INTRODUCTION
Tax is one of the sources of state revenue that has
the largest contribution in the 2020 state budget posture, where in 2020 the
tax sector contributes 78.9% percent of total revenue. Various tools used in
supervisory activities include sectoral taxpayer supervision and transfer
pricing. But most taxpayers still consider tax collection a burden that must be
avoided because economically taxes reduce taxes. Studies have shown that tax
avoidance can reduce company costs and increase shareholder wealth (Hanlon & Heitzman, 2010). Thus, to
determine how much tax avoidance action is required, companies need to exchange
the marginal benefits of tax management for the marginal costs of doing so (Chen, Chen, Cheng, & Shevlin, 2010). One of the
marginal benefits is greater tax savings, while the marginal costs include
potential penalties imposed by the tax administration, implementation costs
(time/ effort and transaction costs of implementing tax transactions), and
agency costs that inevitably accompany tax avoidance activities.
Tiaras and Wijaya (2015)
shows that earnings management has a significant effect on corporate tax
avoidance, in line with the findings of (Amidu, Coffie, & Acquah, 2019; Pajriyansyah & Firmansyah, 2017;
Suyanto & Supramono, 2012). However, this is
different from the research conducted by (Henny, 2019; Syanthi, Sudarma, & Saraswati, 2013)
stated that the old management had no significant effect on tax avoidance. Due
to the differences in the results of the test of this variable, the authors
took this variable to re-confirm the research conducted previously in addition
to the indications of earnings management being used as a tool for tax planning
because managers try to manage earnings with tax motivation
(Scott, 2015).
Profitability is often used as an indicator and
consideration not only used by investors in investing in a company but also the
Directorate General of Taxes to supervise taxpayers and estimate the amount of
tax payments of a company. Based on research conducted by (Jaffar, Derashid, & Taha, 2021; Kim & Im, 2017)
that profitability affects tax avoidance because companies that earn higher
profits pay lower tax rates by doing more planning to reduce the tax burden.
However, this study is different from the results of research conducted by Sitorus (2020)
which states where ROA has no effect on tax avoidance because the level of
profit will make management more conservative towards tax management because
the risk of cost and time sacrificed is irrelevant to the findings if Tax
Complience is carried out. With the differences in research, the author will
re-confirm this variable.
Leverage in a
business context it means borrowing capital for business purposes so as to
obtain optimal ROI (Return on Investment). According to Sjahrial (2009),
leverage means the use of assets and sources of funds by companies that have
fixed costs (fixed expenses), so that the source of funds comes from loans
because loans have interest as a fixed expense with a view to increasing
shareholder potential profits. Thus, DER ratio (Debt to Equity Ratio) for the
effect of leverage on tax avoidance because DER is used as one of the
analytical tools carried out by the fiscus to measure the fairness of the
taxpayer's financial statements attached to the Annual SPT report.
Transfer Pricing is a company policy in determining the transfer price
of company transactions in the form of goods (tangible and intangible) and
services. Transfer pricing is one of Permatasari and Trisnawati (2022)
tax planning strategies. Transfer Pricing practices are carried out in two
groups of transactions, namely intra-company and inter-company transfer pricing
and can be carried out domestically and internationally involving MNC (Multi
National Company) and domestic companies.
According to Hansen and Mowen (2007), transfer pricing
is a special selling price set in exchange between divisions recording the
revenues and costs of a division. Regulations on transfer pricing in general
have been regulated through Article 18 of Law No.36 of 2008 concerning income
tax wherein Article 18 paragraph (3) of the Income Tax Law states that the
Directorate General of Taxes (DGT) has the authority to re-determine the amount
of Taxable Income (PKP) for Taxpayers who have special relationships with other
taxpayers in accordance with the fairness and normality of business affected by
special relationships (arm's length principle).
The existence of conglomerates and group companies
makes this practice easy to do because of the special relationship so that the
determination of unreasonable transfer prices between companies is easy to do
with profit shifting which in the end saves the group's tax payments. Moreover,
transfer pricing has an effect on tax avoidance (Richardson, Taylor, & Lanis, 2013; Sari & Ajengtiyas, 2021).
Tax avoidance as an action to reduce tax obligations
carefully by using loopholes in the provisions of tax law (Jacob & Schütt, 2020). Tax avoidance
practices that are widely used include transfer pricing, the use of debt
instruments and earnings management. Given the importance of the role of taxes
to finance development, it requires optimal supervision by the Directorate
General of Taxes as an institution tasked with securing tax collection
according to the targets planned in the APBN.
This study takes a sampling of companies engaged in
the P3 sector (Plantation, Forestry and Mining) which are listed on the IDX.
The rationale for taking this sampling is because the P3 sector (Plantation,
Forestry and Mining) will support tax revenues in 2021 and there has been a
boom in commodity prices lately, even though it is still in a state of the
COVID-19 pandemic. Based on the description above, it leads to the importance
of revisiting the factors that influence tax avoidance and the purpose of this
study is to find some empirical evidence, (1) analyzing
the influence of Earnings Management on Tax Avoidance of companies that are
included in the P3 sector (Plantation, Forestry and Mining) listed on the IDX
in 2017-2021, (2) analyzing
the effect of Profitability on Tax Avoidance of companies that are included in
the P3 sector (Plantation, Forestry and Mining) listed on the IDX in 2017-2021, (3) analyzing the influence of
Leverage on Tax Avoidance of companies that are included in the P3 sector
(Plantation, Forestry and Mining) listed on the IDX in 2017-2021, (4) analyzing the effect of
Transfer Pricing on Tax Avoidance of companies that are included in the P3
sector (Plantation, Forestry and Mining) listed on the IDX in 2017-2021, and (5) analyzing the Effect of
Earnings Management, Profitability, Leverage and Transfer Pricing on Tax
Avoidance of companies that are included in the P3 sector (Plantation, Forestry
and Mining) listed on the IDX in 2017-2021 simultaneously.
METHOD
The
approach used in this research is descriptive quantitative by analyzing the
data using statistics and then describing the data that has been collected
through the interpretation of the data presented. In this study, the authors
conducted an analysis of the determinants (Earnings Management, Profitability,
Leverage and Transfer Pricing) selected for the problem taken (Tax avoidance)
from secondary data collected on the business sector selected by the
researcher. ETR (Effective Tax Rate) is used as an indicator to measure tax
avoidance in the P3 sector (Plantation, Forestry and Mining).
In
this study, the population is companies that do business in the P3 sector which
are listed on the IDX. From data collection for companies engaged in this
sector, there are 83 companies that have been listed before and between 2018
and 2021 (Sugiyono, 2019).
The
type of data used in this study is panel data which is a combination of cross
section data and time series data, where the same cross section unit is
measured at different times. Data processing and testing was carried out
statistically using the Eviews 12 program. Researchers used secondary data,
namely company financial data that was publicly released through the IDX
website, namely https://www.idx.co.id/usaha-tercatat/laporan-keuangan-
and-annual/ selected from sectors related to P3 (Plantation, Forestry and
Mining).
Descriptive Analysis
Descriptive
statistics provide an overview or description of a data seen from the average
value (mean), standard deviation, maximum, and minimum. In this study,
descriptive statistical analysis focused on the maximum, minimum, mean, and
standard deviation values.
