees-
earnings (Panel B). Both of these associations are statistically sig- timate the regressions using firm-fixed effects, instead of industry
nificant. Confirming earlier results for the whole sample, in Model dummies, and find that the Market Power variable remains highly
5, the internal incentives from CEO equity pay do not play a role in significant with p-values < 0.000 in all models in those two tables.
earnings management. In the last model (Model 6), which includes
Spread, the coefficient is significantly positive only in Panel A indi- 4.3. Multivariate analysis of the influence of industry competition on
cating that the greater the information asymmetry, the more likely earnings management
managers will inflate earnings. In all models, our focus test variable
is robust to the inclusion of these external and internal governance In this study we have underscored and articulated the distinc-
mechanisms in support of the view that none of these governance tion between firm-level pricing power vis-à-vis the industry rivals
factors can be considered substitutes for product market pricing and industry-level competitiveness. Consequently, we have also
power at the firm level within an industry. identified the appropriate measures to capture these different
dimensions of competition at the firm-level and for the industry.
4.2.5. Additional robustness checks In Table 6 we examine how industry competitiveness influences
To test the robustness of our results, we also use the preceding earnings management in that industry. Hence, the analysis is con-
3-year average (i.e., normalized) Market Power. Using the 3-year ducted at the industry level. As described earlier, three different
average accomplishes two objectives. First, a higher Market Power industry-wide competitiveness metrics are utilized: industry med-
over a 3-year period reflects a more sustainable competitive ian Lerner Index, the inverse of the number of firms in the industry,
advantage than that from 1 year, and second, there is less likeli- and industry concentration. We include the same firm characteris-
hood of the association between our product market measure tic variables used in preceding firm level regressions as control
and accruals from being spurious. All our results in this study variables, except in this case the variables reflect the industry

Table 6
Industry competition and earnings management. This table reports the results of OLS regressions examining the impact of market power on discretionary accruals for a sample of
firms meeting our data requirements spanning 1987–2009. The dependent variable is the absolute level of discretionary accrual using modified Jones model with Kothari et al.
(2005) adjustment for firm performance. All the variables are as defined in earlier tables. All variables are averaged for each Fama and French industry grouping with the average
values used in the regressions. Firm characteristic control variables are suppressed for brevity in Panel B. p-Values reported in the parentheses are computed with standard errors
adjusted for industry-level clustering.

Independent variables Industry LI Concentration Inverse of # of firms


Model 1 Model 2 Model 3
Panel A: Total sample
Industry Measure 0.031 2.836 45.355
(0.004) (0.075) (0.029)
Growth 7.738 7.605 7.511
(0.001) (0.001) (0.001)
Log (Market-to-Book) 2.536 3.031 2.625
(0.000) (<0.000) (0.000)
Volatility of Sales 0.014 0.002 0.006
(0.501) (0.912) (0.781)
Size 10.265 12.082 11.357
(0.001) (0.000) (0.001)
Leverage 12.217 13.649 14.002
(0.021) (0.007) (0.005)
Constant 0.752 0.0745 0.081
(0.000) (<0.000) (<0.000)
Year Dummy Yes Yes Yes
Number of Obs. 953 953 953

Independent variables Pos DA Neg DA Pos DA Neg DA Pos DA Neg DA


Industry LI Concentration Inverse of # of firms
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
Panel B: Positive and negative absolute discretionary accruals separately
Industry Measure 0.040 0.030 2.448 3.359 55.074 38.378
(0.001) (0.008) (0.072) (0.085) (0.047) (0.115)
Firm Characteristic Controls Yes Yes Yes Yes Yes Yes
Year Dummy Yes Yes Yes Yes Yes Yes
Number of Obs. 953 953 953 953 953 953

