International Conference on Logistics Engineering, Management and Computer Science (LEMCS 2015)

Type of Earnings Management and Different

Economic Consequences
Tangmei Yuan*
Linyi University
Department of Accounting, Linyi University
Shandong, China
e-mail: yyytm@126.com

Abstract—The purpose of this study is to investigate (Scott,2000)[1]. On the other hand, Healy & Whalen`s(1999)
whether different types of earnings management have earnings management definition contains AEM and REM
different effect on the firms future profitability and market two aspects, “earnings management occurs when managers
value. Using multiple regressions and large sample, this use judgment in financial reporting and in structuring
study finds evidence that accrued earnings management is transactions to alter financial reports to either mislead
positive related with firms future profitability and firms
market value, but real earnings management is negative some stakeholders about the underlying economic
related with firms future profitability and firms market performance of the company or to influence contractual
value. It indicate that accrual earnings management is outcomes that depend on reported accounting numbers.”
effective, real earnings management is aggressive in China. Scott(2000)[1] define accrual earn ings management as
within the permission scope of Generally Accepted
Keywords-Earnings Management; Type; effective;
aggressive; Economic Consequences Accounting Principles, the behavior that maximize
operator's own interests or (and) the company's market
value through accounting policy choice.
I. INT RODUCT ION Roychowdhury(2006)[6] define real earnings management
When talk about of earnings management, the common as departures fro m normal operational practices with the
views regard it would harm to the firms future primary objective of meeting near-term earnings goals.
development, damage the firms health. However, the Meanwhile, we can use Jones or other models to
earnings managements are not all damage the firm future decompose accruals into non-discretionary accrual and
value. According to prior literature, There are two views of discretionary accrual; Use Roychowdhury(2006)[6]
earnings management: one view think that earnings developed models to separate the normal fro m the
management was effective, earnings management can abnormal levels of real operational activities as reflected in
improve earnings informativeness in communicating cash flows fro m operations (CFO), production costs, and
private information. The other view regard as earnings discretionary expenditures. Since Roychowdhury’s work,
management is opportunistic, management using it to subsequent studies dealing with REM issues have provided
reports high earnings opportunistically to maximize h is or evidence supporting that managers engage in real activities
her utility(Scott, 2000)[1]. Subramanyam(1996)[2], Gul et manipulation to meet certain earnings targets.
al.(2000)[3], and Krishnan (2003) research find that the
behavior of discretionary accruals (a proxy for accrual B. AEM: Effective Hypothesis and Opportunistic
earnings management, AEM) is consistent with the Hypothesis
efficient perspective, because discretionary accruals have a According to prior literatures, AEM can be divided into
positive and significant relationship with future two types: efficient earnings management (i.e., to improve
profitability. On the other hand, Burgstahler and Dichev earnings informat iveness in communicating private
(1997)[4] and Balsam et al.(2002)[5] research evidence that information) and opportunistic earnings management (i.e.,
the earnings management is consistent with the management reports earnings opportunistically to
opportunistic perspective. maximize his/her utility) (Scott, 2000)[1]. Several studies
We examine the relationship between earnings find evidence consistent with the opportunistic perspective.
management and firms future profitability and market Burgstahler and Dichev(1997)[4]find that management
value, in order to investigate the earnings management is engages in earnings management to avoid reporting losses
efficient or opportunistic. Our research find that different or earnings decline. Balsam et al.(2002) find a negative
types of earnings management have different effect on the relationship between unexpected discretionary accruals
firms future profitability and market value in China. and stock returns around the earnings announcement date,
this result indicates that the market views discretionary
II. LIT ERAT URE REVIEW accruals as opportunistic. In contrast, other studies find
evidence that is consistent with the efficient perspective.
A. Prior Research on Type of Earnings Management. Subramanyam(1996)[2] concludes that discretionary
Earn ings management can be divided into efficient accruals are efficient because they have a positive and
earnings management and opportunistic earnings significant relationship with future profitability. This
management according to its economic consequences positive relationship describes the ability that discretionary
accruals have to communicate information about a firm's