Verification Analysis
This analysis
aims to determine the relationship between two or more variables, or the method
used to test the truth of a hypothesis Sugiyono (2019) where
in this study hypothesis testing uses panel data regression. Before calculating
the panel data regression, first conducted a test in the panel to use the most
appropriate model in testing the hypothesis using panel data regression.
Panel Data Test
1. Chow test
The Chow test is used to determine
whether the panel data model is regressed with the common effect model or the
fixed effect model (Widarjono, 2017). The hypotheses in this test are as
follows:
Ho : Common Effect Model
H1 : Fixed Effect Model
Information :
a) If the probability value of Chi-square Cross-section
<0.05; then Ho is rejected
b) If the probability value of Chi-square
Cross-section> 0.05; then Ho is accepted
2. Hausman test
The Hausman test
is used to determine whether the panel data model is regressed with a fixed
effect model or with a random effect model (Widarjono, 2017). The hypotheses in
this test are as follows:
Ho : Random Effect Model
H1 : Fixed Effect Model
Information :
a) If the
probability valueCross-section Chi-square< 0.05
; then Ho is rejected
b) If the
probability valueCross-section Chi-square> 0.05
; then Ho is accepted
3. Langrange
Multiplier Test
This test is used to determine
whether the panel data model is regressed with the modelcommon
effector by modelrandom effects (Widarjono,
2017). The hypothesis in this test is as follows:
Ho : Random Effect Model
H1 : Common Effect Model
Information:
a)
If the breusch-pagan probability value < 0.05 ;
then Ho is rejected
b) If the
breusch-pagan probability value > 0.05; then Ho is accepted
Classic assumption test
Classical
assumption test is used to assess the presence or absence of bias on the
results of the regression analysis that has been carried out. The classical
assumption test consists of normality, heteroscedasticity, autocorrelation and
multicollinearity tests (Ghozali, 2013).
1. Normality test
Normality test aims to test
whether in a regression model, the dependent variable, the independent variable
both have a normal distribution or not. Decision making regarding normality is
as follows:
a) If p < 0.05 then the data distribution is not
normal
b) If p > 0.05 then the data distribution is normal
2. Heteroscedasticity Test
Heteroscedasticity test aims to
test whether in the regression model there is an inequality of variance from
the residual of one observation to another observation. If the significance
probability is above the confidence level 0.05 then it does not contain
heteroscedasticity.
3. Autocorrelation Test
The
test that is often used to determine whether or not autocorrelation occurs was
developed by statisticians Durbin and Watson known as the Durbin Watson (DW) d
statistic test. DW test is done by making a hypothesis:
a) Ho
: no autocorrelation ( r = 0 )
b) Ha
: there is autocorrelation ( r 0 ).
The
basis for decision making is as follows:
a) If
DW < DL, then there is a positive autocorrelation
b) If
DL < DW < DU, then there is doubt that autocorrelation occurs
c) If
DU < DW < 4-DU, then there is no autocorrelation
d) If
4-DU < DW < 4-DL, then there is doubt that autocorrelation occurs
e) If
4-DL < DW, then there is a negative autocorrelation.
Information
: DL =
lower limit DW
DU = upper limit DW
4. Multicollinearity Test
Multicollinearity Testaims to
determine whether the regression model found a correlation between the
independent variables (independent). In this study, the VIF value was used to
determine whether in the regression model there was a relationship between the
independent variables. If there is no independent variable that has a VIF value
> 10, then in the regression model there is no multicollinearity problem.
Hypothesis testing
Hypothesis
testing was carried out to prove the effect of the studied variables. The
hypothesis tests used were regression analysis test, T test, F test and the
coefficient of determination (R2).
1. Panel Data Regression Equation
Effect of Earnings Management (EM) (X1), Profitability (Prof) (X2), Leverage
(LEV) (X3) and Transfer Pricing (TP) (X4) on Effective Tax Rate (ETR) (Y)
Panel data regression analysis is
used to determine how much influence the independent variable has on the
dependent variable. The multiple regression equation used is as follows:
2. F test
The F statistic test or the
feasibility test or Goodness of fit is used to determine whether the
independent variable simultaneously affects the dependent variable with a 95%
confidence level (α = 0.05). Simultaneous
research hypotheses as follows:
H0 : b1...b4 = 0; Profit management (EM) (X1), Profitability (Prof)
(X2), Leverage (LEV) (X3) and Transfer Pricing (TP) together not significant effect on Effective Tax Rate (ETR) (Y);
H0 : b1...b4 ≠ 0; Profit management (EM) (X1), Profitability (Prof)
(X2),Leverage (LEV) (X3) and Transfer
Pricing(TP) together significant effect on Effective Tax
Rate (ETR) (Y).
With the decision-making conditions:
a)
Prob value (F statistic) <
0.05 (significance level 5%), then H0 is rejected, which means that the
independent variables have a significant influence on the dependent variable
together.
b)
Prob value (F statistic) >
0.05 (significance level 5%), then H0 is accepted which means that the
independent variables have no effect on the dependent variable together.
3.
T test
Partial test (t test) was conducted with the intention
of partially testing the effect of the independent variables on the dependent
variable with the assumption that other variables are considered constant with
a 95% confidence level (α = 0.05). The research hypothesis partially is as
follows:
a)
H0 : b1 = 0; Profit
management (EM) (X1) does not have a
significant effect on Effective Tax Rate (ETR) (Y);
H1 : b1≠ 0; Profit management(EM) (X1) significant effect on Effective Tax Rate (ETR)
(Y).
b)
H0 : b2 = 0; Profitability (Prof) (X2) does not have a significant effect on Effective Tax Rate (ETR)
(Y);
H1 : b2 ≠ 0; Profitability (Prof) (X2) significant effect on
Effective Tax Rate (ETR) (Y).
c)
H0 : b3 = 0; Leverage (LEV) (X3) does not have a significant effect on Effective Tax Rate (ETR)
(Y);
H1 : b3 ≠ 0; Leverage (LEV) (X3) significant effect on Effective Tax Rate (ETR)
(Y).
d)
H0 : b4 = 0; Transfer Pricing
(TP) does not have a significant effect on Effective Tax Rate (ETR)
(Y);
H1 : b4 ≠ 0; Transfer Pricing
(TP) significant effect on Effective Tax Rate (ETR)
(Y).
With the decision-making conditions:
Prob
value (t-statistic) < 0.05 (significance level 5%), then H0 is rejected, which
means that the independent variable has a significant effect on the dependent
variable partially.
Prob
value (t-statistic) > 0.05 (significance level 5%), then H0 is accepted,
which means that the independent variable has no effect on the dependent variable
partially.
4.
Determination (R2)
The coefficient of determination is used to measure how
far the model's ability to explain variations in the dependent variable is. The
value of the coefficient of determination is between zero and one. A small
value of R2 means that the ability of the independent variables in explaining
the variation of the dependent variable is very limited. R2 is used to
determine how much the independent variable is capable of explaining the
dependent variable Widarjono (2017) or in other words how much is
the ability of the Earnings Management (EM) variable (X1), Profitability (Prof)
(X2), Leverage (LEV) (X3) and Transfer Pricing (TP) in explaining the Effective
Tax Rate (ETR) (Y).