Please cite this article in press as: Datta, S., et al. Product market power, industry structure, and corporate earnings management. J. Bank Finance (2013),
http://dx.doi.org/10.1016/j.jbankfin.2013.03.012
S. Datta et al. / Journal of Banking & Finance xxx (2013) xxx–xxx 11

medians. Standard errors are adjusted for industry-level clustering. firm-year observations during the period spanning 1987–2009,
Various configuration of the following general regression model we document that product market power of firms is inversely re-
are estimated at the industry level: lated to discretionary accruals management. Our findings are con-
sistent with the notion that lack of product market power makes it
Abs Disc Accruals ¼ b0 þ b1 Industry  lev el Competition more likely that a firm’s managers will engage in earnings manage-
þ b2 Growth þ b3 Market  to  book ment, whereas firms that enjoy a more powerful product market
position are less likely to do so. Our results imply that more intense
þ b4 Volatility þ b5 Size þ b6 Lev erage
competition prompts managers to restrict disclosure of informa-
þe ð6Þ tion to rivals and that managerial career concerns outweigh the
disciplinary impact of competition.
We control for business cycle effects by including year dummies in
By documenting product market power as a motive behind
the regression models. In Table 6 (Panel A) we report the results for
earnings management, we identify conditions under which earn-
the total sample, while Panel B displays the regression estimates for
ings (accruals) management is likely to be more aggressive. Specif-
partitioned samples based on the sign of the accruals (positive for
ically, we document that earnings reported by firms with weaker
inflationary earnings management and negative for deflationary
product market pricing power within an industry or firms in com-
earnings management). Model 1 (Panel A), which uses Industry LI
petitive industries are prone to greater earnings manipulation than
(industry median Lerner Index) as the focus explanatory variable,
their counterparts with (a) greater pricing power, and (b) in less
shows a significantly negative association with discretionary accru-
competitive industries. This study contrasts with much of the pre-
als. This result indicates that the industry-level competitiveness is
vious literature which focuses on documenting earnings manipula-
positively related to the degree of accruals management in that
tions that ‘‘games’’ the capital markets prior to a certain event such
industry. Given that industries characterized by thin margins are
as SEO or to boost executive bonuses. Our findings also do not sup-
associated with greater competitive pressure, the results imply that
port the view that competition serves as a disciplinary mechanism
more competitiveness leads to more earnings management, consis-
by providing more information transparency.
tent with the findings obtained when using firm-specific market
Overall, the knowledge derived from this study provides addi-
power. Consistent with Model 1, the negative and significant coef-
tional tools to assess the degree of earnings management by firms,
ficient estimate in Model 2 for Concentration also indicates that
thus enabling standard setters, financial market regulators, ana-
the less the competition in an industry, the less the degree of accru-
lysts and investors to make more informed legislative, regulatory,
als management in that industry. Testing the relevance of the in-
resource allocation, and investment decisions.
verse of the number of firms in an industry to accruals in Model 3
also reveals that greater competitive pressures lead to more
accruals.
Panel B of Table 6 presents results for absolute value of positive
Appendix A.
and negative accruals separately. The findings confirm that the re-
sults in Panel A hold for the two categories of accruals for each
A.1. Variable definitions
measure of industry competition. However, while the firms operat-
ing in a competitive, heavily-populated, industry show greater pro-
Institutional holdings for each firm is measured as the aggre-
pensity to inflate earnings with a significantly negative coefficient
gate shares held by all institutional investors as reported in the
(Model 5), the coefficient estimate in Model 6, although negative, is
13-F quarterly files divided by the total number of shares outstand-
not statistically significant at conventional levels indicating lack of
ing. We assume that in institutional holdings remains unchanged
a link between negative accruals and number of firms in the
in the intervening months till the subsequent quarter holdings
industry.
data become available in 13-F quarterly files.
Overall, our analysis at the industry level, which is in line with
Number of Analyst is number of valid estimates used to com-
Hypothesis 2A, supports the view that greater competition in an
pute mean monthly earnings forecast for each firm in IBES histor-
industry leads to a higher degree of earnings management for
ical summary files. In regressions, we use log (1 + num of analysts)/
the average firm in that industry as compared to firms in industries
100.
with lower competitive pressure. Our results are consistent with
Spread is computed as average of the daily closing (ask price-
view that more intense competition prompts the firm to manage
bid price)/closing price for each firm in each fiscal year. This vari-
earnings more. Thus, our empirical evidence does not support the
able is divided by 100 in regressions.
view that under more intense competitive pressure, full disclosure
CEO EBC is computed as the value of equity-based compensa-
prevails through unraveling of information. These findings also im-
tion (value of granted options and restricted stockholdings) to total
ply that career concerns outweigh the disciplinary impact of
CEO pay package.
competition.
Concentration is computed as the sum of 4 largest firms’ sales
as a fraction of aggregate sales of all firms that exist in the same
5. Conclusions industry in COMPUSTAT files in each fiscal year. We use Fama–
French 49 industry classification to define an industry.
This study adds an important new dimension to the earnings GIM Index is an index measuring shareholder rights devised by
management literature by establishing a link between product Gompers et al. (2003) incorporating 24 governance provisions and
market power of firms and their degree of earnings management. state antitakeover laws. Regression variable is GIM/100.
Specifically we establish a link between (a) firm level product mar- Growth is calculated as the change in total assets (COMPUSTAT
ket pricing power and the degree to which firms manage their re- item AT) scaled by 1 year lagged assets.
ported earnings and (b) industry competitiveness and the degree of Leverage is computed as 1 year lagged total long-term debt
earnings management for the industry. All our findings are highly (COMPUSTAT item DLTT) divided by total assets (COMPUSTAT item
robust to alternative measurements of variables, as well as control- AT).
ling for (a) internal and external disciplinary governance mecha- Market capitalization is computed as the product of number of
nisms, (b) executive compensation, and (c) the firm’s information shares outstanding and the market price of the share from CRSP
environment. Based on a comprehensive sample of 43,628 monthly files.