© 2015. The authors - Published by Atlantis Press 1792

future profitability to the public. Gul et al. (2000)[3] and system with no direct cash flo w consequences, real
Krishnan (2003), following Subramanyam(1996)[2], also earnings management generally sacrifices firms’ long-run
find consistent evidence. cash flows in order to inflate short-run reported earnings.
Jeong-Bon Kim(2013)[9] study documents evidence that
C. REM: Information Communicate Hypothesis and REM is positively associated with the implied cost of
Value Damage Hypothesis equity after accounting for the effect of AEM. Widely
The view of "Information Co mmunicate Hypothesis" accepted view considers real activities manipulation have
consider real earnings management could conveying to the more harm to the firms health compared to accrued
market a signal of better performance (Gunny, 2010). earnings management. It is said that accrued earnings
Bartov (1993) documents that managers alter real activities management is more efficient, but real earnings
such as selling fixed assets to avoid debt covenant management is more opportunistic. So, put forward the
violations, suggesting one potential benefit of real earnings following hypothesis:
management to shareholders. As Roychowdhury (2006)[6] H1: accrued earnings management is positive related
suggest, not all deviations from normal real activities are with firms future profitability and firms market value.
intended to avoid earnings disappointments. For examp le, H2: real earnings management is negative related with
managers may cut R&D budgets either because other long- firms future profitability and firms market value.
term projects are more likely to be successful or because
they are faced with decreasing returns on R&D. Gunny B. Research Methodology and Research Model
(2010) finds that REM is positively associated with future Following Subramanyam (1996)[2], we use the
period earnings and cash flow performance for the firms following research model to test Hypothesis. If the
that just meet or beat their earnings benchmarks, wh ich coefficient on AEM(1) more than zero the H1 is confirmed,
indicates that these firms use REM to signal future firm and if the coefficient on REM(r) less than zero, the H2 is
prospects. confirmed.
"Value Damage Hypothesis" views regard real
earnings management as more likely to entail substantial FutProf it /MV= a0 + a1 DACit +arREM it +a3 NDACit
costs to shareholders. So me research find that Real +a4 NMCFOit +a5 Sizeit +a6 Lev it +∑λ j IDj
earnings management is an opportunistic behavior that +∑kYeark + εit (Model 1)
benefits managers while impairing shareholder interests
(e.g., Dechow and Sloan,1991[7]; Bens et al., 2002). Stein’s FutProfit / MV= a0 +a1 DACit +a2 REit +a3 NDACit
(1988) predict ion that real earnings management, a more +a4 NMCFOit +a5 Sizeit +a6 Lev it +∑λ j IDj
costly earnings management activity that allows managers +∑kYeark +εit (Model 2)
to meet earnings targets by altering real activities, thus
signaling the firm’s future performance. After that, Where: A. Dependent variables
Zang(2006)[8], Bhojraj et al.(2009), Leggett et al.(2009) all FutProfit /MV: Future Profitability and market value,
find evidence that REM have negative effect on firms measured by each of the following variables: (1)
future profitability, support the viewpoint of REM is not OIt+1 =Next year operation income; (2) NIt+1 =Next year net
the optimu m decision making. Using a sample of SEO income; (3)ROA t+1 =next year return of asset; (4)
firms, M izik and Jacobson (2007) find that to temporarily MVt+1 =Next year market value; (5)TQt+1 =Next year
inflate stock prices at the time of SEOs, managers engage TobinQ .
in boosting reported earnings via cutting marketing B. Independent variables are as defined in Table 1
expenses, but in the long run, such managerial myopia Earn ings are decomposed into four variables: non-
leads to a decline in stock market performance. Francis et discretionary accruals (NDAC), discretionary accruals
al.(2011) find that REM has more interpretation than AEM (DAC), normal CFO(NM CFO), and abnormal CFO(-abn
for stock "price diving" after issue SOX. Cohen and CFO). DAC and REM is the variable of interest and if the
Zarowin (2010) document significant post-SEO earnings type of earnings management is efficient, the coefficient
declines that are attributable to abnormal real activities (aa, ar) will be positive, otherwise, it will be negative.
around SEOs. In addition, Zhao et al.(2012) find that Other variables relate to firm size, Lev, ID, Year are
abnormal real activ ities in the absence of just meeting included as control variables.
earnings targets are negatively associated with firms’
future performance. Kim and Sohn(2013) research find C. Main Variables Measure
that REM is positively associated with the implied cost of Following the previous research, we measure earnings
equity after accounting for the effect of AEM, and the use management(EM) use different earnings management
of REM could be costly to a firm because the market sees proxy that will be calculated by using different models.
through its cash flow consequences and is able to factor 1) Discretionary Accruals (DAC) Proxy
this into the increased cost of capital. We measure accrual earnings management use
Discretionary Accruals (DAC) pro xy that will be
III. HYPOT HESES DEVELOPMENT AND RESEARCH calculated by using the modified Jones model (Jones,
M ET HODOLOGY 1991)[10]. This model has been found to be the most
powerful in detecting earnings management among
A. Hypotheses Development competing models (Dechow et al., 1995; Subramanyam,
AEM within-GAAP discretionary accounting choices 1996[2]; Bartov et al.,2002[11], as well as effective
are less costly and less resource-consuming. Unlike (Davidson et al., 2004) and reliable(Guay et al.,1996).
accruals management affects the output of the accounting DeFond and Jiambalvo(1994), Subramanyam(1996),

