Professional Documents
Culture Documents
61
Rita Yuliana: The Islamic capital …
vestment decision must pass rigorous assess- risk in order to get the flow of resources to the
ments. The Financial Accounting Standard does appropriate qualified companies only and market
not limit the selection of accounting policies and mechanisms to operate efficiently and effectively.
procedures chosen by the manager to report the In relation to the above description, it is still
earnings. Moreover, the process of selecting ac- limited for the empirical testing on the market
counting policies and procedures are complex and response to the actions of real earnings manage-
challenging, not just based on considerations ment. For example, in Brazil Cupertino, Martinez
about providing high quality information for in- & da Costa Jr. 2015) found that the market did not
vestors. Therefore, investors have to analyze the understand the risks of such action. This is evi-
earnings quality carefully. denced from the market response that was posi-
One important thing for analyzing the quality tive on companies conducting earnings manage-
of earnings is the level of the aggressiveness of ment through sales activities and production.
REM because according to Roychowdhury (2006) Based on the above phenomenon and empiri-
REM reduces earnings quality. Therefore, the in- cal findings, this study examines the existence of
vestors should understand that consequence in real earnings management in Indonesia capital
order to get good return. Generally, earnings market, followed by the test of differences be-
management is an accounting policy choice that is tween the Islamic capital market and conventional
used in order to pack earnings to make it look capital markets. This research aims to show the
attractive. Therefore, investors should have high differences in the quality of corporate earnings in
effort to analyze the earnings in order to get a reli- both types of capital markets. The study also iden-
able assessment of earnings quality. tifies the company that conducted real earnings
The statement is an important argument that management by using techniques Earnings Distri-
the market should respond wisely because the bution Analysis (EDA) as the study Roychowd-
company's future earnings are not affected by the hury (2006) and Gunny (2010). The technique re-
form of earnings but by the essence of the earnings fers to the Prospect Theory arguments by Kahne-
information itself. Therefore, in order to obtain man & Tversky (1979).
good information, the market should examine the The next test is to show the market response
quality of earnings information, including meas- over the aggressiveness of real earnings
ures of earnings management. There are arguments management on each type of capital markets.
underlying the study of market response on the Those responses were seen through the value of
earnings management. For example, the measure Cumulative Abnormal Return (CAR) around the
reduces the quality of earnings and threatens the date of publication of earnings. Purpose of the test
company’s value. By doing so, the market imposes is to show the level of investors’understanding on
sanctions on companies doing earnings manage- the quality of earnings. Really, earnings respond-
ment, e.g., they should not buy their shares. ed better quality than the earnings that are less
Recent studies identify one form of earnings qualified.
management based on real activity, not just an By assuming the efficient capital markets, this
accrual (accounting) manipulation. In this case, research also examines the capital market efficien-
Mulford (2002, p. 144) stated that real earnings cy in Indonesia. If the market responses to good
management is the management of real activity earnings quality, it can be said that the market is
originating from business decisions that manage- aware of, understand and respond properly to the
ment used to improve earnings. Another propo- information so that it can be concluded that the
nent is Roychouwdhury (2006) who stated that Indonesian capital market semicircular strong.
such measures have a direct impact on cash flow. The evidence is important significance for issuers
In addition such measures also have their indirect to encourage them to present quality financial
impact on the accrual. information. That understanding also implies an
Real earnings management is new and consi- effort to improve the competency of accountants
dered more secure than earnings management and auditors as the authorities producing financial
based accrual although they have the same big statements.
risk for companies and investors. This cause real
earnings management to change the company's 2. THEORETICAL FRAMEWORK AND HYPO-
normal business. Besides that it can affect the val- THESES
uation of the company's performance. Therefore, Real Earnings Management
the market should determine the actions of the According to Roychowdhury (2006), there are
62
Journal of Economics, Business, and Accountancy Ventura Vol. 20, No. 1, April – July 2017, pages 61 – 69
three types of real earnings management as inde- This study used two techniques of data analy-
pendent variable such as real earnings manage- sis, i.e. regression and difference test. Regression
ment through increased sales (MLR sales), in- was used to analyze the influence of the status of
creased production (MLR Production), and reduc- the company, namely the suspect/non-suspect
tion of discretionary costs (MLR Cost Discretio- and ICM/non ICM on the level of real earnings
nary). The value of real earnings management management. Multiple regression models were
through each of these real activities was obtained used to include some control variables, namely the
from the difference between their respective real size of the company (SIZE), earnings (Net Income)
activities with the real normal activities. Normal and Return on Assets (ROA). Those variables are
activity is calculated by using the model formu- proxy of some of the company's fundamental per-
lated by Roychowdhury (2006) in each group-year formance.