RESULTS AND DISCUSSION
A. Data
Description
In
this study, an analysis will be carried out regarding the effect of Earnings
Management (DA), Profitability (ROA), Leverage (DER) and Transfer Pricing (TP)
on Tax Avoidance (ETR) in companies included in the P3 sector (Plantation, Forestry
and Mining) which listed on the IDX in 2017 – 2021. Before analyzing the
factors that are thought to have an effect on tax avoidance, they will first be
analyzed descriptively.
Table 1
Descriptive statistics
|
Variables
|
Mean
|
Maximum
|
Minimum
|
Std.
Deviation
|
N
|
|
Tax
Avoidance (ETR)
|
23.47
|
47.86
|
1.75
|
7.47
|
110
|
|
Earnings
Management (DA)
|
-9.63
|
32.95
|
-112.26
|
31.55
|
110
|
|
Profitability
(ROA)
|
11.23
|
52.02
|
1.06
|
10.72
|
110
|
|
Leverage
(DER)
|
89.76
|
369.07
|
2.06
|
79.34
|
110
|
|
Transfer
Pricing (TP)
|
25.56
|
86.25
|
0.000
|
25.91
|
110
|
Source:
Processed data (2022)
Based
on the results in table 1 the amount of data used is 110 data consisting of 22
companies in 2017-2021 with an explanation of the results of descriptive
statistics on each variable as follows:
1. Tax Avoidance (ETR)
Tax avoidance or
ETR for companies included in the P3 sector (Plantation, Forestry and Mining)
listed on the IDX in 2017-2021 has an average of 23.47% with a standard
deviation of 7.47%. This shows that the average ratio of P3 sector corporate
income tax expense to income before tax reaches 23.47%. standard deviation of
7.47% which shows the level of variation in the distribution of data.
The lowest ETR is
a company with a SMAR issuer code of 1.75% in 2017 and the highest ETR is a
company with an ITMG issuer code of 47.86% in 2020. The average value of the
company's ETR during the 2017 - 2019 range is slightly below the effective tax
rate that applies according to the provisions, namely 25% and in 2020 - 2021
some companies have an ETR above the effective tax rate of 22%. This is because
in 2017 – 2019 some of these companies still have compensation for losses and
payment of other liabilities. The Covid-19 pandemic has caused most companies
to experience a decrease in tax payments but the payments are still above the
effective tax rate in effect.
2. Earnings Management (DA)
Earnings
management or DA in companies included in the P3 sector (Plantation, Forestry
and Mining) listed on the IDX in 2017-2021 has an average of -9.63% with a
standard deviation of 31.55%. The lowest earnings management is the company
with the issuer code MDKI of -112.261 in 2017 and the highest DA is the company
with the issuer code CEKA which is 32.95% in 2021. The most earnings management is
negative, namely – 9.63% which means that the average company does not carry
out earnings management.
3. Profitability (ROA)
Profitability or ROA of companies
included in the P3 sector (Plantation, Forestry and Mining) listed on the IDX
in 2017-2021 has an average of 11.23% with a standard deviation of 10.72%. This
shows that the average profitability of the P3 sector companies is mostly from
the sample of companies that perform poorly with a standard deviation of
10.72%. The lowest profitability is the company with the issuer code TKIM of
1.06% in 2027 and the highest profitability is the company with the issuer code
BYAN which is 52.02% in 2021 with the ROA calculation.
4. Leverage
(DER)
Leverage or
DER for companies that are included in the P3 sector (Plantation, Forestry and
Mining) listed on the IDX in 2017-2021 have an average of 89.76% with a
standard deviation of 79.34%. This shows that the average leverage of the P3
sector companies is quite high and the ratio of total debt to equity used by
the companies is quite high. This indicates that the average companies in the
P3 sector use debt instruments to finance their investment and operational
activities. The lowest DER is a company with a TOBA issuer code of 2.06% in
2019 and the highest DER is a company with an ITMG issuer code, which is
369.01% in 2020 with the DER calculation.
5. Transfer
Pricing (TP)
Transfer
Pricing or TP for companies that are included in the
P3 sector (Plantation, Forestry and Mining) listed on the IDX in 2017-2021 have
an average of 26.56% with a standard deviation of 25.91%. This shows that the
use of Receivable Relations instruments in P3 sector companies is relatively
small. The lowest transfer pricing is 0.00007%, namely companies with LSIP
issuer codes in 2019 and the highest is LSIP at 86.25% in 2021.
B. Panel Data Estimation Model
Panel
data regression is a regression model that uses panel data or data pools from a
combination of times series data and cross section data. There are several
models that can be used to estimate the panel data, namely the common effect
model approach, fixed effect model and random effect model (Widarjono, 2017)
1.
Common
Effect Model
The
Common Effect Model approach assumes that the behavior of data between
companies is the same at various times (Widarjono, 2017). Combined data is
considered to be part of the unity of observations so that in estimating the
parameters of this model, we can use OLS (Ordinary Least Square).
Table 2
Common Effect Model
|
Dependent Variable: ETR
|
|
Method: Least Squares Panel
|
|
Date: 07/02/22 Time: 22:03
|
|
Sample: 2017 2021
|
|
Periods included: 5
|
|
Cross-sections included: 22
|
|
Total panel (balanced) observations: 110
|
|
|
|
|
|
|
|
|
|
Variable
|
Coefficient
|
Std. Error
|
t-Statistic
|
Prob.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
C
|
24,42338
|
1.538215
|
15.87774
|
0.0000
|
|
|
DA
|
0.018040
|
0.021704
|
0.831206
|
0.4077
|
|
|
ROA
|
-0.006276
|
0.065828
|
-0.095333
|
0.9242
|
|
|
DER
|
0.020171
|
0.008962
|
2.250641
|
0.0265
|
|
|
TP
|
-0.094978
|
0.027714
|
-3.427057
|
0.0009
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Source:
Data processed 2022
2.
Fixed
effect model
This
approach assumes that there are intercept differences within the firm but the
same over time. The regression coefficient (slope) remains between companies
and time (Widarjono, 2017).
Table 3
Fixed Effect Model
|
Dependent Variable: ETR
|
|
Method: Least Squares Panel
|
|
Date: 07/02/22 Time: 22:03
|
|
Sample: 2017 2021
|
|
Periods included: 5
|
|
Cross-sections included: 22
|
|
Total panel (balanced) observations: 110
|
|
|
|
|
|
|
|
|
|
Variable
|
Coefficient
|
Std. Error
|
t-Statistic
|
Prob.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
C
|
22.57026
|
1.800131
|
12,53812
|
0.0000
|
|
|
DA
|
0.020948
|
0.017972
|
1.165596
|
0.2471
|
|
|
ROA
|
-0.186099
|
0.090467
|
-2.057082
|
0.0428
|
|
|
DER
|
0.031894
|
0.012208
|
2.612660
|
0.0106
|
|
|
TP
|
0.012249
|
0.041652
|
0.294088
|
0.7694
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Source:
Data processed 2022
3.
Random
Effect Model
This
approach is used to estimate the possibility that the disturbance variables
will be interconnected between time and individuals (Widarjono, 2017). The use
of this model must meet the requirements, namely the number of cross sections
must be greater than the number of research variables.