Please cite this article in press as: Datta, S., et al. Product market power, industry structure, and corporate earnings management. J. Bank Finance (2013),
http://dx.doi.org/10.1016/j.jbankfin.2013.03.012
12 S. Datta et al. / Journal of Banking & Finance xxx (2013) xxx–xxx

Market Power calculated as Sales (COMPUSTAT item sale) less Erickson, M., Wang, S., 1999. Earnings management by acquiring firms in stock for
stock mergers. Journal of Accounting and Economics 27, 149–176.
cost of goods sold (COMPUSTAT COGS) less sales, general and
Fama, E.F., 1980. Agency problems and the theory of the firm. Journal of Political
administrative expenses (COMPUSTAT item XSGA) divided by Economy 88, 288–307.
sales, which is then industry-adjusted by subtracting sales- Fan, J., Wong, T., 2002. Corporate ownership structure and the informativeness of
weighted price–cost margin of all firms within an industry. accounting earnings in East Asia. Journal of Accounting and Economics 33, 401–
425.
Median Industry Lerner Index is the median value of a firm’s Gal-Or, E., 1985. Information sharing in oligopoly. Econometrica 53, 329–343.
Lerner index in an industry in a fiscal year using all firms that exist Gaspar, J., Massa, M., 2006. Idiosyncratic volatility and product market competition.
in COMPUSTAT files. Lerner index for each firm is calculated as Journal of Business 79, 3125–3152.
Gertner, R., Gibbons, R., Scharfstein, D., 1988. Simultaneous signaling to the capital
Sales (COMPUSTAT item sale) less cost of goods sold (COMPUSTAT and product markets. RAND Journal of Economics 19, 173–190.
COGS) less sales, general and administrative expenses (COMPU- Giroud, X., Mueller, H., 2011. Corporate governance, product market competition,
STAT item XSGA). We use Fama–French 49 industry classification and equity prices. Journal of Finance 66, 563–600.
Gompers, P., Ishii, J., Metrick, A., 2003. Corporate governance and equity prices.
to define an industry. Quarterly Journal of Economics 118, 107–155.
Number of firms in an industry is the actual number of all Graham, J., Harvey, C., Rajgopal, S., 2005. The economic implications of corporate
firms in COMPUSTAT files in an industry in a fiscal year. We use financial reporting. Journal of Accounting and Economics 40, 3–73.
Grullon, G., Michaely, R., 2007. Corporate Payout Policy and Product Market
Fama–French 49 industry classification to define an industry. Competition. Working Paper. Rice University.
Market-to-book is the ratio of market capitalization to the Guadalupe, M., Pérez-González, F., 2010. Competition and Private Benefits of
book value of the firm (COMPUSTAT item CEQ). Regression variable Control. Working Paper. INSEAD.
Guidry, F., Leone, A., Rock, S., 1999. Earnings-based bonus plans and earnings
is log(1 + market-to-book)/100.
management by business unit managers. Journal of Accounting and Economics
Size is the natural logarithm of market capitalization/1000 (in 26, 113–142.
regressions). Harris, M.S., 1998. The association between competition and managers’ business
Volatility of sales is calculated as the standard deviation of segment reporting decisions. Journal of Accounting Research 36, 111–128.
Hart, O., 1983. The market mechanism as an incentive scheme. The Bell Journal of
sales over the preceding 3-year period scaled by 1 year lagged as- Economics 14, 366–382.
