The impact of the foreign ownership on the real earnings management of firms in vietnam

Tran Thi Thuy Linh et.
al. | 1

The impact of the foreign ownership on
the real earnings management of firms
in Vietnam
TRAN THI THUY LINH
University of Economics HCMC –

NGUYEN THANH NHA


HOANG THI PHUONG ANH


DANG THI VY NGOC


Abstract
The paper analysed the impacts of the foreign ownership on the
behaviour of the real earnings management of firms in Vietnam and the
difference in the quality of operating income between firms with high and
low levels of the foreign ownership. Data of 167 firms listed on the
Vietnam Stock Exchange for the duration from 2011 to 2015 was taken as
the case study and the regression models which are OLS, Robust and
2SLS were applied.
Results of this study demonstrated that the foreign ownership
contributed to the behaviour of the real earnings adjustment of firms; a
high level of the foreign ownership was found to positively affect the real
earnings management of firms in Vietnam. In addition, this study
supported the knowledge spillover theory that highly skilled foreign

investors had the tendency to assist their owned firms to improve the
earnings quality on financial reports. Therefore, it was important for
financial managers to develop appropriate strategies to manage and
improve the quality of financial reports according to the International
Financial Reporting Standards (IFRS). This helped increase competitiveness
and transparency of firms to attract more foreign investors on the
Vietnam Stock Exchange.
Keywords: foreign ownership; real
earnings
adjustment;
real
earnings management; earnings quality; abnormal levels of operating
cash flows.


1. Introduction
In emerging markets, foreign investors play an important role in
improving the quality of financial reports of domestic firms
(Huang and Shiu, 2009). They can promote the management of
domestic firms by using the voting right to influence decisions
of managers or to announce divestment from domestic firms
(Aggarwal et al., 2005; Gillan and Starks, 2003). According to the
report of World Bank in 2016, the foreign investment increased by
40%, which was approximately $1.8 trillion in 2015; this was the
highest since the global financial crisis in 2008. Therefore,
foreign investors and investment greatly affect the financial
management of firms.
In Vietnam, for the past 10 years, the Government has
developed and loosened regulations of attracting foreign
investment. Together with the admission of the international

trade organisation, this has contributed significantly to the
integration of Vietnam into the world market. Vietnam has
removed barriers for foreign investors and facilitate both indirect
and direct capital investment. Using 167 firms listed on Vietnam
Stock Exchange as the case study, this paper focuses on
analysing impacts and relationship between the foreign
ownership and earnings quality of these firms. The objectives of
this study are to uncover: (1) influences of the foreign
investment on the earnings management of firms and (2)
impact of high and low levels of the foreign ownership on the
earnings quality of firms.
2. Background theories and empirical studies
2.1.

Background theories

The study of impacts of the foreign ownership on the earnings
quality was based on the asymmetric information theory and the
knowledge spillover theory.
The asymmetric information theory was studied by George
Akerlof, Michael Spence and Joseph Stigliz (1970). Regarding to
the earnings management, the asymmetric information occurs
when managers take the advantage of the fact that foreign
investors have very little information of situation of firms and
domestic market and, therefore, make fraudulent acts, adjust
profit and indicators on financial reports. These activities lead to
the fact that investors, particular foreign investors, cannot control


behaviours of firms; they are classified as the moral hazard. In

addition, the standards and regulations


of accounting in a particular country may cause restrictions on
the understanding of foreign shareholders. Therefore, in a firm
with a large number of foreign shareholders, managers can
manage and adjust income for their own purposes. This is
called the asymmetric information which is less favourable by
foreign investors than domestic investors. As pointed out by
Choe et al. (2001) and Dvorak (2005), in the Indonesian market,
domestic investors have more profit compared to foreign
investors. Thus, according to the asymmetric theory, the real
earnings adjustment is more likely on firms with higher level of the
foreign ownership.
The knowledge spillover theory
Alfred Marshall (1890) was the first researcher studying
impacts of the kwowledge spillover theory, which was extended
by the economist Kenneth Arrow (1962) and Paul Romer (1986).
Later, Edward Glaeser, Hedi Kallal, José Scheinkman and Vàrei
Shleifer (1992) developed a theory called MAR spillover. Accoring
to Carlino (2001), the spillover facilitates exchange of ideas and
promotion of creativity and innovation which are the important
factors for sustainable economic development of countries.
Based on this theory, the concentration of firms in a similar
industry helps spread knowledge and initiate growth and
innovation. Regarding to issues of the real earnings
management, knowledge of accounting and management is
enhanced via foreign investors, which enables firms to
effectively monitor business activities and report financial status.
Therefore, the knowledge spillover theory suggests that firms

with high level of the foreign ownership tend to perform more
real earnings adjustment in a comparison with ones having low
level of the foreign ownership.
2.2.