Table
4
Random Effect Model
|
Dependent Variable: ETR
|
|
Method: Panel EGLS (Cross-section random effects)
|
|
Date: 07/02/22 Time: 11:27
|
|
Sample: 2017 2021
|
|
Periods included: 5
|
|
Cross-sections included: 22
|
|
Total panel (balanced) observations: 110
|
|
|
|
|
|
|
|
|
|
Variable
|
Coefficient
|
Std. Error
|
t-Statistic
|
Prob.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
C
|
23.69971
|
1.847346
|
12.82906
|
0.0000
|
|
|
DA
|
0.018838
|
0.017627
|
1.068724
|
0.2876
|
|
|
ROA
|
-0.074135
|
0.072953
|
-1.016199
|
0.3119
|
|
|
DER
|
0.024640
|
0.010077
|
2.445057
|
0.0161
|
|
|
TP
|
-0.053857
|
0.032017
|
-1.682115
|
0.0955
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
C. Panel
Data Regression Model Structure Test
In
the panel data regression model to determine the effect of Earnings Management
(DA), Profitability (ROA), Leverage (DER) and Transfer Pricing (TP) on Tax
Avoidance (ETR) in companies included in the P3 sector (Plantation, Forestry
and Mining) which listed on the Indonesia Stock Exchange in 2017-2021, the best
model to be used will first be selected from 3 models, namely the common effect
model, fixed effect model and random effect model. The tests used to select the
best model are the Chow test, the Lagrange multiplier test and the Hausman
test.
1.
Chow test
Chow test was
conducted to determine the right model between the common effect or fixed
effect. Here are the hypotheses on the chow test:
H0: common
effectis the best model
H1:Fixed
Effect is the best model
Test statistical analysis results chow using
a significant level of 5%.
Table 5
Test Chow
|
Redundant Fixed Effects Tests
|
|
|
Equation: Untitled
|
|
|
|
Test cross-section fixed effects
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Effects Test
|
Statistics
|
df
|
Prob.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Cross-section F
|
4.324746
|
(21.84)
|
0.0000
|
|
Cross-section Chi-square
|
80.623195
|
21
|
0.0000
|
|
|
|
|
|
|
|
|
|
|
|
|
In
Table 5 the p-value is 0.000 <0.05 so H0 is rejected and H1 is accepted so
that it can be seen that the Fixed effect model is the best model to be used as
a panel regression model.
2.
Lagrange Multiplier Test
The Lagrange Multiplier
test is carried out to determine the right model between common effects or
random effects. The following is the hypothesis on the lagrange multiplier
test:
H0: Common effectis the best model
H1: Random Effectis the best model
Test statistical analysis result slagrange
multiplier using a significant level
of 5% can be seen in table 6.
Table
6
Lagrange Multiplier
|
Lagrange Multiplier Tests for Random Effects
|
|
Null hypotheses: No effects
|
|
Alternative hypotheses: Two-sided (Breusch-Pagan) and one-sided
|
|
(all others) alternatives
|
|
|
|
|
|
|
|
|
|
|
|
|
Hypothesis
Test
|
|
|
Cross-section
|
Time
|
Both
|
|
|
|
|
|
|
|
|
|
|
|
Breusch-Pagan
|
23,59601
|
0.065808
|
23.66182
|
|
|
(0.0000)
|
(0.7975)
|
(0.0000)
|
|
Honda
|
4.857572
|
0.256531
|
3.616217
|
|
|
(0.0000)
|
(0.3988)
|
(0.0001)
|
|
King-Wu
|
4.857572
|
0.256531
|
2.178143
|
|
|
(0.0000)
|
(0.3988)
|
(0.0147)
|
|
Standardized Honda
|
5.504626
|
0.809906
|
0.482538
|
|
|
(0.0000)
|
(0.2090)
|
(0.3147)
|
|
Standardized King-Wu
|
5.504626
|
0.809906
|
-0.304448
|
|
|
(0.0000)
|
(0.2090)
|
(0.6196)
|
|
Gourieroux, et al.
|
--
|
--
|
23.66182
|
|
|
|
|
(0.0000)
|
|
|
|
|
|
In
Table 6 the p-value obtained is 0.000 <0.05 so that H0 is rejected and H1 is
accepted, it can be seen that the Random effect model is the best model to be
used as a panel regression model.
3. Hausman
test
Hausman
test is conducted to determine the right model between random effects or fixed
effects. Here are the hypotheses on the Hausman test:
H0: Random effects is the best model
H1: Fixed Effect is the best model
Test statistical analysis results hausman
using a significant level of 5% can be
seen in table 7.
Table 7
Hausman test
|
Correlated Random Effects - Hausman Test
|
|
Equation: Untitled
|
|
|
|
Test cross-section random effects
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Test Summary
|
Chi-Sq. Statistics
|
Chi-Sq. df
|
Prob.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Cross-section random
|
7.497989
|
4
|
0.1118
|
|
|
|
|
|
|
|
|
|
|
|
|
In
Table 7 the p-value obtained is 0.1118 > 0.05 so that H0 is accepted and H1
is rejected, it can be seen that the random effects model is the best model to
be used as a panel regression model. Table 4.15 is a summary of the final model
used, namely the Random effect model because the majority of the 3 tests chose
the Random effect model as the best model.
Table 8
Panel Data Estimation
|
Determinants of
the Estimated Model (ETR)
|
|
Chow
test
Fixed
effect
|
Lagrange
Test
Random
effects
|
Hausman
test
Random
effects
|
|
Estimation model
used:
Random
effects
|
Source:
Processed data (2022)
D. Classic
assumption test
Classical
assumption test is used to determine whether the regression model shows a
significant and representative relationship. The classical assumption tests
performed were normality, multicollinearity and heteroscedasticity tests.
1. Normality test
The
normality test was carried out to determine the distribution of a residual from
the regression model of the influence of Earnings Management (DA),
Profitability (ROA), Leverage (DER) and Transfer Pricing (TP) on Tax Avoidance
(ETR) in companies included in the P3 sector (Plantation, Forestry). and
Mining) which are listed on the IDX in 2017 – 2021 have a normal distribution
or not. The normality test will use the probability value of Jarque Bera. A
data is normally distributed if it has a Sig value > 0.05. Table 4.16 shows
that the research data is normally distributed because it has a value of Sig
(0.091) > 0.05.
Table 9
Normality test
|
Jarque-Bera
|
Sig
|
Information
|
|
4,778
|
0.0917
|
Normal
Distribution
|
Source:
Processed data (2022)
Figure 1. Normality Test
Histogram Image
2. Multicollinearity Test
The
multicollinearity test aims to find out whether there is a correlation (strong
relationship) between the independent variables (Independent). A good
regression model should not have a correlation between the independent
variables or there should be no multicollinearity. How to check the case of
multicollinearity is by looking at the VIF value > 10. In table 15 it can be
seen that the data does not have a VIF value of more than 10 so that there are
no cases of high correlation or multicollinearity between independent variables
in the ETR regression model.
Table 10
Multicollinearity Test
|
Variance Inflation Factors
|
|
Date: 07/02/22 Time: 11:37
|
|
Sample: 2017 2021
|
|
|
Included observations: 110
|
|
|
|
|
|
|
|
|
|
|
|
|
Coefficient
|
Uncentered
|
Centered
|
|
Variable
|
Variance
|
VIF
|
VIF
|
|
|
|
|
|
|
|
|
|
|
|
C
|
3.412686
|
2.662131
|
NA
|
|
DA
|
0.000311
|
1.024449
|
1.001969
|
|
ROA
|
0.005322
|
1.535757
|
1.012144
|
|
DER
|
0.000102
|
1.672406
|
1.034112
|
|
TP
|
0.001025
|
1.586498
|
1.022251
|
|
|
|
|
|
|
|
|
|
|
Source:
Processed data (2022)
3. Heteroscedasticity Test
Heteroscedasticity test was conducted to
determine the existence of deviations from the requirements of classical
assumptions in linear regression. To find out whether there is
heteroscedasticity, the Glejser test can be carried out, namely by regressing
the absolute value of the residual on the independent variable. If the
significance probability is above the 0.05 confidence level, it does not
contain heteroscedasticity.