1793
develop the Jones time series model to the cross section costs, which is the sum of the cost of goods sold and the
model and find the measurement effect of cross section change in inventory (Co mpustat item COGS + change in
model is better than the time series model. So, this article INVT), and DEXP denotes dis cretionary expenses
selects cross section data perform OLS regression. computed by the sum of advertising expenses, selling,
The model involves the estimation of earn ings general and administrative expenses (Co mpustat items
management as the difference between the firm's actual XAD + XSGA).
and expected accruals. Accruals are measured as the Abnormal CFO, abnormal Prod, and abnormal Dis
difference between reported earnings and operating cash EXP, denoted by abn CFO, abn Prod, and abn DEXP,
flows, the pro xy "DA C" is the difference between reported respectively, are the differences between actual values of
net income and operating cash flows. Expected accruals CFO, Prod, and DEXP (all lagged-asset deflated) and their
(NTACC) were co mputed by regressing total accruals in normal levels (i.e., the fitted values of Eqs. (2), (3) and (4),
the firm's 2-digit SIC-code industry on total assets, respectively). To make the direction of the discretionary
revenues, property, plant, and equipment, and accounts expenditure-based measure consistent with abn Prod, we
receivable. construct our first proxy for abnormal real activ ities (-abn
Intuitively, the discretionary accruals proxy will be CFO) as the abnormal CFO mu ltiplied by -1, such that a
calculated by using the flowing modified Jones model (Eq. higher value of -abn CFO indicates more severe
1): manipulation of real activities; We construct our third
proxy for abnormal real activities (-abn DEXP) as the
DACit =ANACit =TACit -NTACit = NIit – CFOit -NTACit abnormal d iscretionary expenditures multip lied by -1, such
= TACit /A it-1 -[a0 (1/A it-1)+a1 [⊿REVit -⊿ RECit ]/A it-1 +a2 that a higher value of -abn DEXP indicates that it is more
(PPEit /A it-1 )] (Eq. 1) likely the firm cuts discretionary expenditures.
In order to capture the overall level of real activ ities
Where, TAC is total accruals; DAC: Discretionary manipulation, consistent with Zang(2006)[8], we also
accruals; NI: Net income; CFOit: cash flows fro m compute comprehensive metrics of abnormal real activities
operating activities of firm i for period t; ANAC: as our fourth proxy for abnormal real activities by
Abnormal accruals; NTA C: Normal total accruals, aggregating all three individual measures (i.e., abn Prod, -
expected total accruals by modified Jones model; Ait-1: abn DExp and -abnCFO) into one measure(RE). As is true
Beginning total asset for firm i in year t; ΔREVit : Change for the individual measures, the higher the values of the
in revenue fro m year t-1 to yeart (REVt −REVt-1); PPEit: comprehensive metrics, the more likely that the firm is
gross property, plant, and equipment in year t; ΔRECit: engaging in real activities manipulation.
Change in net accounts receivables in year t from year t−1 TABLE I. DEFINITIONS OF VARIABLES USED IN THE MODELS
to year t (RECit−RECit-1).
2) The Estimate of the Real ActivitiesVariables Definitions
DAC = T he discretionary accruals calculated using modified Jones model
Manipulation(REM)
As in prior studies (e.g., Roychowdhury, 2006; Cohen-abn CFO = T he level of abnormal cash flows from operations multiplied by -1
The level of abnormal production costs, where production costs are
et al., 2008; Cohen and Zarowin, 2010), we measure a abn = defined as the sum of cost of goods sold and the change in
firm’s deviations from its normal business practices by PROD
inventories
computing its abnormal production costs (conducting -abn
= T he level of abnormal discretionary expenses multiplied by -1
overproduction to report a lower cost of goods sold in the DEXP
Comprehensive metrics of abnormal real activities by aggregating
current period), abnormal discretionary expenditures RE =
abn Prod, -abnDExp and -abn CFO into one measure.
(reducing discretionary expenditures in the current period), OI = Operation income.
and abnormal cash flows from operations(offering NI = Current year net income scaled by total sales
excessive sales discounts or lenient credit terms to ROA = Return on assets
temporarily boost sales revenues in the current period). MV = Market value.
We using the following industry-year linear regression T Q = T obin Q.
Eqs. (2), (3) and (4) to decompose the actual CFO, CFO = Cash flows from operating activities
production costs, and discretionary expenses into the NMCFO = Normal cash flows from operating activities
normal, expected portion and the abnormal, unexpected
Size = Natural logarithm of total assets
portion, respectively. Where each industry is defined by its Lev = leverage ratio, equals to total liabilities divided by total assets
two-digit Standard Industrial Classification (SIC) code:
ID = Industry dummy variable, a control based on two-digit SIC codes
dummy variables for years 2007-2010 (A control based on calendar
CFOit /A it-1 =β1 (1/A it-1 )+β 2 (Sit /A it-1 )+β 3 (ΔSit /A it-1 )+ε t (Eq. 2) Year =
year).
t = 2007, 2008, 2009, 2010, 2011
Prod it /A it-1=β1 (1/A it-1 )+β 2 (Sit/A it-1 )+β 3 (ΔSit /Ait-1)
+β4 (ΔSit-1 /A it-1 )+εt (Eq. 3) t+1 = 2008, 2009, 2010, 2011, 2012