Empirical studies

Xiao et al. (2004) argued that foreign investors faced more
challenges due to the asymmetric information than the domestic
investors did and managers had more information of status of
firms and market compared to foreign investors. Therefore, real
earnings adjustment is still likely in firms having high foreign
ownership. Similarly, Klai and Omri (2001) showed that the
presence of foreign investors reduced the quality of financial
reports. Adebiyi and Olowookere (2006) argued that firms with
high level of foreign ownership are associated with high real
earning adjustment. On the contrary, other studies showed that,
with better knowledge and information, foreign investors are more


capable of spreading their expertise to domestic firms to improve
the quality and


efficiency of their business productivity. According to Seaholes
(2000), foreign financial institutions perform better than
domestic ones, since the foreign ownerships creates significant
economic benefits. Based on studies in Canada, Italy, Germany,
Japan, UK and US, Hejazi and Safarian (199) pointed out that the
foreign ownership has positive spillover effects on efficiency and

technology. Hallward-Driemeiter et al. (2002) conducted surveys
in Indonesia, South Korea, Malaysia, Philippines and Thailand and
found that foreign- owned enterprises in East Asia are more
productive compared to other firms. The study of Ho et al. (2010)
showed that, in small enterprises, the more the foreign ownership,
the greater the relationship between technology investment and
firm productivity; therefore, foreign investors tend to bring IT
professionals to help small enterprises. Abor and Biekpe (2007)
demonstrated that foreign ownership has positive impact on
the quality and productivity of small and medium enterprises.
According to Aydin, Sayim and Yalama (2007), foreign-owned
enterprises contribute more to enhance the financial performance
and improve the performance compared to owned enterprises in
the county. Gillan and Starks (2003) argued that foreign investors
are the key role in promoting changes in the corporate
governance of firms through direct monitoring by using their
voting rights to affect managerial decisions or indirect monitoring
by threatening to sell their shares. Aggarwal et al. (2011)
investigated impacts of financial institutions on the corporate
governance system and argued that a firm with investment
from foreign financial institutions has a stronger foundation to
promote the development of management. According to BenNasr, using data collected from firms in 45 different countries,
Boubakri and Cosset (2015) showed that the higher level the
foreign ownership the higher the quality of income. Alaryan
(2015) pointed out that the foreign ownership plays a vital role in
ensuring the quality of profit. Taking Jordan as an example, he
found that the knowledge spillover theory was supported and
that the higher foreign ownership leads to more control over the
management of domestic business income. Gue et al. (2005)
studies foreign shareholders and real earnings adjustment on

financial reports and suggested that foreign shareholders have
positive effects on the quality of the real earnings. The higher
the foreign ownership, the lower the real earnings adjustment; this
is the factor to restrict real earnings adjustment.


3. Research methodology
3.1.

Research data

Research data was collected from the following websites:
www.hsx.vn, The main
sample of data used in this study was 167 firms listed on Ho Chi
Minh Stock Exchange (HOSE) in the duration from 2011 to 2015
and contained 835 observations in 5 main major disciplines
including Consumer Discretionary, Industrials, Real Eastate,
Materials and Consumer Staples. Firms were categorised
according to the Global Industry Classification Standards (GICS).
Table 1
Foreign ownership of different disciplines
N
o
1

Disciplines

2

Consumer

Discretionary
Industrials

3

Real Estate

4

Materials

5

Consumer Staples

Number of
companies
2
5
6
3
2
7
3
3
19

Total

167

Frequen
cy
98

Averaged FO
13.99%

301

13.56%

142

13.44%

182

12.77%

112
835

11.79%
13.18%

FO is the percentage of held by foreign
The
collected the data
shares

investors.
authors
from
in the duration from 2011 to 2015. The classification
was based on GICS and
firms relating to utilities and finance
were neglected. Source: Stata analysis
performed by the authors.

3.2.

Research hypothesis

This study was based on the work of Guo et al. (2015) which
was focused on the Big Bang Accounting Reform in Japan.
Hypothesis 1: Firms with high levels of the foreign ownership
involve in the management of the real earnings more than the
ones with low levels of the foreign ownership, which supports
the knowledge spillover theory.