Table 11
Heteroscedasticity Test
|
Dependent Variable: ABSRESSION
|
|
|
Method: Panel EGLS (Cross-section random effects)
|
|
Date: 07/04/22 Time: 11:39
|
|
|
Sample: 2017 2021
|
|
|
|
Periods included: 5
|
|
|
|
Cross-sections included: 22
|
|
|
Total panel (unbalanced) observations: 109
|
|
Swamy and Arora estimator of component variances
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Variable
|
Coefficient
|
Std. Error
|
t-Statistic
|
Prob.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
C
|
5.079975
|
1.190987
|
4.265347
|
0.0000
|
|
DA
|
0.006517
|
0.014049
|
0.463902
|
0.6437
|
|
ROA
|
-0.061366
|
0.050060
|
-1.225848
|
0.2230
|
|
DER
|
0.004442
|
0.006937
|
0.640259
|
0.5234
|
|
TP
|
0.014841
|
0.021722
|
0.683252
|
0.4960
|
|
|
|
|
|
|
Source:
Processed data (2022)
In table 11 the results of the
probability value (F-Statistic) on the variables DA, ROA, DER and TP > 0.05
so it can be concluded that the panel data regression model on the ETR variable
of the P3 sector companies in 2017-2021 does not have heteroscedasticity
symptoms.
4. Autocorrelation Test
The autocorrelation test aims to test whether
in a linear regression model there is a correlation between the confounding
error in period t and the error in period t-1. A good regression model is one
that is free of autocorrelation. To detect autocorrelation, it can be done
through the Durbin-Watson test. DW test is done by making a hypothesis:
Ho : no autocorrelation ( r = 0 )
Ha : there is autocorrelation ( r 0 ).
The
basis for decision making is as follows:
a) If DW < DL, then there is a positive
autocorrelation
b) If
DL < DW < DU, then there is no doubt that autocorrelation occurs
c) If
DU < DW < 4-DU, then there is no autocorrelation
d) If
4- DU < DW < 4-DL, then there is no doubt that autocorrelation occurs
e) If
4-DL<DW, then there is a negative autocorrelation.
Information
:
DL
= lower limit DW
DU
= upper limit DW
DW = Durbin Watson
The following is the result of the autocorrelation
calculation using the Durbin Watson value:
Table 12
Durbin Watson Statistical Test
|
Durbin-Watson stat
|
Information
|
|
1,829
|
No Autocorrelation
|
Based on the results of the Durbin Watson test with
the amount of data (n) = 110 at a significant level = 5% and k = 5 independent
variables, it is obtained that the value of DL = 1.614 and DU = 1.765 According
to the provisions of the Durbin Watson test if the Durbin-Watson value is
between the values of dU and 4 -dU (1.765 < 1.829 < 2.235), it can be
concluded that there is no autocorrelation in the regression model.
E. Panel Data Regression
Equation
Panel data regression is a combination of cross
section data and time series data, where the same cross section unit is measured
at different times. So in other words, panel data is data from the same
individuals who are observed over a certain period of time. The following are
the results of the estimation of the influence of Earnings Management (EM)
(X1), Profitability (Prof) (X2), Leverage (LEV) (X3) and Transfer Pricing (TP)
(X4) on the Effective Tax Rate (ETR) (Y) using Random Effect Models:
Table 13
Coefficient of Random Effect Model
|
Variable
|
Coefficient
|
Std. Error
|
t-Statistic
|
Prob.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
C
|
23.69971
|
1.847346
|
12.82906
|
0.0000
|
|
DA
|
0.018838
|
0.017627
|
1.068724
|
0.2876
|
|
ROA
|
-0.074135
|
0.072953
|
-1.016199
|
0.3119
|
|
DER
|
0.024640
|
0.010077
|
2.445057
|
0.0161
|
|
TP
|
-0.053857
|
0.032017
|
-1.682115
|
0.0955
|
The panel data regression
equations compiled from the analysis are as follows:
ETRit = 23.69
+0.018Emit) – 0.074 (Profit) +0.024 (LEVit) – 0.053 (Tpit) + eit
Information :
ETRit = Tax Avoidance
EMit = Earnings
Management
Profit = Profitability
LEVit = Leverage
Tpi t = Transfer
Pricing
= Constant
eit = Error or Interruption
Variable
1, 2, 3, 4 = Regression coefficient
The above equation can be
interpreted as follows:
1) is 23.69 which means if Earnings Management (EM)
(X1), Profitability (Prof) (X2), Leverage (LEV) (X3) and Transfer Pricing (TP)
(X4) are zero, then Tax Avoidance (ETR) (Y ) will be worth 23.69 units.
2) The regression coefficient of the Earnings
Management (EM) (X1) variable is 0.018, which means that if there is an
increase in Earnings Management (EM) (X1) by 1 unit (assuming other variables
are constant), then Tax Avoidance (ETR) (Y) will increase. as big as0.018unit.
3) The regression coefficient of Profitability (Prof)
(X2) variable is -0.074 which means if there is a change in Profitability
(Prof) (X2) by 1 unit (assuming other variables are constant), then Tax
Avoidance (ETR) (Y) will decrease by -0.074 units.
4) The regression coefficient of the Leverage (LEV)
(X3) variable is 0.024, which means that if there is a change in the increase
in Leverage (LEV) (X3) by 1 unit (assuming other variables are constant), then
Tax Avoidance (ETR) (Y) will increase by 0.024 unit.
5) The regression coefficient for the Transfer Pricing
(TP) variable is -0.053, which means that if there is a change in the Transfer
Pricing (TP) (X4) by 1 unit (assuming other variables are constant), then Tax
Avoidance (ETR) (Y) will decrease by - 0.053 units.
F. Hypothesis test
Hypothesis
testing aims to measure the effect of Earnings Management (DA), Profitability
(ROA), Leverage (DER) and Transfer Pricing (TP) on Tax Avoidance (ETR) in
companies included in the P3 sector (Plantation, Forestry and Mining)
registered in IDX 2017-2021 partially or simultaneously. The hypothesis test
used consisted of F test and T test.
1. F test
The F
statistical test or the feasibility test or Goodness of fit was used to test
whether there was a simultaneous significant effect on the ETR regression
model. The significance level (ɑ) used in this study was 5%. This means that if
the p-value (Sig) <5%, the independent variable as a whole has an influence
on the dependent variable and is feasible to use.