sets (COMPUSTAT data item AT). Haushalter, D., Klasa, S., Maxwell, W.F., 2006. The influence of product market
dynamics on a firm’s cash holdings and hedging behavior. Journal of Financial
Economics 84, 797–825.
References Hayes, R., Lundholm, R., 1996. Segment reporting to the capital market in the
presence of a competitor. Journal of Accounting Research 34, 261–279.
Healy, P., Wahlen, J., 1999. A review of the earnings management literature and its
Adams, B., Carow, K.A., Perry, T., 2009. Earnings management and initial public
implications for standard setting. Accounting Horizons 13, 365–383.
offerings: the case of the depository industry. Journal of Banking and Finance
Hermalin, B., Weisbach, M., 2012. Information disclosure and corporate governance.
33, 2363–2372.
Journal of Finance 67, 195–233.
Aghion, P., Dewatripont, M., Rey, P., 1999. Competition, financial discipline and
Hoberg, G., Phillips, G.M., 2010. Real and financial industry booms and busts.
growth. Review of Economic Studies 66, 825–852.
Journal of Finance 65, 45–86.
Aghion, P., Bloom, N., Blundell, R., Griffith, R., Howitt, P., 2005. Competition and
Holmström, B., 1982. Moral hazard in teams. Bell Journal of Economics 13, 324–
innovation: an inverted-U relationship. The Quarterly Journal of Economics 120,
340.
701–728.
Hribar, P., Collins, D., 2002. Errors in estimating accruals: implications for empirical
Akdogu, E., MacKay, P., 2012. Product markets and corporate investment: theory
research. Journal of Accounting Research 40, 105–134.
and evidence. Journal of Banking and Finance 35, 439–453.
Jagannathan, R., Srinivasan, S., 1999. Does product market competition reduce
Allen, F., Gale, D., 2000. Corporate Governance and Competition, Published in
agency costs? The North American Journal of Economics and Finance 10, 387–
Corporate Governance: Theoretical and Empirical Perspectives, Edited by X.
399.
Cambridge University Press, Vives, pp. 23–94.
Jones, J., 1991. Earnings management during import relief investigations. Journal of
Bagnoli, M., Watts, S.G., 2007. Oligopoly, Disclosure and Earnings Management.
Accounting Research 29, 193–228.
Working Paper. Purdue University.
Kale, J.R., Loon, Y.C., 2011. Product market power and stock market liquidity. Journal
Balakrishnan, K., Cohen, D., 2011. Product Market Competition, Financial
of Financial Markets 14, 376–410.
Accounting Misreporting and Corporate Governance: Evidence from
Karuna, C., 2007. Industry product market competition and managerial incentives.
Accounting Restatements. Working Paper. New York University.
Journal of Accounting and Economics 43, 275–297.
Bartov, E., Gul, F., Tsui, J., 2001. Discretionary-accrual models and audit
Klein, A., 2002. Audit committees, board of director characteristics and earnings
qualifications. Journal of Accounting and Economics 30, 421–452.
management. Journal of Accounting and Economics 33, 375–400.
Bebchuk, A., Cohen, A., Ferrell, A., 2004. What matters in corporate governance?
Kothari, S., Leone, A., Wasley, C., 2005. Performance matched discretionary accruals.
Review of Financial Studies 22, 783–827.
Journal of Accounting and Economics 39, 161–197.
Becker, C., DeFond, M., Jiambalvo, J., Subramanyam, K., 1998. The effect of audit
Kothari, S., Loutskina, E., Nikolaev, V., 2006. Agency Theory of Overvalued Equity as
quality on earnings management. Contemporary Accounting Research 15, 1–24.