Hypothesis 2: Firms with high levels of the foreign ownership
involve in the management of the real earnings less than the
ones with high levels of the foreign ownership, which supports
the asymmetric information theory.
3.3.

Description of variables

3.3.1.

Independent variables

The foreign ownership (FO) of the firm, which was studies in the
studies of Dahlquist et al. (2003) and Leuz et al. (2009), is the
ratio of the number of shares held by foreign investors (which
include private foreign investors, supporting funds and financial
institutions) to the total amount shares of a firm. For one firm,
this variable is the measurement of influences of foreign
investors on the domestic market at the end of each year. FO is
defined as
$%&'() +, -ℎ/)(- ℎ+01 '2 ,+)(345 356(-7+)!" =

8+7/0 5%&'() +, -ℎ/)(- +, 7ℎ( ,

9100

3)&
In addition, the regression model contains other variables
showing the economic
characteristics of the firm, which include:
The size (SIZE) of the firm: The size of the firm has an
important effect on the management of the real earnings
(Roychowdhury 2006). SIZE is given by:
<=>? = @5(8+7/0 /--(7-)
LEVERAGE: Previous studies showed that firms use the leverage
to avoid breaking
debt agreement when performing any activities relating the real
earning management (Dichev and Skinner, 2002). This action is
important since most of firms are supported by banks and

creditors can hold important roles in monitoring business activities
of firms. LEVERAGE is calculated as
@+54
7()&
03/'3037(8+7/0 /--(7@?C?DEF?

=
Return on assets (ROA): To control effects of financial
performances on the real


earning management, the regression model includes ROA which is
a reliable index to justify financial performances of firms and to
show efficiency in deploying assets.
=5G+&( '(,+)( (97)/+)135/)2 37(&D"E =

8+7/0 /--(7-


Growth of assets (GROWTH): Roychowdhury (2006) showed
that potential sustainable growth of assets can lead to substantial
changes in the real earnings management.
8+7/0 /--(7- 35 2(/) 7
−1
FD"H8I

8+7/0 /--(7- 35 2(/) 7
−1

=

Absolute values of discretionary accruals (ABS_DA): Cohen et al. (2008)
showed
the existence of the trade-off between the real earnings
management and the saving management. Therefore, ABS_DA
was included in their study. DA is estimated based on Kothari et
al. (2005) while ABS_DA is the remaining calculated from
MN
8EKL =

KLO

+ MP∆E<Table
2
Descriptions of variables in the equation of TA
Notation

Variables
Total
Accruals

TA

Equations
∆<ℎ+)7 7()& E--(7- − ∆ ∆Y(Z)(G3/73+5
W/-ℎ + Sℎ+)7 7()&
=56(-7&(57
8+7/0

E--(7-LOP − ∆<ℎ+)7 7()& 1('7 −

1

[

Total

\]]^_]`a

ass
ets

O[

??
<
E--(7LOP

∆SALES

LOP

Differe
nc e
sales
PP
E

RO
A

8+7/0 E--(7-K,LOP

Tangible
fixed
Assets
Return
on
Assets

8/543'0( !39(1 E--(7-

8+7/0 E--(7-LOP
=5G+&(
(97)/+)135/)2
37()&-'(,+)(
8+7/0
E--(7sLOP

Source: Authors’ summary.

According to Chung et al. (2004), the discretionary accruals can
help managers adjust the real earnings. This variable is,
therefore, used to increase the accuracy of the regression model
and to emphasise the importance of the income management.


Dummy variable SUSPECT: The study of Roychowdhury (2006)

pointed out that firms with 0 < ROA < 0.5% are more likely to
adjust the real earnings and are associated to suspicious cash
flow, low discretionary expenses and high production costs. Therefore,



0,

0 < D"E < 0.5%

1,
+7ℎ()Dummy variable LOSS: This variable was used in the
regress model because
Roychowdhury (2006) showed that small loss and slightly negative
variation in earnings can motivate firms to adjust their earnings to
satisfy their previous profit.
0,
D"E < 0
@"<< =
3.3.2.