Table 14
F test
|
Dependent Variable: ETR
|
|
|
Method: Panel EGLS (Cross-section random effects)
|
|
Date: 07/02/22 Time: 11:27
|
|
|
Sample: 2017 2021
|
|
|
|
Periods included: 5
|
|
|
|
Cross-sections included: 22
|
|
|
Total panel (balanced) observations: 110
|
|
Swamy and Arora estimator of component variances
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Variable
|
Coefficient
|
Std. Error
|
t-Statistic
|
Prob.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
C
|
23.69971
|
1.847346
|
12.82906
|
0.0000
|
|
DA
|
0.018838
|
0.017627
|
1.068724
|
0.2876
|
|
ROA
|
-0.074135
|
0.072953
|
-1.016199
|
0.3119
|
|
DER
|
0.024640
|
0.010077
|
2.445057
|
0.0161
|
|
TP
|
-0.053857
|
0.032017
|
-1.682115
|
0.0955
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Effects
Specification
|
|
|
|
|
|
|
SD
|
Rho
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Cross-section random
|
4.702043
|
0.4205
|
|
Idiosyncratic random
|
5.519699
|
0.5795
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Weighted
Statistics
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
MSE root
|
5.481885
|
R-squared
|
0.088831
|
|
Mean dependent var
|
10.90788
|
Adjusted
R-squared
|
0.054120
|
|
SD dependent var
|
5.769173
|
SE of
regression
|
5.610888
|
|
Sum squared resid
|
3305,617
|
F-statistics
|
2.559155
|
|
Durbin-Watson stat
|
1.829393
|
Prob(F-statistic)
|
0.042858
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
In Table 14 obtained
the value of F count (3.703) > F table (4.105) (2.458) or the value of Sig
(0.0428) <0.05, which means H0 is rejected and H1 is accepted. This means
that there is a significant effect of the variable earnings
management, profitability, leverage and transfer pricing simultaneously on tax
avoidance of P3 sector companies, so that Hypothesis 5 is accepted.
2. T test
T test or partial test is used to test
how the influence of each independent variable (Independent) namely earnings
management, profitability, leverage and transfer pricing individually on the
dependent variable (dependent) namely tax avoidance. The significance level
applied is 5%, so if the value of Sig < 0.05 then the independent variable
individually has a significant influence on the dependent variable. T test
results can be seen in the following table:
Table 16
T test
|
Dependent Variable: ETR
|
|
|
Method: Panel EGLS (Cross-section random effects)
|
|
Date: 07/02/22 Time: 11:27
|
|
|
Sample: 2017 2021
|
|
|
|
Periods included: 5
|
|
|
|
Cross-sections included: 22
|
|
|
Total panel (balanced) observations: 110
|
|
Swamy and Arora estimator of component variances
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Variable
|
Coefficient
|
Std. Error
|
t-Statistic
|
Prob.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
C
|
23.69971
|
1.847346
|
12.82906
|
0.0000
|
|
DA
|
0.018838
|
0.017627
|
1.068724
|
0.2876
|
|
ROA
|
-0.074135
|
0.072953
|
-1.016199
|
0.3119
|
|
DER
|
0.024640
|
0.010077
|
2.445057
|
0.0161
|
|
TP
|
-0.053857
|
0.032017
|
-1.682115
|
0.0955
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Effects
Specification
|
|
|
|
|
|
|
SD
|
Rho
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Cross-section random
|
4.702043
|
0.4205
|
|
Idiosyncratic random
|
5.519699
|
0.5795
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Weighted
Statistics
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
MSE root
|
5.481885
|
R-squared
|
0.088831
|
|
Mean dependent var
|
10.90788
|
Adjusted
R-squared
|
0.054120
|
|
SD dependent var
|
5.769173
|
SE of
regression
|
5.610888
|
|
Sum squared resid
|
3305,617
|
F-statistics
|
2.559155
|
|
Durbin-Watson stat
|
1.829393
|
Prob(F-statistic)
|
0.042858
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
a) Earnings Management (DA)
Earnings management
has a positive coefficient value of 0.018 with a sig value of 0.2876 > 0.05
or t count 1.068 < t table (105) 1.98. This shows that H0 is accepted and H1
is rejected, which means that the Earnings Management variable has no partial
effect on tax avoidance and is not in accordance with the proposed hypothesis,
meaning that hypothesis 1 is not accepted.
b) Profitability (ROA)
Profitability has a
negative coefficient value of -0.074 with a sig value of 0.3119 > 0.05 or t count
-1.016 < t table (105) -1.98. This shows that H0 is accepted and H1 is
rejected, which means that the profitability variable has no partial effect on
tax avoidance and is not in accordance with the proposed hypothesis, meaning
that hypothesis 2 is not accepted.
c) Leverage
(DER)
Leverage has a positive coefficient value of 0.020
with a sig value of 0.0265 <0.05 or t count 2.445 > t table (105) 1.98.
This shows that H0 is rejected and H1 is accepted, which means that the
leverage variable has a partial effect on tax avoidance and is in accordance
with the proposed hypothesis, meaning that hypothesis 3 is accepted.
d) Transfer
Pricing (TP)
Transfer
Pricing has a negative coefficient value of -0.053
with a sig value of 0.095 > 0.05 or t count -1.682 < t table (105) -1.98.
It means that the transfer pricing variable has no partial effect on tax
avoidance in a negative direction so that H0 is accepted and H1 is rejected, so
that Hypothesis 4 is not accepted.
3. Coefficient of Determination
(R2)
The
coefficient of determination is used to measure how much influence the
independent variable has on the dependent variable in the study. The greater
the R-Square value is close to one, the ability of the independent variable to
explain the dependent variable is very informative. Table 4.22 shows an
R-Square of 8.8% so it can be concluded that the variables of earnings
management, profitability, leverage and transfer pricing are able to explain
the dependent variable, namely ETR of 8.8% while the remaining 91.2% is
explained by other variables outside of this study.
G. Discussion
1. The Influence of Profit management to Tax evasion
Earnings Management Variable
(X1) has a coefficient value of 0.018 with a significance
probability of 0.028. The significance value is greater than the significance
level (α) of 0.05 so that X1 is rejected. Thus it can be said that Earnings
Management has no significant effect on the direction of the Positive
relationship to Tax Avoidance. Therefore, H1: Earnings Management has a positive
effect on Tax Avoidance is rejected. The results show that earnings management
has no effect on tax avoidance by companies in the P3 sector. This means that
P3 sector companies are considered to prefer to use earnings management for
accounting purposes by increasing commercial profits rather than for tax
purposes. In addition, the sample companies are companies listed on the Stock
Exchange so that the Company will try to display financial statements with the
best performance because all parties need information from the financial
statements so that the presentation of information in the financial statements
is required to comply with applicable regulations including in accordance with
tax provisions. The results of this study are supported by research conducted
by Lee and Swenson (2011) who conducted research on
earnings management on discretionary spending in America, The results of other
studies which show that Earnings Management has no effect on tax avoidance are
research conducted by Henny (2019) and Syanthi et al. (2013)
because the company is proven not to do earnings management in tax planning
because companies that carry out earnings management will have higher profits.
persistent compared to companies that do not perform earnings management. In
addition, Sari and Ajengtiyas (2021) who took a sample of mining
companies on the IDX produced the same results where earnings management had no
effect on tax avoidance because earnings management did not affect management's
decision to carry out tax planning which would minimize the tax burden.