industry. The following model to calculate each of The second analysis technique is a difference
the normal activities: test used to analyze the CAR differences between
a. Normal Sales (NS) the two groups of samples based on the level of
𝑁𝑆𝑡 1 𝑆𝑡 𝛥𝑆𝑡 aggression in real earnings management. The di-
= 𝛼0 + 𝛼1 + 𝛽1 + 𝛽2 +
𝐴𝑡−1 𝐴𝑡−1 𝐴𝑡−1 𝐴𝑡−1
vision level of aggressiveness of real earnings
𝜀𝑡 . (1)
management is determined based on the ranking
Where:
of the value of each real earnings management
𝑁𝑆𝑡 = operation cash flow in year t
activities, namely the top 25% (Q1) with 25% in
𝐴𝑡−1 = assets in year t-1
the bottom (Q4). The test is also applied to two
𝑆𝑡 = sales year t
sample groups, namely non ICM and ICM to indi-
𝛥𝑆𝑡 = difference between sales in year t to the
cate the market response on real earnings man-
year t-1
agement in both types of capital markets.
b. Normal Production (NP)
𝑁𝑃 𝑡 1 𝑆𝑡 𝛥𝑆𝑡
= 𝛼0 + 𝛼1 + 𝛽1 + 𝛽2 + 3. RESEARCH METHOD
𝐴𝑡−1 𝐴𝑡−1 𝐴𝑡−1 𝐴𝑡−1
𝛽3
𝛥𝑆𝑡−1
+ 𝜀𝑡 . (2) Population and Sample
𝐴𝑡−1 The samples consist of companies listed in Indo-
Where: nesia Stock Exchange in 2011 and 2012. They were
𝑁𝑃𝑡 = cost of goods sol + Δ inventories at year t drawn from all kinds of industries except the fi-
𝐴𝑡−1 = assets in year t-1 nancial and banking industry. In addition, the
𝑆𝑡 = sales year t samples were also restricted to companies that
𝛥𝑆𝑡 = difference between year t to the year t-1 have a complete financial data, no corporate ac-
𝛥𝑆𝑡−1 = difference in sales between years t-1 tp tion, active stock data, as well as the financial
year t-2 statements are presented in nominal amount.
c. Normal Discretionary Expense (NDE) Samples were grouped into two, namely a
𝑁𝐷𝐸𝑡 1 𝑆𝑡−1
= 𝛼0 + 𝛼1 + 𝛽1 + 𝜀𝑡 . (3) representative sample of Islamic Capital Market
𝐴𝑡−1 𝐴𝑡−1 𝐴𝑡−1
Where : (ICM) and a representative sample of the Conven-
𝑁𝐷𝐸𝑡 = discretionary expense (research and tional Capital Market (CCM). Samples of ICM
development cost+ advertising cots+ sales cost, represented by issuers included in the List of Is-
general, and administrative) in year t lamic Securities (LIS), while samples of CCM are
𝐴𝑡−1 = assets in year t-1 issuers that are not included in DES. Samples were
𝑆𝑡−1 = sales year t-1 also differentiated between companies suspected
of doing real earnings management (suspect) who
Cumulative Abnormal Return (CAR) do not. The determination of the suspect compa-
Measurement of CAR as dependent variable in nies was based on the ratio of earnings per asset
this study use abnormal return which is obtained beginning of the year, which is between 0 and
from the difference of the stock return with the 0,005 as research Roychouwdhury (2006).