1,
Dependent variables

D"E > 0

In this study, the variable of the real earnings management
were defined based on studies of Roychowdhury (2006) and
Cohen et al. (2008). These studies pointed out that the real

earnings management are defined by four indices which are
abnormal levels of operating cash flow (ACFO), abnormal
discretionary expenses (ADIS), abnormal production costs
(APROD) and a combined measure (COM_REM)
Abnormal levels of operating cash flows (ACFO): According to the study
of Roychowdhury (2006) which was based on Dechow et al.
(1998), managers can use excessive price discounts or credit
terms to boost temporary sales, which lead to the current period
earnings increase. However, current period cash flows may not go
up.
Abnormal discretionary expenses (ADIS) is a subjective cost that is not
necessary for business operation (Roychodhury 2006). This cost
was defined as the summation of advertising expenses, R&D
expenses, and selling, general and administrative expenses.
Although these expenses are not essential to business operation,
managers can use them to increase earnings by reducing current
expenditures.
Abnormal production costs are the third index used by
Roychowdhury (2006). The study of Pan (2009) showed that firms
try to increase their profit by reducing discretionary expenses
and cost of goods sold. Then, managers can increase production
more than being required to decrease average cost per unit,
which leads to lower the reported cost of goods sold (COGS) and
inflate operating profit margins. The excessive production also
leads to higher production and inventory holding costs. For a given
level of sales, this increases annual production costs and lowers
current period cash flows. The production costs were defined as


the summation of COGS and changes in inventory during the

year.


A combined measure (COM_REM): According to Cohen et al
(2008), a combined measure was defined as aggregating three
aforementioned variables: ACFO, ADIS, and APROD. In detail, this
variable is defined as ACFO + ADIS – APROD.
Table 3
Description of four indices relating to the variable of the real
earnings management (REM)
Notation

Previous studies

Description

ACFO

Roychowdhury (2006)

ADIS

Roychowdhury (2006)

Abnormal levels of operating cash
flows
Abnormal discretionary expenses)

APROD

Roychowdhury (2006)

COM_REM

Cohen et al. (2008)

Abnormal production costs
A
combined measure of the variable
REM:
ACFO + ADIS – APROD

Source: Authors’ summary.

This table listed original studies of the four indices and
estimated results in the following 3 equations by the OLS
regression model:
1
∆W!"L
+s ∗ +s ∗ +s ∗ +u 1
=
s
ELO
P

N

P

S

U

ELOP

L

ELOP

ELOP
1
Y=+s ∗ +s ∗ +
=
u
s
N
P
S
ELO
ELOP

(2)
L

ELOP

P

1
∆TD"YL
+s ∗ +s ∗ +s ∗ +
=
s
s
N
P
S
U
ELO
P

ELOP

ELOP

∆∗
+
u
w

ELOP

(3)

L

ELOP

The impact of foreign ownership regression for the four criteria
for the expression of
real earnings management. By equation:


D?y = sN + sP ∗ !" + sS ∗ <=>? + sU ∗ @?C?DEF? + sw ∗
D"E + sz
∗ E{<_YE + s} ∗ FD"H8I + s~ ∗ @"<< + s ∗ W8
+ uL (4)
The authors used the panel data, the Ordinary Least Square
(OLS) regression model,
the descriptive statistics and the correlation between variables
to uncover statistically meaningful variables. The authors
performed Durbin-Watson test, White test and Heteroskedasticity
test: Breuch-Pagan-Godfrey to correct the OLS regression. In
addition, the authors used the robust and the two-stage least
squares (2SLS) regression.