However, the results of this study contradict those of (Amidu et al., 2019; Pajriyansyah & Firmansyah, 2017; Tiaras &
Wijaya, 2015). The
difference in this research is because the research concludes that earnings
management carried out by companies in the form of income decreasing is aimed
at avoiding government regulations where these regulations are directly related
to company profits so that earnings management actions carried out by companies
are more towards tax management so that the tax burden paid by the company is
reduced. In addition, the difference in reporting according to accounting and
tax is a gap to carry out earnings management, but because of the pressure to
show good company performance there will be a conflict of interests of the
company to carry out management profit for tax purposes (Pajriyansyah & Firmansyah, 2017). The difference in
this research is because the research concludes that earnings management
carried out by companies in the form of income decreasing is aimed at avoiding
government regulations where these regulations are directly related to company
profits so that earnings management actions carried out by companies are more
towards tax management so that the tax burden paid by the company is reduced.
2. The Effect of Profitability on Tax Avoidance
Profitability variable (X2) has a coefficient value of
-0.074 with a significance probability of 0.3119. The significance value is
greater than the significance level (α) of 0.05 so that X2 is rejected. Thus,
it can be said that profitability has no significant effect on the direction of
the negative relationship to tax avoidance. Therefore, H1: Profitability has a
positive effect on tax avoidance. As the company's ability to generate profits
increases, the company's operating profit will also increase and taxes payable
will also increase. The results of this study reveal the behavior of ROA which
has a negative relationship with Tax Avoidance. The direction of the negative
relationship means that Profitability has a behavior that is not in the
direction of Tax Avoidance (ETR) where when ROA increases, Tax Avoidance will
decrease and vice versa if ROA decreases, Tax Avoidance will increase. The
results of this study are supported by research conducted by Mbroh, Monney, and Bonsu (2019)
which examines the relationship between tax avoidance and corporate
profitability in the country of Ghana resulting in a negative relationship
between tax avoidance and profitability (ROA) because good corporate governance
is required to generate profitability compared to do tax evasion.
However, the results of this study contradict those of
Jaffar et al. (2021)
and Kim and Im (2017). The difference in
the interpretation of the research results is because according to Jaffar et el
(2021) who examined companies in Malaysia that generate higher profits pay
lower tax rates because these companies do more tax planning to reduce the tax
burden. The company focuses on developing a tax strategy to reduce its income
tax liability but not on profits in its financial statements.
3. The Effect of Leverage on Tax Avoidance
Profitability variable (X3) has a coefficient value of
0.024 with a significance probability of 0.095. The significance value is
smaller than the significance level (α) of 0.05 so that X3 is accepted. Thus,
it can be said that leverage is significant with a positive relationship
towards tax avoidance. Therefore, H1: l Leverage has a positive effect on tax
avoidance. As the company's debt increases, the interest expense that must be
paid increases and this increase in interest expense naturally causes
additional costs which ultimately reduce taxable income. The results of this
study reveal the behavior of DER which has a direct relationship with Tax
Avoidance. The direction of the unidirectional relationship means that an
increase in leverage will affect an increase in tax avoidance (ETR becomes
smaller) and vice versa, a decrease in DER will reduce tax avoidance (ETR will
be larger). Leverage reflects the complexity of a financial transaction where
high leverage increases the company's ability to avoid taxes Pajriansyah
(2017). In addition, there are indications that companies that have
a high tax burden will tend to finance debt to reduce the tax burden through
interest payments because the capital structure of equity financing does not
receive tax incentives in Indonesia. The results of this study are in line with
those carried out by (Kim & Im, 2017; Pajriyansyah & Firmansyah, 2017) but different from the research conducted by Tiaras and Wijaya (2015)
and Jaffar et al. (2021). This difference
is due to differences in interpretation which state that the company does not
use debt to avoid tax. Research conducted by Ugbogbo, Omoregie, and Eguavoen (2019)
which examines the determinants of tax avoidance in Nigeria makes the
hypothesis that leverage does not affect tax aggressiveness because companies
that have high leverage are not directly motivated to do tax avoidance and
other studies show that there is no significant effect of leverage on tax
avoidance against companies that take advantage of tax shelters. This
difference is due to differences in interpretation which state that the company
does not use debt to avoid tax. Research conducted by Ugbogbo, Omoregie, and Eguavoen (2019)
which examines the determinants of tax avoidance in Nigeria makes the
hypothesis that leverage does not affect tax aggressiveness because companies
that have high leverage are not directly motivated to do tax avoidance and
other studies show that there is no significant effect of leverage on tax
avoidance against companies that take advantage of tax shelters. This difference
is due to differences in interpretation which state that the company does not
use debt to avoid tax. Research conducted by Ugbogbo, Omoregie, and Eguavoen (2019)
which examines the determinants of tax avoidance in Nigeria makes the
hypothesis that leverage does not affect tax aggressiveness because companies
that have high leverage are not directly motivated to do tax avoidance and
other studies show that there is no significant effect of leverage on tax
avoidance against companies that take advantage of tax shelters.
4. The Effect of Transfer Pricing on Tax Avoidance
The Transfrer Pricing (X4) variable has a coefficient
value of -0.053 with a significance probability of 0.095. The significance
value is greater than the significance level (α) of 0.05, meaning that the
Transfer Pricing variable does not significantly affect tax avoidance in a
negative direction. Therefore, H1: l Transfer has a positive effect on Tax
Avoidance is rejected. This negative direction indicates that an increase in
Transfer Pricing activities will reduce tax avoidance activities (ETR will
increase). Taxpayers engaged in the P3 sector have a tendency to carry out
Transfer Pricing using relational receivables instruments but with other
accounting purposes that will increase asset capitalization as a group so that
overall company performance looks good. This is inseparable because the samples
taken are companies that have been listed on the stock exchange because their
financial statements are exposed to all parties so that the company will
display the best performance of their company. This research is supported by
research conducted by Widyanto, Kristanto, and Sucahyo (2019)
which results in the conclusion that transfer pricing has no effect on tax
avoidance due to tax policies issued by the government that provide tax
incentives to avoid transfer pricing, namely the tax amnesty program that makes
several companies deposit their funds abroad. repatriate assets to Indonesia.
There has been an increase in supervision over Transfer Pricing transactions
conducted by the Directorate General of Taxes in recent years with the issuance
of Regulation No. PMK. 213/PMK. 03/2016 which requires companies to disclose
affiliate transactions through TP DOC and the regulatory procedures that adopt
the OECD Guidelines play a role in reducing transfer pricing practices in
Indonesia. However, it is different from the results of research conducted by Amidu et al. (2019) and Sari and Ajengtiyas (2021). the company operates in many countries
and has a tendency to exploit loopholes in tax laws that differ between
countries.
5. Effect of Earnings Management, Profitability, Leverage
and Transfer Pricing on tax avoidance
Earnings Management, Profitability, Leverage and
Transfer Pricing variables simultaneously affect tax avoidance with a
probability value of 0.04and an R-Square of 8.8%, which means that the
determinant of tax avoidance from this research model is only able to explain
8.8% and the remaining 91.2%. influenced by other variables. The low value of
R-Square is due to the large number of selected independent variables that have
less significant effect on tax avoidance.
Earnings Management Variable; based on the
significance test Earnings management has no effect on tax avoidance and from descriptive
statistical data the average company that performs earnings management is
-9.63%. The negative coefficient indicates that the average company does not
carry out earnings management.
Profitability Variable; based on the significance test
Profitability has no effect on tax avoidance and from descriptive statistical
data the average company that performs well is 11.93%. With the company's
performance tht is not too good but the company has an average tax payment
above the effective tax rate in effect in 2020-2021 where the average effective
tax rate for P3 sector companies is 23.5% while the applicable tax rate is 22%.