expected return, for 3 days i.e. t-1, t0 and t + 1. The This study uses secondary data such as finan-
calculation of expected return uses a market mod- cial data and stock data. They were obtained from
el. Related to the calculation of estimated parame- the Financial Statements, Indonesia Stock Exchange
ters for calculating the expected return, this study database, www.yahoo.finance.com, and Osiris.
uses a range of observations around the publica-
tion date of the financial statements, i.e., t-60 and Research Variables
t-10. The variables of this research are:
63
Rita Yuliana: The Islamic capital …
Table 1
Samples Description
Industries ICM Non ICM Total
Agriculture 17 7 24
Mining 26 11 37
Basic Industry and Chemicals 50 23 73
Various Industries 26 17 43
Consumer Goods Industry 33 9 42
Property Real Estate and Construction 55 11 66
Utilities and Transportation Infrastructure 22 25 47
Trade in Services and Investment 84 33 117
Total 313 136 449
Table 2
Descriptive Statistics
Variables (Mean) ICM Non ICM t Stat Total Samples
CAR -0.004 0.009 -31.253 0.000
MLR Sales 0.018 -0.016 -12.788 0.008
MLR Production -0.145 0.763 -7.487 0.130
MLR Cost Discretionary 0.015 -0.029 -10.521 0.001
Size 4.125 4.131 90.652 4.127
NI 0.025 -0.056 -11.253 0.000
ROA 0.006 -0.012 -13.204 0.000
N 313 139 449
64
Journal of Economics, Business, and Accountancy Ventura Vol. 20, No. 1, April – July 2017, pages 61 – 69
25
20
15
10
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Figure 1
EDA Illustration
ble 2. The table shows the average value. In gener- shown from the average value of three of the op-
al, both groups of samples are significantly differ- posite sign. The table shows that the ICM group
ent at α level of 5%. General financial performance did MLR in Sales and MLR in Discretionary Ex-
of each sample group appears on the SIZE varia- pense more aggressive than non-ICM group. The
ble, NI, and ROA. Company size (SIZE) of ICM opposite occurs in MLR in Production.
group is smaller than the non-ICM group. How- In general, the level of aggressiveness in real
ever, earnings performance (NI) and profitability earnings management in the ICM group was
(ROA) of ICM group is better than the non-ICM higher than non-ICM group. This empirical find-
group. ing becomes one of the important arguments con-
An overview of the market response (CAR) in cerning the finding that the CAR in the ICM group
each group showed that the market responded is negative. It means that the market understands
negatively on the earnings information in the ICM that the companies included in the ICM conduct
group, otherwise the market response to the earn- real earnings management aggressively and they
ings information non ICM group is positive. Based are punished by the market in the form of a nega-
on the positive accounting theory, this fact shows tive CAR. In contrast, markets reward companies
preliminary evidence that there are differences in to the non-ICM (in the form of positive CAR) on
response to market on the earnings information in actions did not do real earnings management ag-
both types of markets, which, of course requires gressively.
further explanation. Based on these facts, there is This research identifies the companies that
prediction that the market also uses the status of were alleged to do real earnings management
the company as important information in making (suspect-firm) through EDA techniques (Roy-
investment decisions. chouwdhury 2006; Gunny 2010; Aflatooni & Mo-
Table 2 also presents a description of the level karomi 2013; and Chen, Lin & Weng 2013). Firms
of aggressiveness in real earnings management. In that suspected of real earnings management re-
general, there are differences in the level of ag- ported positive earnings despite the small value.
gressiveness of real earnings management in each Some of these studies showed that suspect-firm
sample group, both for MLR in Sales, MLR in Pro- conducted real earnings management (Roy-
duction, or MLR in Discretionary Expense. This is chouwdhury 2006 and Gunny 2010).
65
Rita Yuliana: The Islamic capital …
Table 3
The Results of Regression Test
MLR Sales MLR Production MLR Discretionary Expense
Intercept -0.158*** 7.683*** 0.181***
Suspect -0.037 -0.535 -0.073***
ICM 0.029*** -1.007 0.032***
SIZE 0.035*** -1.654*** -0.048***
NI 0.025 1.389 0.152***
ROA 0.153*** -1.224 -0.032
F 5.956 3.312 15.277
Adj R Square 5.24% 2.51% 13.74%
; ; and ***, significant at α 10%; 5%; and 1%.