630 |
ICUEH2017

4. Research results
4.1.

Descriptive statistics and correlation analysis

Table 4
Foreign ownership in different years
Year

N

2011

167

2012

Mean
11.70%

Std.
dev
13.88%

Max
74.67%

167

12.13%

13.87%

2013

167

13.29%

2014

167

2015
Total

Q1

Q3

Min

1.12%

Media
n
6.04%

19.42%

0.00%

49.00%

1.35%

6.01%

18.19%

0.03%

14.37%

49.33%

1.47%

7.30%

19.88%

0.00%

14.37%

15.27%

49.85%

1.97%

8.39%

20.59%

0.03%

167

14.41%

15.08%

49.00%

2.26%

9.39%

21.19%

0.02%

835

13.18%

14.52%

74.67%

1.53%

7.34%

20.23%

0.00%

Source: Authors’s compilation by stata 12.0

Table 4 presents the variation of the foreign ownership in
different years. There were 167 firms each year in the sample of
data. The averaged ownership of foreign investors in the duration
from 2011 to 2015 was 13.18%. The averaged ownership was
increasing and reached 14.37& in 2015. In a comparison with Vo
(2015), the averaged foreign ownership of Vietnam increased
from 10% measured in 2007 to 51.3% measured in 2015. This
result showed that the foreign investors play a more important
role in firms in Vietnam.
Table 5 reports results of the statistics of all samples (835
observations). The mean values (MEAN) of ACFO, ADIS, APRO and
COM_REM were 0.000. These mean values are positive and much
smaller than expected value; in addition, they are very small that
was close to 0. Therefore, a good agreement can be concluded.
When using all samples, the mean and median of the foreign
ownership was estimated to be 13.25% and 7.3% respectively,
which shows that the distribution of the foreign ownership of
firms in Vietnam was unbalanced. The averaged value of
ABS_DA was 0.121. The mean and median of total assets was
159 billion nad 26 billion VND. The averaged long-term
liabilities was accounted for 12.2% of the total assets. 5.87% of

the total firms have negative profit and 9.82% of the total firms
were suspicious due to very small positive profit.


Table 5
Descriptive statistics of all samples
Variable

Ob
Mean
Std. Dev.
s
ACFO
83
0.000
0.151
5
ADIS
83
0.000
0.07
5
8
APROD
83
0.000
0.18
5
6
COM_REM

83
0.000
0.33
5
3
FO
83
0.132
0.14
5
5
SIZE
83
2.383
1.09
5
2
LEVERAGE
83
0.122
0.14
5
5
ROA
83
0.058
0.08
5
4
ABS_DA

83
0.121
0.15
5
5
GROWTH
83
0.127
0.32
5
2
LOSS
83
0.071
0.25
5
6
SUSPECT
83
0.098
0.29
5
8
Source: Authors’s compilation by stata 12.0

Min

Max

-1.480

0.720

-0.210

0.350

-0.670

0.000

1.1
90
1.2
10
0.750

0.240

5.540

0.000

0.670

-0.650

0.780

0.000

-0.690

1.3
10
3.660

0.000

1.000

0.000

1.000

-2.730

Table 6 describes the correlation matrices of all samples. In
addition to 4 dependent variables, there was weak correlation
between variables in the regression model. The variable FO
possessed the positive correlation with ACFO, ADIS and
COM_REM and negative correlation with APROD, which supported
the knowledge spillover theory that the foreign ownership
positively affects the real earnings management. The variables
SIZE and ROA had the strongest correlation with FO, 30.5% and
15.6% respectively. In general, there were good correlation
between ROA and other REM variables: 0.352 with ACFO, 0.184
with ADIS, -0.415 with APROD and 0.436 with COM_REM.


632 | ICUEH2017

4.2. The correlation matrix of the variables in the model
Table 6
Correlation matrix of variables
ACFO
ACFO

ADIS

APROD

1

COM_RE
M

F
O

SIZE

LEV

ROA

ABS_DA

GROW

LOSS

ADIS

0.122

APROD

-0.547

-0.464

COM_REM

0.791

0.545

-0.917

FO

0.130

0.270

-0.184

0.224

SIZE

-0.081

-0.023

0.069

-0.081

0.305

LEVERAGE

-0.031

-0.142

0.138

-0.125

-0.043

0.327

ROA

0.352

0.184

-0.415

0.436

0.156

-0.174

-0.238

ABS_DA

-0.157

-0.037

0.135

-0.155

-0.050 -0.027

0.025

0.006

GROWTH

-0.228

0.124

0.214

-0.196

0.014

0.128

0.150

0.110

0.450

LOSS

-0.045

-0.058

0.118

-0.100

0.009

-0.012

0.028

0.001

-0.084

1

SUSPECT

-0.122

-0.119

0.135

-0.158

0.112

0.116

0.173

0.402
0.228

-0.043

-0.080

0.091

SUS

1
1
1
1

Source: Authors’s compilation by stata 12.0

1
1
1
1
1

1


Tran Thi Thuy Linh et.
al. | 633

Table 6 reports the correlation between variables in the
research model. In addition to the four dependent variables, the
variables in the model are almost not strongly correlated. The
variable foreign ownership- FO is positively correlated with the
variables ACFO, ADIS, COM_REM but which is negatively

correlated with the variable APROD. It is a good indicator of the
knowledge spillover theory that foreign ownership has a positive
impact on income management.
Table 7
Determine the sample mean/median difference.
Variable

All
sample
Obs.