an Explanation for the Accrual Anomaly. Working Paper. MIT Sloan School of
Beneish, M., Vargus, M., 2002. Insider trading, earnings quality, and accrual
Management.
mispricing. The Accounting Review 77, 755–791.
Lang, M., Lundholm, R., 1996. Corporate disclosure policy and analyst behavior. The
Bergstresser, D., Philippon, T., 2006. CEO incentives and earnings management.
Accounting Review 71, 467–492.
Journal of Financial Economics 80, 511–529.
Lee, C., Li, L., Heng, Y., 2006. Performance, growth and earnings management.
Bowen, R., Rajgopal, S., Venkatachalam, M., 2008. Accounting discretion, corporate
Review of Accounting Studies 11, 305–334.
governance and firm performance. Contemporary Accounting Research 25 (2),
Lee, C., Lusk, E., Halperin, M., 2010. CFO resignations: their underlying performance
351–405.
and behavioural context. International Journal of Behavioral Accounting and
Collins, D., Hribar, P., 2000. Earnings-based and accrual-based market anomalies:
Finance 1, 312–334.
one effect or two? Journal of Accounting and Economics 29, 101–124.
Lerner, A., 1934. The concept of monopoly and the measurement of monopoly
Cremers, K., Nair, V., Peyer, U., 2008. Takeover defenses and competition. Journal of
power. Review of Economic Studies 1, 157–175.
Empirical Legal Studies 5, 791–818.
Leuz, C., Verrecchia, R., 2000. The economic consequences of increased disclosure.
Curry, B., George, K., 1983. Industrial concentration: a survey. The Journal of
Journal of Accounting Research 38, 91–124.
Industrial Economics 31, 203–255.
Lindenberg, E., Ross, S., 1981. Tobin’s q ratio and industrial organization. Journal of
Datta, S., Iskandar-Datta, M., Sharma, V., 2011. Product market pricing power,
Business 54, 1–32.
industry concentration and analysts’ earnings forecasts. Journal of Banking and
Maksimovic, M., Pichler, P., 2001. Technological innovation and initial public
Finance 35, 1352–1366.
offerings. Review of Financial Studies 14, 459–494.
Dechow, P., Sloan, R., Sweeney, A., 1995. Detecting earnings management. The
Milgrom, P., Roberts, J., 1992. Economics, Organization and Management. Prentice
Accounting Review 70, 193–226.
Hall, Upper Saddle River, NJ.
Dechow, P., Sloan, R., Sweeney, A., 1996. Causes and consequences of earnings
Minton, B., Schrand, C., 1999. The impact of cash flow volatility on discretionary
manipulation: an analysis of firms subject to enforcement actions by the SEC.
investment and the costs of debt and equity financing. Journal of Financial
Contemporary Accounting Research 13, 1–36.
Economics 54, 423–460.
Demsetz, H., 1973. Industry structure, market rivalry, and public policy. Journal of
Nalebuff, B., Stiglitz, J., 1983. Prices and incentives: towards a general theory of
Law and Economics 16, 1–10.
compensation and competition. Bell Journal of Economics 14, 21–43.
Diamond, D.W., Verrecchia, R.E., 1991. Disclosure, liquidity, and the cost of capital.
Narayanan, N.P., 1985. Managerial incentives for short-term results. Journal of
Journal of Finance 46, 1325–1359.
Finance 40, 1469–1484.
Domowitz, I., Hubbard, R., Petersen, B., 1986. Business cycles and the relationship
Peress, J., 2010. Product market competition, insider trading and stock market
between concentration and price-cost margins. RAND Journal of Economics 17,
efficiency. Journal of Finance 6, 1–43.
1–17.