DEXP/A it-1 =β 1 (1/A it-1 )+β 2 (Sit /A it-1 )+ε t (Eq. 4) IV. DAT A

A. Sample selection
In the above Eqs. (2), (3) and (4), for each firm i and
We obtained data from Wind and CSMAR database for
year t, CFO is cash flow fro m operations, Ait-1: total
all firms that appeared during the period 2007–2011, we
assets for firm i in year t-1. Sit the firm i’s revenues in year
initially obtained all listed firms in the CSMA R 12350
t, ΔSit the firm i’s change in revenues between year t-1 and
firm-years. The following criteria are applied in selecting
year t, and εt is the error term. Prod refers to production

1794
 firms for the sample: (1) Excluding in financial firms, B. Analysis effects of REM
insurance industry firms, real estate industry firms, The effects of -abnCFO and abnPROD are all
communication and cultural industries firms significantly negative associated with firms future
(Roychowdhury,2006)[6]; (2) Excluding Incomp lete data profitability and market value in table 3; In table 4：The
firms; (3) Excluded owner's equity is less than or equal to
effects of RE are all significantly negative associated with
zero firms； Main business income is less than or equal to firms future profitability and market value.
zero firms; (4) Excluded industry unavailable firms. The coefficient on -abn DEXP are not significant
Because we further require that financial data is available associated with OIt+1, NIt+1, M Vt+1 and TQt+1, it only
fro m database for at least 8 firms operating in the same 2- negative related with ROAt+1 significant at 1% level, it
digit SIC industry, in order to compute earnings indicate that -abn DEXP has not significant effect on firms
management as the deviation of firms accruals relative to future profitability and the firms value in China capital
the industry norm. (5) Excluded continuous 3 years data market.
were unavailable firms. We have the final sample This indicate that managers manipulate real earnings
composed of 5984 firm-year observations. management(mainly abn CFO, abn PROD) affect firms
B. Descriptive Statistics future operation activities and firms value significantly.
This is consistent with views of Value Damage
Table 2 shows the mean, standard deviation, variance,
Hypothesis(Dechow and Sloan, 1991[7]; Bens et al., 2002;
minimu ms and maximu ms all variables of sample. The Zang, 2006[8]; Bhojraj et al., 2009; Mizik and Jacobson,
mean of every earnings management proxy is 0.0054, -
2007; Leggett et al., 2009; Cohen and Zarowin, 2010;
0.0092 ， -0.0156 ， -0.0088, -0.0336 respectively, the Francis et al., 2011; Zhao et al., 2012; Kim and Sohn,
maximu m is 0.3580, 0.3580, 0.5249, 1.7592, 0.3193, 2013). Suggesting a general value-destroying effect of
1.9779 respectively. managers manipulate real earnings management activities.
TABLE II. P RESENTS DESCRIP TIVE STATISTICS FOR ALL SAMP LE The results(list in Table 3, Table 4) for other control
variables show that: large firms have lower market value;
N MIN MAX MEAN VAR ST D and Firms with high LEV are significantly negative
OI t+1 5984 -1.3450 9.3595 0.0517 0.1718 0.0300 correlated with the firms future profitability and market