Variable Leverage; based on the significance test, it
is known that leverage has a significant effect on tax avoidance because the average
descriptive statistics of companies engaged in this sector have a large debt of
79.34%. The large leverage of companies engaged in this sector cannot be
separated from the high financing for capital expenditure needs, especially for
land acquisition and the company's initial investment costs. From the data
recap, almost all companies engaged in this sector have large affiliated and
non-affiliated debts, causing interest payments which naturally reduce the
amount of taxable income that must be paid by the company.
Variable Transfer Pricing; based on the significance
test of transfer pricing, it has no effect on tax avoidance and from
descriptive statistical data the average transfer pricing of the company is
26.56%. This shows that the transaction of relational receivables has no effect
on the company's taxable profit because some of the relational receivables are
ultimately not aimed at tax avoidance but rather for other accounting purposes.
The sensitivity of the transfer pricing measure can be measured with another
approach as has been done by Richardson et al. (2013)
who studied the determinants of the aggressiveness of transfer pricing in
Australia by making a more complex equation including the control variable of
the industrial sector and the independent variable of profitability, intengible
assets, MNC, Leverage, tax heaven utility. The same thing as done by (Amidu et al., 2019)
who examined the relationship between transfer pricing and tax avoidence in the
country of Ghana which used a similar approach to Richardson et al. (2013)
but by adding control variables for firm age and liquidity. In addition to the
similarity of the variables used in the measurement of the two studies, the
sample is companies that operate in a multinational manner.
CONCLUSION
Earnings Management
has no significant effect on the direction of the Positive relationship to Tax
Avoidance. This means that P3 sector companies are considered to prefer to use
earnings management for accounting purposes by increasing commercial profits
rather than for tax purposes.
Profitability
has no significant effect on tax avoidance because the higher the profitability
of companies in the P3 sector increases the amount of tax payments to taxpayers
in this sector. Companies prefer to do tax planning for their tax obligations
to the maximum when they have good profitability to avoid the costs incurred
when compliance is carried out by the DGT.
Leverage effect on tax avoidance because
companies use more debt instruments in financing investment and working capital
because debt will generate interest expense which naturally will reduce taxable
profit.
Transfer Pricing does
not have a significant effect on tax avoidance. This shows that taxpayers
engaged in the P3 sector carry out transfer pricing through relational receivables
instruments not with the aim of tax avoidance but more for other accounting
purposes such as transfers between business units of one group which will
increase asset capitalization so that the company's performance will look good.
Earnings
Management, Profitability, Leverage and Transfer Pricing variables
simultaneously affect the tax avoidance of companies engaged in the P3 sector
listed on the IDX.
REFERENCES
Amidu, M., Coffie, W., & Acquah, P. (2019). Transfer
pricing, earnings management and tax avoidance of firms in Ghana. Journal of
Financial Crime. Google Scholar
Chen, S., Chen, X., Cheng, Q., & Shevlin, T. (2010). Are
family firms more tax aggressive than non-family firms? Journal of Financial
Economics, 95(1), 41–61. Elsevier
Hanlon, M., & Heitzman, S. (2010). A review of tax
research. Journal of Accounting and Economics, 50(2–3), 127–178. Elsevier
Hansen, D. R., & Mowen, M. M. (2007). Managerial
accounting. South-Western. Google Scholar
Henny, H. (2019). Pengaruh Manajemen Laba Dan Karakteristik
Perusahaan Terhadap Tax Avoidance. Jurnal Muara Ilmu Ekonomi Dan Bisnis,
3(1), 36–46. Google Scholar
Jacob, M., & Schütt, H. H. (2020). Firm valuation and the
uncertainty of future tax avoidance. European Accounting Review, 29(3),
409–435. Google Scholar
Jaffar, R., Derashid, C., & Taha, R. (2021). Determinants
of tax aggressiveness: Empirical evidence from Malaysia. The Journal of
Asian Finance, Economics and Business, 8(5), 179–188. Google Scholar
Kim, J. H., & Im, C. C. (2017). The study on the effect
and determinants of small-and medium-sized entities conducting tax avoidance. Journal
of Applied Business Research (JABR), 33(2), 375–390. Google Scholar
Lee, N., & Swenson, C. (2011). Earnings management
through discretionary expenditures in the US, Canada, and Asia. International
Business Research, 4(2), 257. Google Scholar
Mbroh, N., Monney, A., & Bonsu, M. O.-A. (2019). Corporate
tax avoidance and firm profitability. Google Scholar
Pajriyansyah, R., & Firmansyah, A. (2017). Pengaruh
leverage, kompensasi rugi fiskal dan manajemen laba terhadap penghindaran
pajak. Keberlanjutan, 2(1), 431–459. Google Scholar
Permatasari, M., & Trisnawati, E. (2022). Deferred Tax On
Real Profit Management With Tax Planning As Moderating. Jurnal Akuntansi,
26(2), 280–305. Google Scholar
Richardson, G., Taylor, G., & Lanis, R. (2013).
Determinants of transfer pricing aggressiveness: Empirical evidence from
Australian firms. Journal of Contemporary Accounting & Economics, 9(2),
136–150. Elsevier
Sari, R. H. D. P., & Ajengtiyas, A. (2021). Pengaruh
Profitabilitas, Transfer Pricing, dan Manajemen Laba Terhadap Tax Avoidance. Konferensi
Riset Nasional Ekonomi Manajemen Dan Akuntansi, 2(1), 898–917. Google Scholar
Scott, W. R. (2015). Financial accounting theory sevent
edition. United States: Canada Cataloguing. Google Scholar
Sitorus, R. (2020). Pengaruh Transfer Pricing Dan Sales
Growth Terhadap Tax Avoidance Dengan Profitabilitas Sebagai Variabel
Moderating. Media Akuntansi Perpajakan, 5(1), 18–31. Google Scholar
Sugiyono. (2019). Metode Penelitian Kuantitatif,
Kualitatif R&D. Bandung: Alfabeta. Google Scholar
Suyanto, K. D., & Supramono, S. (2012). Likuiditas,
leverage, komisaris independen, dan manajemen laba terhadap agresivitas pajak
perusahaan. Jurnal Keuangan Dan Perbankan, 16(2). Google Scholar
Syanthi, N. T., Sudarma, M., & Saraswati, E. (2013).
Dampak Manajemen Laba terhadap Perencanaan Pajak dan Persistensi Laba. EKUITAS
(Jurnal Ekonomi Dan Keuangan), 17(2), 192–210. Google Scholar
Tiaras, I., & Wijaya, H. (2015). Pengaruh likuiditas,
leverage, manajemen laba, komisaris independen dan ukuran perusahaan terhadap
agresivitas pajak. Jurnal Akuntansi, 19(3), 380–397. Google Scholar
Ugbogbo, S. N., Omoregie, N. A., & Eguavoen, I. (2019).
Corporate determinants of aggressive tax avoidance: Evidence from Nigeria. IOSR
Journal off Business and Management (IOSR-JBM). Google Scholar
Widyanto, C. R., Kristanto, A. B., & Sucahyo, U. S.
(2019). The influence of tax aggressiveness as an intervening variable on the
relationship between managerial ownership and firm value. International
Journal of Recent Technology and Engineering, 7(6), 965–971. Google Scholar