* **
Table 4
Different Test Results: CAR
Q1 Q4 t Stat Entire Samples
MLR Sales 0.007 -0.011 -14.399 -0.001
MLR Production 0.018 0.003 -13.951 0.011
MLR Discretionary Expense 0.008 0.010 -14.137 0.009
66
Journal of Economics, Business, and Accountancy Ventura Vol. 20, No. 1, April – July 2017, pages 61 – 69
Table 5
CAR Difference Test Results: ICM and non ICM
Q1 Q4 t Test
ICM:
MLR Sales -0.001 -0.002 -2.267***
MLR Production 0.000 0.025 2.185***
MLR Discretionery Exp. 0.002 0.003 -1.690***
Non ICM:
MLR Sales 0.065 -0.011 1.158
MLR Production 0.060 -0.012 -0.706
MLR Discretionery Exp. -0.016 0.056 2.107***
; ; and ***, significant at α 10%; 5%; and 1%.
* **
straight increase in the frequency of reporting of real earnings management. It also shows that
small positive earnings. On the other hand, it was ICM variable is in a positive significance in the
found rarely the frequency of reporting small neg- regression where the dependent variable is Sales
ative earnings. Based on these findings, the com- MLR and MLR Discretionary fee. In the test re-
panies that are in a group of 16 allegedly real sults in which the MLR sales as the dependent
earnings management (suspect), whose evidence variable, the coefficient of ICM is positive signifi-
requires further analysis as part of further discus- cant indicates that the ICM group are more ag-
sion. gressive in conducting real earnings management
This section describes the findings about the than non-ICM firms. Furthermore, empirical evi-
determinants of the real earnings management dence also shows that the coefficient of ICM in
which are derived from the status of the company, the regression test of MLR Cost Discretionary
namely suspect/non suspect and ICM/non ICM. show a positive and significant value as well. It
Based on the identification of suspect-firm using implies that the ICM group did MLR Discretio-
EDA techniques, it is necessary to prove that the nary expense more conservatively than non-ICM
firm is conducting in real earnings management. group.
Regression test results as shown in Table 3 indi- The coefficients have opposite signs with
cate that the suspect-firm significantly perform Suspect coefficient, meaning that the behavior of
real earnings management through reduction of the two groups differ in conducting MLR Discre-
discretionary expense, while the two other real tionary expense. In addition, the company's status
earnings management activity is not the case. Re- as a member of the Islamic capital market becomes
gression coefficients for Suspect variables in the one of relatively strong argument in explaining
equation of real earnings management through the variation in real earnings management. There-
reduction of discretionary expense, showed a neg- fore, this study complements the use phenomenon
ative value (-0.073). The meaning of these values is of Prospect Theory in real earnings management.
that the suspect-firms have a tendency to do real This section identifies the results in real earn-
earnings management through reduction of dis- ings management. The real earnings management
cretionary expense more aggressively than the is affected by the status of the company, whether
non suspect-firms suspect/non Suspect and ICM/non ICM, this
Empirical evidence shows that the use of study examines the level of investor awareness of
Prospect Theory to explain the motivation of real the risks in the real earnings management actions.
earnings management is only valid on the condi- Before testing the market response to the actions
tion of MLR through reducing discretionary ex- of real earnings management in each status, it first
pense, while the other two types MLR (sales and has served the market response to the overall ac-
production) requires further explanation. There- tions of real earnings management, as shown in
fore, this study includes the company's status as a Table 4. It shows the results of different test mar-
member of ICM as a determinant of the real earn- ket response, which is measured by CAR.
ings management. The samples are distinguished by the level of
Regression test results showed further evi- aggression in real earnings management con-
dence regarding the effects of the status of com- ducted. Q1 group is a group of companies that
panies included in the ICM to the aggressiveness conduct real earnings management aggressively.
67
Rita Yuliana: The Islamic capital …
The group was 25% (112 firm-years) top ranks. that the Sales MLR is a business strategy that
Instead, Q4 was 25% firm-years which are at the should be appreciated and its practice applied in
lowest rank. Thus, both have different levels of companies with sharia label.
aggressiveness in real earnings management. This conclusion is supported by the results of
Table 4 shows that there are significant CAR different test CAR on each sample group which is
differences between the groups of Q1 to Q4 in all distinguished by their status, ICM and non ICM,
of real earnings management activities. However, as shown in Table 5. The results of the different
these differences have a different interpretation tests show that in general, ICM respons differently
when it is associated with the understanding that three activity MLR based on the aggresiveness,
real earnings management reduces the quality of while the conventional capital market is only able
earnings. According to this understanding, the to distinguish aggressiveness of MLR Discretio-
CAR in Q1 group should be lower than the Q4 nary expense only.
groups. Table 5 shows that in the Islamic capital mar-
Empirical evidence shows that this only ap- ket, the average of CAR in the sample MLR ag-
plies to the CAR different test results on samples gressive selling (Q1) is higher as compared to the
of MLR Discretionary fee. It means, that the mar- samples with levels that are not MLR of aggres-
ket understands the risks arising from the MLR sive sales (Q4). The same was found in samples of
discretionary expense. Table 4 shows the average non ICM although the differences were not signif-
of CAR in the group Q1 is lower than the group icant. The empirical fact is in line with previous
Q4, with a significant difference. These results are findings, namely that the market responded posi-
consistent with previous regression test results tively to the MLR Sales. Thus, these findings rein-
which prove that the cause of the MLR Discretio- force the conclusion that, with the assumption of
nary expense is an opportunist action by manager efficient capital markets and the positive account-
that turn a negative earnings into a positive earn- ing theory, MLR Sales does not reduce the quality
ings (see Table 3). Thus, these findings are consis- of earnings because the market actually appre-
tent with Capital Market Hypothesis and Positive ciates the action.
Accounting Theory. Further empirical evidence is that the Islamic
In different tests on Sales MLR and Produc- capital market gives sanction on MLR Production
tion MLR, the empirical facts show precisely the which is shown the value of CAR in Q1 (0.000)
opposite result. CAR in Q1 group is higher than lower than the CAR in Q4 (0.025), with significant
the CAR in Q4 group. The findings are also consis- differences. The evidence shows that the Islamic
tent with previous findings that show empirical capital markets understand that the MLR produc-
evidence that in the Indonesian capital market, tion is somethings that pose a risk it feasible to get
Sales MLR and Production MLR were not affected a penalty. The fact is different with different test
by the company's status as a suspect/non suspect results on samples using non-ICM firms, because
that the assertion that real earnings management it turns the conventional capital market did not
reduce earnings quality unsupported empirical prove any CAR differences between sample Q1 to
evidence. Based on these facts, it is natural for the Q4.
market to respond MLR Sales and Production The empirical evidence of market response on
MLR as things that are not bearing risk and do not MLR of Discretionary Expense shows solid evi-
deserve the penalty of the market. dence that it reduces the quality of earnings, both
Especially for testing on samples of MLR in Islamic and conventional capital markets. This
Sales, it shows interesting results. CAR with dif- is supported by the CAR in the group Q1 lower
ferent test results shows that the CAR in compa- and significantly different from the CAR in Q4
nies that conduct sales MLR aggressive is positive group. From the three types of MLR, MLR of Dis-
and well above the average of the entire sample. cretionary Expenses, consistently, show empirical
The significance of these findings is that the mar- evidence that supports the conclusion. It is stated
ket actually appreciates the Sales MLR action. The that the real earnings management reduces the
findings are consistent with empirical facts on the quality of earnings. This finding is consistent with
previous regression test that Sales MLR is not af- the findings of Cupertino, Martinez & da Costa Jr.
fected by opportunistic behavior of managers (as (2015) which showed empirical evidence that in
evidenced from Suspect variables were not signifi- Brazil, the market responds correctly only real
cant), but it is conducted by a group of companies earnings management actions on MLR Discretio-
included in the DES. Thus, the market considers nary fee.
68
Journal of Economics, Business, and Accountancy Ventura Vol. 20, No. 1, April – July 2017, pages 61 – 69
69