value . They are all consistent with the previous literature.
NI t+1 5984 -1.3075 9.8013 0.0609 0.1763 0.0310
Based on above analysis, AEM(discretionary accruals
CFO t+1 5984 -1.7688 9.4864 0.0627 0.1707 0.0290 manipulation)in China firms is consistent with the efficient
ROA t+1 5984 -1.2915 0.5322 0.0340 0.0722 0.0050 perspective, on the contrary, REM is consistent with the
MVt+1 5984 0.0000 78.1955 2.3060 2.0825 4.3370 Value Damage Hypothesis. It is say that abnormal accrual
T Q t+1 5984 0.4196 21.8956 1.9546 1.3689 1.8740
has positive significantly affect on firms future
profitability and market value, but real earnings
DAC 5984 -0.4132 0.3580 0.0054 0.0849 0.0070 management has negative significantly affect. Therefore,
-abnCFO 5984 -0.4168 0.5249 -0.0092 0.0885 0.0080 the H1 and H2 are all confirmed.
abnPROD 5984 -1.1387 1.7592 -0.0156 0.1360 0.0190
TABLE III. REGRESSION RESULT OF MODEL 1
-abn
5984 -0.4042 0.3193 -0.0088 0.0673 0.0050
DEXP
5984 -1.4575 1.9779 -0.0336 0.2304 0.0531 OIt+1 NIt+1 ROAt+1 MVt+1 TQt+1
RE
Intercept .070*** .139*** -.059*** 14.614*** 11.677***
NDAC 5984 -0.7999 2.2878 -0.0084 0.0651 0.0040 (1.578) (3.048) (-3.580) (30.026) （39.664）
DAC .383*** .388*** .305*** 2.928*** 1.325***
NMCFO 5984 -0.3267 0.6064 0.0438 0.0497 0.0020
(9.948) (9.772) (21.053) (6.895) （5.165）
Size 5984 18.1624 28.2833 21.7569 1.2332 1.5210- abnCFO -.520*** -.527*** -.396*** -3.833*** -2.137***
(-11.89) (-11.671) (-24.02) (-7.937) （-7.322）
lev 5984 0.0071 0.9981 0.5021 0.1910 0.0360
abnPROD -.072*** -.068*** -.039*** -1.024*** -.767***
(-3.511) (-3.202) (-5.000) (-4.521) （-5.605）
V. REGRESSION RESULT S OF M ODEL - .000 -.018 -.045*** .060 -.293
abnDEXP (.011) (-.483) (-3.273) (.147) （-1.190）
We report the regression results in Table 3 and Table 4.
NDAC .277*** .258*** .162*** 1.254*** 1.089***
A. Analysis effects of AEM (7.992) (7.209) (12.434) (3.276) (4.706）
NMCFO .579*** .583*** .413*** 3.464*** 1.797***
Table 3 note that AEM(DAC) are all positive(0.383, (10.931) (10.672) (20.694) (5.924) (5.084)
0.388, 0.305, 2.928, 1.325,) and significant(t=9.948, 9.772, Size -.001 -.004** .004*** -.580*** -.450***
21.053, 6.895,5.165; p=0.000) at 1% level associated with (-.653) (-2.120) (5.524) (-25.942) (-33.276)
Lev -.064*** -.051*** -.049*** -.452*** -.481***
firms next year OI, NI, ROA, M V and Tobin Q. In Table 4: (-4.797) (-3.675) (-9.654) (-3.065) (-5.397)
The effects of AEM are also all significantly positive ID Control Control Control Control Control
associated with firms future profitability and market value. Year Control Control Control Control Control
This is consistent with Subramanyam(1996)[2], Gul, et adjR 2 .095 .086 .273 .250 .366
al.(2000)[3], and Krishnan (2003) views of d iscretionary 23.513*** 21.180*** 81.258*** 72.234*** 124.478***
F
accruals is consistent with the efficient perspective,
max VIF 3.363 3.363 3.363 3.363 3.362
opposite to Burgstahler, et al.(1997) and Balsam, et
Note: 1)Model: FutProf it / MV= a0 +a1 AEM it
al.(2002) v iews of the earnings management is
+arREM it +a3 NDACit +a4 NMCFOit +a5 Sizeit +a6 Lev it
opportunistic perspective.
+∑λ jIDj +∑kYeark + εit ; 2) A ll variables are as defined in