Please cite this article in press as: Datta, S., et al. Product market power, industry structure, and corporate earnings management. J. Bank Finance (2013),
http://dx.doi.org/10.1016/j.jbankfin.2013.03.012
S. Datta et al. / Journal of Banking & Finance xxx (2013) xxx–xxx 13

Perry, S., Williams, T., 1994. Earnings management preceding management buyout Shleifer, A., Vishny, R., 1997. A survey of corporate governance. Journal of Finance
offers. Journal of Accounting and Economics 18, 157–179. 52, 737–783.
Peteraf, M., 1993. The cornerstones of competitive advantage: a resource-based Skinner, D., Sloan, R., 2002. Earnings surprises, growth expectations, and stock
view. Strategic Management Journal 14, 179–191. returns or don’t let an earnings torpedo sink your portfolio. Review of
Porter, M., 1980. Competitive Strategy: Techniques for Analyzing Industries and Accounting Studies 7, 289–312.
Competitors. Free Press, New York. Stivers, A.E., 2004. Unraveling of information: competition and uncertainty. BE
Raith, M., 2003. Competition risk and managerial incentives. American Economic Journal of Theoretical Economics 4, 1–30.
Review 93, 1425–1436. Symeonidis, G., 2002. The Effects of Competition: Cartel Policy and the Evolution of
Rangan, S., 1998. Earnings management and the performance of seasoned equity Strategy and Structure in British Industry. The MIT Press, Cambridge, MA.
offerings. Journal of Financial Economics 50, 101–122. Teoh, H., Welch, I., Wong, T., 1998. Earnings management and the long-run
Ronen, J., Yaari, V., 2008. Earnings management, emerging insights in theory, performance of initial public offerings. Journal of Finance 53, 1935–1974.
practice, and research. Springer Series in Accounting Scholarship 3, ISBN 978-0- Tirole, J., 1998. The Theory of Industrial Organization. ISBN:978-0-262-20071-4.
387-25769-3. Verrecchia, R., 1983. Discretionary disclosure. Journal of Accounting and Economics
Rotemberg, J., Scharfstein, D., 1990. Shareholder-value maximization and product- 5, 365–380.
market competition. Review of Financial Studies 3, 367–391. Verrecchia, R., Weber, J., 2006. Redacted disclosure. Journal of Accounting Research
Salop, S., 1979. Strategic entry deterrence. American Economic Review 69, 335–338. 44, 791–814.
Scharfstein, D., 1988. The disciplinary role of takeovers. Review of Economic Studies Yu, F., 2008. Analyst coverage and earnings management. Journal of Financial
55, 185–199. Economics 88, 245–271.
Schmidt, K., 1997. Managerial incentives and product market competition. Review Yu, Q., Du, B., Sun, Q., 2006. Earnings management at rights issues thresholds:
of Economic Studies 64, 191–213. Evidence from China. Journal of Banking and Finance 30, 3453–3468.
Shleifer, A., 2004. Does competition destroy ethical behavior? American Economic
Review 94, 414–418.

Please cite this article in press as: Datta, S., et al. Product market power, industry structure, and corporate earnings management. J. Bank Finance (2013),
http://dx.doi.org/10.1016/j.jbankfin.2013.03.012

You might also like