Mean

Low foreign

High foreign

Mean

ownership
group
Obs.
Mean

ownership
group
Obs.
Mean

difference

t-statistics

ACFO

835

0.000

466

-0.02

369

0.03

0.05***
(7.43)

ADIS

835

0.000

466

-0.01

369

0.02

0.02***
(5.72)

APROD

835

0.000

466

0.03

369

-0.04

-0.07*** (5.56)

COM_REM

835

0.000

466

-0.07

369

0.08

SIZE

835

2.383

466

2.1
3

369

2.71

0.15*
** (
6.59)
0.44***
(7.92)

LEV

835

0.122

466

0.1
3

369

0.12

ROA

835

0.058

466

0.04

369

0.08

ABS_DA

835

0.121

466

0.1
3

369

0.11

GROWTH

835

0.127

466

0.12

369

0.13

-0.006
(-0.67)
0.03**
* (
6.00)

-0.01* (1.28)
0.01
( 0.56)

LOSS
SUSPECT

835
835

0.071
0.098

466
466

0.07
0.1
3

369
369

0.07
0.06

-0.005
(-0.29)
-0.07*** (3.35)

This table examines the mean difference between low and high foreign
ownership groups. The variables are defined in section 3.3. Enterprises are
classified as high foreign ownership, if the foreign ownership is


greater than and equal to 10%; and the foreign ownership is low if vice versa.
*, **, and *** denote significance level of 10%, 5% and 1% respectively
Source: Authors’s compilation by stata 12.0

Table 7 shows the results of the mean/median difference
between the two groups of high, low and comparable foreign
ownership. The author finds that the average of abnormal cash
flows from operations (ACFO), abnormal discretionary expenses
(ADIS), and the comprehensive real earnings management
measure (COM_REM) are notably higher for the high foreign
ownership firms than the for the low foreign ownerships firms,
which indicates that high foreign ownership firms less likely
involve in real earnings management. In contrast, the
abnormal production cost (APROD) is significantly lower for high
foreign ownership firms than for low foreign ownership firms,
suggesting that a greater foreign investment base will be less
likely impacted by real earnings management. High foreign
ownership firms achieve the everage of SIZE, ROA, and GROWTH
higher than low foreign ownership firms.Therefore, we can
conclude that foreign investors prefer large firms with a high of
return on total assets also a high annual asset growth.
This study also finds that the average value of the absolute
value of discretionary accruals (ABS_DA) of the low foreign
ownership firms are higher than the high foreign ownership firms.
Assume that firms with more foreign investors are less likely to

use discretionary accurals for earnings management, which is
suitable for the findings of Chung et al. (2004). In addition, the
paper mentions that foreign investors hold disproportionate
shares in firms, they will hold more shares in firms with better
accounting performances and with low leverage, which is
consistent with the evidence from Kang and Stulz (1997).
Furthermore, the results show that high foreign ownership firms
have higher growth opportunities than low foreign ownership
firms. Moreover, high foreign invested enterprises are less likely
to be classified as suspected enterprises with earnings
adjustment activities.
4.3. Results of the OLS regression model
4.3.1.
Regression Results for 4 Real Earnings Management
(REM)
Table 8 presents the estimated parameters from the real
earnings management models by results of the OLS regressions.


Table 8
Summary of measurement of indicators showing real earnings
CFOt/At-1

DISt/At-1

PRODt/At-1

(1)

(2)

(5)

0.036*
**
(7.71
)
12.99*
**
(15.4
2)

-0.058***
(-7.49)
30.82*
**
(184.7
7)
0.967***
(1.46)

Equation
Intercept

0.02***
(2.41)

1/At-1
15.36***
(4.58)

-0.00
(-0.12)

St/At-1

St-1/At-1

0.03***
(8.44)

∆St/At-1

0.022
(0.72)

-0.03***
(-2.57)

∆St-1/At-1
R

0.011
(-5.57)

2

3.19%

28.82%

98.19%

Note: Estimated results in the following 3 equations by the OLS regression model:
%

%

%

= )* + )'
$&'

$&'

45/$

∗

1

%

= )* + )' %

$&'

∗

9:#
4$

+ )0 %

+ ).