1795
 Table 1; 3)***,** and * denote the significance level at Those evidences indicate that different types of earnings
1%，5% and 10% respectively. management have different effect on the firms future
profitability and market value, accrual earnings
TABLE IV. REGRESSION RESULT OF MODEL 2 management is effective, real earnings management is
OIt+1 NIt+1 ROA t+1 M Vt+1 TQ t+1 aggressive in China.
Intercept 0.015 0.082* -0.1*** 14.29*** 11.52*** REFERENCES
(0.349) (1.825) (-6.19) (29.7) (39.62)
[1] SCOTT R W. Financial accounting theory[M].2nd Ed. New Jersey:
DAC 0.129*** 0.132*** 0.114*** 1.227*** 0.505*** Prentice Hall, 2000.
(4.663) (4.641) (10.73) (4.037) (2.755)
[2] SUBRAMANYAM K R. The pricing of discretionary
RE -0.13*** -0.13*** -0.10*** -1.23*** -0.88***
accruals[J].Journal of Accounting and Economics ,1996,22:
(-11.8) (-11.7) (-23.6) (-10.3) (-12.3) 249−281.
NDAC 0.290 0.271*** 0.173*** 1.33*** 1.126*** [3] Gul, F. A., S. Lung, & B. Srinidhi. The effect of investment
(8.305) (7.533) (12.88) (3.468) (4.866) opportunity set and debt level on earnings–returns relationship and
NM CFO 0.386*** 0.388*** 0.265*** 2.196*** 1.185*** the pricing of discretionary accruals. Unpublished Working Paper,
(7.898) (7.694) (14.08) (4.091) (3.656) 2000.
Size 0.002 -0.000 0.007*** -0.56*** -0.44*** [4] BURGST AHLER D, DICHEV I. Earnings management to avoid
(1.066) (-0.44) (9.063) (-25.5) (-33.2) earnings decreases and losses[J]. Journal of Accounting and
lev -0.10*** -0.08*** -0.07*** -0.69*** -0.60*** Economics,1997, 24:99−126.
(-7.46) (-6.20) (-14.3) (-4.830) -6.88 [5] BALSAM S, BARTOV E, MARQUARDT C. Accruals
ID Control Control Control Control Control management, investor sophistication, and equity valuation:
Evidence from 10-Q filings[J]. Journal of Accounting
Year Control Control Control Control Control Research,2002,40(4): 987−1012.
adj R2 0.081 0.073 0.229 0.245 0.364 [6] ROYCHOWDHURY S. Earnings management through real
F 21.22*** 19.04*** 69.44*** 75.7*** 132.*** activities manipulation[J]. Journal of Accounting and
Economics,2006,42: 335–370
M ax 1.369 1.369 1.369 1.369 1.369 [7] DECHOW P M, SLOAN R G, SWEENEY A P. Causes and
VIF consequences of earnings manipulation: An analysis of firms
Note: 1)Model: FutProfit /MV=a0 +a1 AEM it +arREit subject to enforcement actions by SEC[J]. Contemporary
+a2 NDACit +a3 NMCFOit +a5 Sizeit +a6 Lev it +∑λ jIDj Accounting Research,1991,13(1):1−36.
+∑kYeark + εit ; 2) All variables are as defined in Table 1; [8] ZANG A Z. Evidence on the tradeoff between real manipulation
3)***,** and * denote the significance level at 1%，5% and accrual manipulation[J]. Working paper, Duke University,2006.
and 10% respectively. [9] Kim Jeong-Bon. Real earnings management and cost of
capital[J].Account Public Policy,2013.
VI. CONCLUSIONS [10] JONES, J. J. Earnings management during import relief
investigation[J].Journal of Accounting Research, 1991,
This paper examines that accrued earnings 29:193−228.
management(A EM) has positive and significant related to [11] BART OV E, GIVOLY D, HAYN C. The reward to meeting or
the firms future profitability and market value, on the beating analysts’ forecasts[J]. Journal of Accounting and
contrast, mostly real earnings management pro xies(abn Economics,2002, 33: 173–204.
CFO, abn PROD, RE) have negative and significant
related to the firms future profitability and market value.

1796