$&'

$&'

∗

/$&'

+ ). %
∗
1

= )* + )' ∗

+ 2$ 1

$&'

∗

$&'

+ 2$ (2)
/$

+ ). ∗

+ )0

∆/

+ ); ∆/

∗

$

∗

+ 2$ (3)

$&'

the table shows the results of regression coefficient and the t test value in
parentheses.
*, **, and *** denote significance level of 10%, 5% and 1%
respectively Source: Authors’s compilation by stata 12.0

From the regression coefficient of equations (1, 2, 3), the
author re-estimates the dependent variable by the model, the
differences between the estimated data from the model and real


collected data are actual earnings management variables such as
ACFO, ADIS, APROD correspondingly with equations (1), (2), (3).

4.3.2.

Results of the regression model

Table 9
Results of the OLS regression model of foreign ownership and real
earnings quality.
Equation
_CONS

ACFO

ADIS

APROD

(4.1)

(4.2)

(4.3)

0.001

0.047

-0.054***

(4.4)
0.10
0.363***

FO

0.084**

0.140***

SIZE

-0.006**

0.002

-0.007

LEVERAGE

0.00
1
0.149***

-0.055***

0.000

ROA

0.745***

0.09
6
1.772***

ABS_DA
GROWTH

0.04
-0.127***

LOSS

0.052***

SUSPECT
OBS.

-0.026
835

-0.137***

COM_REM

0.061
*
-0.055***

-0.965***

0.044***

0.141***

-0.225***

0.019
0.031

0.06
6
-0.067**

0.00
0.01
835

0.021

835

-0.116*

835

22.1
13.4
26.2

28.3
%
%
%
%
Note: This table summarizes the results of foreign ownership regression for
the four criteria for the expression of real earnings management. All variables
are defined in section 3.3. By equation:
:=> = )* + )' ∗ "# + ). ∗ /5?= + )0 ∗ @=A=:%B= + ); ∗ :#% + )C ∗ %D/_4%
+ )F
R

2

∗ B:#GHI + )J ∗ @#// + )K ∗ /L/9=!H + 2$ (4)
the table shows the results of regression coefficient and the t test value in
parentheses.
*, **, and *** denote significance level of 10%, 5% and 1% respectively
Source: Authors’s compilation by stata 12.0

The research performs regression based on the OLS regression
model examines the relationship between foreign ownership and
real earnings management according to equation (4), and
according to different real earnings management solutions. Table
9 reports the results of the regression analysis. Model 4.1
(column 1), the study uses the abnormal cash flows from
operations (ACFO) as the dependent variable, the ACFO model is
calculated based on the foreign ownership function and other

characteristics of the companies. Test statistics and significance
levels are based on the standard errors,


adjusted by clustering at the firms and year levels. Consequently,
our standard error is conservative, since explanatory variables are
determined at the firm level (Bertrand et al., 2004). The
coefficient of foreign ownership (0.084) is positive at fivepercent level, supporting the hypothesized knowledge that high
foreign ownership firms will reduce the adjustment activities that
are related to the cash flow. The results also suggest that the
coefficients on SUSPECT are negative (-0.026). This is
economically significant and is consistent with the findings in
Roychowdhury's (2006) indicate that firms tend to encroach on
sales and reduce the abnormal cash flows from business
operations.
Model 4.2 (column 2), the research investigates the real
earnings management through the abnormal discretionary
expenses (ADIS), also shows that high foreign ownership firms
are less likely involved in real earnings management by reducing
the abnormal discretionary expenses with the coefficient of 0,14
at the 1% significance level. Equation 4.3 (column 3) reports the
result of the abnormal production cost (APROD). The result
shows that the coefficients of foreign ownership are negatively
related to abnormal production costs. The coefficient of foreign
ownership is -0.137, at the 1% level presents a high foreign
ownership, thus, the managers are less likely boosting production
costs. This result further supports the knowledge spillover
hypothesis
by
showing

that
foreign
investors
limit
management’s
overproduction
and
opportunistic
sales
manipulation. Equation 4.4 (column 4) shows the regression
results using the combined score COM_REM as the dependent
variable. In further support of the knowledge spillover hypothesis,
the coefficient on foreign ownership (0.363) is significantly
positive at the 1% significance level. These outcomes show
positive connections between foreign ownership and real earnings
management.
4.3.3.

Verification of the robustness of the OLS regression

According to a study by Jun Gou (2015), the FO variable is
likely to encounter endogenous problems with variables in the
research model. Therefore, the author performs variance testing
changes and endogenous FO explanatory variable. In addition, to
solve this problem, the author re-examines the model with the
Robust regression to overcome variance variation and to
conduct the 2SLS regression. The results of the stability test of
the model: