Professional Documents
Culture Documents
ABSTRACT
This paper investigated the propensity and the magnitude of tax avoidance risk exposure
among different taxpayers by analyzing both enterprise-related and government-related
variables. Providing far-reaching analysis and examining a relatively unexplored area of
conforming tax avoidance, this study employs two measurements of tax avoidance
including non-conforming and conforming tax avoidance. In the brain area of empirical
analysis, this paper combined a fixed-effect model to control omitted variable bias
together with adoption of adopts heteroskedasticity and autocorrelation-consistent stan-
dard errors (HAC/clustered SE). The results depict that the magnitude of tax avoidance
risks varies depending on the characteristics of taxpayers. Higher risks were found in
so-called foreign-controlled enterprises and foreign invested enterprises. With respect to
entities’ sector, this study also demonstrates that the propensity of higher risk exposure
was depicted in financial and mining sector relative to full sample taxpayers.
Keywords: tax avoidance, risk assessment, evasion
ABSTRAK
Paper ini meneliti kecenderungan dan besarnya risiko penghindaran pajak atas bebera-
pa tipe Wajib Pajak dengan menggunakan data terkait perusahaan dan terkait pemer-
intah. Analisis yang lebih menyeluruh menggunakan pendekatan penghindaran pajak
conforming dan non-conforming. Paper ini menggunakan fixed effect model untuk
mengkontrol variable yang tidak dimasukkan dalam model dengan mengadopsi
heteroskedasticity dan autocorrelation-consistent SE. hasil penelitian ini menunjukkan
bahwa besarnya risiko penghindaran pajak bervariasi tergantung tipe Wajib Pajak.
Risiko terbesar ditemukan pada perusahaan investasi luar negeri dan perusahaan yang
dikontrol oleh entitas di luar negeri. Terkait analisis risiko penghindaran pajak sektoral,
studi ini menkonfirmasi kecenderungan risiko yang lebih besar muncul pada sector
keuangan dan pertambangan.
Kata kunci: penghindaran pajak, penilaian risiko, penggelapan
1
Agung Endika Satyadini, Restu Rea Erlangga, Brigitta Steffi VDKG / Who Avoids Taxes? (2019) 1-26
2
Agung Endika Satyadini, Restu Rea Erlangga, Brigitta Steffi VDKG / Who Avoids Taxes? (2019) 1-26
argued that this definition overrides the Effective tax rate (ETR) was also consid-
distinction between common activities ered as one of the most effective ways to
that are tax-favored and those that are measure tax avoidance. Dyreng et al.
tax planning, aiming specifically to (2008) suggested that the book ETR,
reduce taxes and targeted tax benefits formulated as the total tax expense divid-
from aggressive lobbying activities. ed by pre-tax income, be broadly used to
Following Hanlon & Heitzman (2010), measure a firm’s tax burden. In terms of
since ambiguity of whether a transac- tax planning measurement, which is cap-
tion is permissible or not, this study tured in the financial performance, Mills,
does not make a distinction between Erickson, and Maydew (1998) suggested
legal tax avoidance and illegal tax eva- that the ETR might be a powerful indicator
sion1. From the context of relation of the effectiveness of a company’s tax
between avoidance risk and client planning activities. Robinson, Sikes, &
behavior, Satyadini et al (2019) empha- Weaver (2010) emphasized that the value
sized that affecting client behavior and of the ETR represents tax avoidance activi-
actively steering the population towards ties that directly affect net income.
low risk will allow customs authorities to According to Gupta and Newberry (1997),
concentrate more on controlling Rego (2003), Zimmerman (1983), Omer,
resources on high risks. Molloy, & Ziebart (1993), Armstrong,
Blouin, & Larcker (2011), and Jacob (1996),
2.2 Tax Avoidance Measurement lower values of the ETR represent higher
levels of tax avoidance.
As suggested by Satyadini (2018), there However, Hanlon & Heitzman
is abundant literature on tax avoidance (2010) argued that these studies captured
measurement. One aspect to define only non-conforming tax avoidance, in
the intention of tax avoidance is aggres- which transaction for tax and accounting
siveness. However, aggressiveness is purposes would be reported differently;
difficult to measure because it is not while conforming tax avoidance, in which
uniform and depends on variation in tax avoidance practices would simultane-
dutifulness and honesty (Slemrod, ously reduce financial accounting income,
2007). Therefore, how to quantify tax was not captured in these studies. Frank et
aggressiveness is still puzzled. al. (2009) developed a model of ETR
Prior studies have suggested differential, occupying the gap between
several methods to measure tax statutory tax rate and ETR, and used
aggressiveness, for example, Frank et al. permanent difference measurement.
(2009) used the total book-tax differ- Afresh, Hanlon & Heitzman (2010) argued
ences (DTAX) and Wilson (2009) adopt- that this model also captured only
ed the discretionary permanent non-conforming tax avoidance because
book-tax differences (BTD). permanent difference was a function of
ETR2.
1 Several
related literatures are borrowed from Satyadini (2018).
2Permanent difference, denoted as PERMDIFF by Frank et al (2009), is essentially the difference between the effective and statutory
tax rates multiplied by pre-tax accounting income for the estimation.
3
Agung Endika Satyadini, Restu Rea Erlangga, Brigitta Steffi VDKG / Who Avoids Taxes? (2019) 1-26
Providing empirical evidence of con- in various forms, including but not limited
forming tax avoidance, Salihu, Obid, & to financial secrecy, combined with vary-
Annuar (2013) used the ratio of cash ing degrees of refusal to co-operate with
taxes paid to the operating cash flow as other jurisdictions in exchanging informa-
tax avoidance proxy, which captures the tion (Shaxson, 2011).
reduction in both financial accounting,
tax incomes and accruals simultaneous- 2.4 Size of Enterprise
ly. They replicated the findings of
Hanlon & Heitzman (2010) showing that Several studies revealed a unique correla-
this ratio provided measurement of tion between size of enterprise and tax
conforming tax avoidance. avoidance. Markle & Shackelford (2012)
provided evidence that the impact of
2.3 Low or No Tax Jurisdiction enterprises’ size on ETR’s has been uncon-
vincing. Other studies by Rego (2003),
The term 'tax haven' has been widely Omer et al. (1993), and Zimmerman (1983)
used since the 1950s, yet there is no revealed a negative correlation between
consensus as to what it means (Palan, size of an enterprise and ETRs. Conversely,
2009). According to OECD (2018), tax using size of an enterprise as a single
haven in the "classical" sense refers to a function of total sales, Noor, Fadzillah, &
country which imposes a low or no tax Mastuki (2010) found a positive correla-
and is used by corporations to avoid tax tion. Noor et al (2010) measured tax
which otherwise would be payable in a avoidance under the Official Assessment
high-tax country. According to OECD System (OAS) and the Self-Assessment
report in 1998 and 2001, tax havens System (SAS). The result suggested that
have the following key characteristics ETR was positively correlated with size
including (1) No or only nominal taxes; during both the OAS and SAS regimes.
(2) Lack of effective exchange of infor- IOn the other hand, studies by Gupta and
mation; and (3) Lack of transparency in Newberry (1997), Armstrong et al. (2011)
the operation of the legislative, legal or and Mills et al. (1998) concluded that there
administrative provisions. However, was no relation. Moreover, Slemrod (2007)
Dharmapala & Hines (2009) argued that suggested that according to the U.S. Gen-
tax havens are typically small, well-gov- eral Accounting Office, the IRS estimated
erned states that impose low or zero tax that big enterprises tended to have lower
rates on foreign investors. Moreover, non-compliance rates than smaller enter-
Dharmapala & Hines calculated that for prises3.
a country with a population under one
million, the likelihood of becoming a tax
haven rises from 24 percent to 63
percent. Another characteristic from tax
havens is that they often offer secrecy,
3Permanent difference, denoted as PERMDIFF by Frank et al (2009), is essentially the difference between the effective and statutory
tax rates multiplied by pre-tax accounting income for the estimation.
4
Agung Endika Satyadini, Restu Rea Erlangga, Brigitta Steffi VDKG / Who Avoids Taxes? (2019) 1-26
2.5 Avoidance Risk on Specific Sector Therefore, it is necessary to look deep into
mining industry regarding tax avoidance
In the light of tax avoidance discussion, measurement.
literature suggested the involvement of
several segments. The practice of tax 2.6 Audit Penalty
avoidance is aided by a very lucrative
tax avoidance industry, staffed by pro- Audit intensity, penalties, and risk aversion
fessional accountants, lawyers and are closely related to reduction of tax
finance experts (Sikka and Willmott, aggressiveness. Slemrod (2007) pointed
2010). The International Consortium of out that optimal tax evasion leans on the
Investigative Journalists (2016) on their probability of getting caught, amount of
report “Panama Papers” stated how penalty and level of risk aversion. He also
major banks have driven and helped emphasized Becker’s (1968) model of eco-
the creation of hard-to-trace compa- nomics of crime, how enterprises maxi-
nies in offshore havens. More than 500 mized their expected utilities by consider-
banks, their subsidiaries, and their ing possible penalties in the equal way as
branches created more than 15,000 other contingent costs. Similarly, in the
offshore companies for their customers context of tax compliance decision made
through Mossack Fonseca. This has by an enterprise, Alm & McKee (1998) pro-
come to a conclusion i.e. financial vided an argument that tax aggressive-
industry has played a big part in tax ness under uncertain circumstances was
avoidance. rigorously correlated with the fear of the
On the other hand, the 2014 possibility of being caught and penalized.
Global Financial Integrity (GFI) Report Broadly identical with Becker’s
entitled “Illicit Financial Flows from (1968) model, Allingham & Sandmo (1972)
Developing Countries: 2003-2012”, puts established a model under the assump-
Indonesia as the seventhplace country tion of Vonn Neumann – Morgenstern4
in the world with the highest illicit finan- axiom for behavior under uncertainty5 . As
cial flows. The report estimates that the an extension of this research, Yitzhaki
total illicit financial flows in Indonesia for (1974) specified that if the penalty rate was
2003-2012 reached $187,844 million proportional with the tax understated
(IDR 1,690 trillion, average exchange (rather than income understated), the tax
rate IDR 9,000/US$) or reached $18,784 rate would provide no effect on the spec-
million per year. Using the data, PWYP ulation to carry out avoidances since
Indonesia (2015) estimated illicit finan- reward-to-risk was unchanged. Here, if
cial flows in Indonesia for 2014 reached the marginal benefit of evasion as the
IDR 227.7 trillion or equal to 11.7%. function of income understated and tax
While the Mining sector in Indonesia rate is smaller than the marginal cost of
accounted for 10.5% of the total illicit detection (function of penalty rate, tax rate
financial flows in Indonesia, which was and audit intensity), the optimum level of
estimated to reach IDR 23.89 trillion in tax evasion will be zero6.
2014 (PWYP Indonesia, 2015).
5
Agung Endika Satyadini, Restu Rea Erlangga, Brigitta Steffi VDKG / Who Avoids Taxes? (2019) 1-26
Another literature about the impact of Tax gap may be defined from the institu-
penalties on tax avoidance demonstrat- tional perspective of tax authorities as “the
ed different results. Beck, Davis, & Jung difference between tax collected and the
(1991) and Park & Hyun (2003) provided tax that should be collected” (HMRC 2012,
evidence of positive correlation p. 3). That the amount of tax collected is in
between penalty and tax compliance. practice less than it should be if all taxpay-
Conversely, Alm, Jackson, & Mckee ers were fully compliant with their tax obli-
(1992) provided evidence of negative gations, i.e. if they filed complete and
correlation. Other studies suggested accurate tax returns and paid all due taxes
that the result is fluctuated according to is often referred to as the ‘tax gap’. Tax
taxpayer’s characteristics; for instance, evasion, depending on the country,
Witte & Woodbury (1985) acknowl- accounts for 80%-90% of the total tax
edged high tendency of negative gap, which also increases as a result of tax
correlation between penalty and tax avoidance, excessive corruption or low
avoidance for small and medium effectiveness of tax administration (Harre-
taxpayers, but positive correlation for mi 2014, p. 365). Fisman and Wei (2004)
large taxpayers. stated, one particularly important issue is
understanding the relationship between
2.7 Tax Rate Differentials tax rates and tax evasion. The relationship
between tax rates and evasion is positive,
Prior studies indicated that ETR was but this depends on assumptions of risk
closely related to strategy as a response aversion and the punishment for evasion
to tax differentials, especially for multi- (Allingham and Sandmo, 1972). In this
national enterprises. Rego (2003) pro- brain area, tax gap may be caused by
vided an evidence that the magnitude numerous elements including (1) unclear
of multinational operations was nega- legislation, negligent omissions, differenc-
tively correlated to book ETR, he sug- es in interpretation, lack of knowledge,
gested that the multinational enterpris- and non-deliberate errors leading to
es tend to avoid taxes. From the mana- differences between the tax intended to
gerial accounting point of view, multina- be collected and the amount actually
tional enterprises are able to allocate collected; (2) insolvencies, whose conse-
profits, losses, and expenses based on quence is the impossibility for tax authori-
geographical strategy. In this strategy, ties to collect the taxes on bankrupt com-
profit-center companies are usually panies, even though there is a tax liability;
located in low or no tax jurisdictions. and (3) taxpayers’ deliberate actions such
Conversely, cost-center companies are as tax fraud, tax evasion and tax avoid-
usually located in high tax countries. ance.
4Neumann-Morgenstern utility theorem suggested that based on certain axiom of rational behavior, risky choices will be respond-
ed with maximizing higher utilities in the future.
5 In this model, Allingham & Sandmo (1972) emphasized that total contingency cost if the taxpayers were caught was the
undeclared amount (gap between actual income W and declared income X), at a penalty rate π and larger than tax rate θ.
6 Harvey S Rosen and Ted Gayer (2014) suggested that the model also predicts that evasion decreases when marginal tax rates
reduced, since a lower value of t decreases the marginal benefit of evasion, and when there is no intersection between marginal
cost curve of detection and the marginal benefit curve for evasion, the effective dollars of underreporting will be zero.
6
Agung Endika Satyadini, Restu Rea Erlangga, Brigitta Steffi VDKG / Who Avoids Taxes? (2019) 1-26
8Sincethe targeted population on this research is homogeneous upper-middle-taxpayers, size of company is not mandatory
relevant with age of company. This issue was already addressed in OLS assumption model (iid: independently, identically distribut-
8
Agung Endika Satyadini, Restu Rea Erlangga, Brigitta Steffi VDKG / Who Avoids Taxes? (2019) 1-26
9
Agung Endika Satyadini, Restu Rea Erlangga, Brigitta Steffi VDKG / Who Avoids Taxes? (2019) 1-26
3.3 Empirical Model for Non- To calculate total assets, this study uses
Conforming Tax Avoidance taxpayers’ encrypted ID as the lookup key.
Determinants LAUDITRESULT is the amount of
underpayment assessment letters issued
To investigate the determinants of to respective taxpayers at a period before
non-conforming tax avoidance, this respective years including principal, fines
paper employs ETR as a tax avoidance and additional penalties (in natural log) as
measurement. This model estimates the a result of tax examination process. The
significance of determinants correlated amount of underpayment assessment
to changes (reduction) of tax income as letter includes tax assessment for all taxes,
a ratio of accounting income by using but limited to tax underpayment assess-
the following regression equation: ment (SKPKB), tax additional underpay-
ment assessment (SKPKBT) and notice of
tax collection (STP)10. By way of illustration,
LAUDITRESULT is derived from Audit
Assignment (SP2) data which is catego-
rized as “special audit”, not “routine audit”
of the relatively homogeneous
Where
upper-middle-taxpayers. It is reasonable
i : taxpayers’ ID since the special audit assignments are
t : year
Zi: taxpayers’ ID fixed-effects
conducted based on preliminary risk anal-
Tt: year fixed-effects ysis and indication of non-compliance,
meaning that the samples of audit results
Dependent Variable are closely related to indication of tax
ETR is effective tax rate, formulated as avoidance. Interestingly, this research finds
total tax expense divided by pre-tax no observation of overpayment audit
income. This study estimates the value result (SKPLB).
of ETR using tax return’s main data TAXRATEGAP is the difference
(form 1771). between Indonesia’s statutory tax rates
Independent Variables and counterpart’s tax rate. The counter-
LSIZE is the size of company, measured part is a country where the related party
by total assets (in natural log)9. To pro- of an Indonesian taxpayer is located11.
vide precise estimation, this study com- Information captured in tax return sum-
pares the amount of total assets mary is the largest related party (scale 1)
according to tax return attachment based on Form 1771 Annex 3A and 3A-2.
(Form 1771 Special Annex 8A-1) and
total assets according to financial state-
ment transcript.
9A number of literature occupied definition of size of enterprises as a function of enterprises’ total assets, revenues and value of equities. However,
since ETR is closely related to income statement, hence to minimize simultaneous causality bias, this paper adopts a function of enterprise’s size from
balance sheet item.
10As regulated by Indonesian Law Number 16 of 2000 concerning General Provision and Tax Procedures, DGT authorized to issue tax assessment letter
including tax underpayment assessment (SKPKB), additional tax underpayment assessment (SKPKBT), tax overpayment assessment (SKPLB), nil tax
assessment (SKPN), and collection letter (STP). Avoiding misinterpretation, the sample selection for this variable is limited to non-zero and positive
assessment letters, due to difficulty to distinguish between “no audit” or “no findings” if the value is zero. However, the number of excluded zero-value
observations are relatively small (97 observations), hence it might not alter the overall estimation.
10
Agung Endika Satyadini, Restu Rea Erlangga, Brigitta Steffi VDKG / Who Avoids Taxes? (2019) 1-26
11 In corporate tax return, tax payers are required to fulfil the appendix 3A of Tax Return 1771. The information covers the name of affiliated party,
addresses, and transfer pricing methodology applied to that taxpayers. The affiliated party is not always a foreign company. For local company
TAXRATEGAP will be equal to zero.
11
Agung Endika Satyadini, Restu Rea Erlangga, Brigitta Steffi VDKG / Who Avoids Taxes? (2019) 1-26
12
Agung Endika Satyadini, Restu Rea Erlangga, Brigitta Steffi VDKG / Who Avoids Taxes? (2019) 1-26
13
Agung Endika Satyadini, Restu Rea Erlangga, Brigitta Steffi VDKG / Who Avoids Taxes? (2019) 1-26
From the point of view of preventive (1997), Rego (2003), Zimmerman (1983),
strategy, tax assistance has a positive Omer et al. (1993) and Armstrong et al.
significant correlation with ETR, which (2011) the lower value of ETR represents
means that assistance will increase tax the higher level of tax avoidance. Hence, it
compliance, or in other words, taxpay- suggests that company with a bigger size
ers tend to avoid less taxes. The similar has a weaker tendency of avoiding taxes.
positive significant correlation with ETR This result is consistent with Noor et al.
is also demonstrated by the age vari- (2010), who found a positive significant
able, which suggests that older compa- correlation between company’s size (as
nies tend to avoid less taxes. However, single function of total assets) and ETR.
the tax rate differences between host Moreover, this result is also persistent with
country and related party’s country Slemrod (2007).
have a negative significant correlation The possible explanation why a
with ETR, which suggests that if the tax bigger enterprise in this study tends to
rate gap is positive and larger (tax rate more comply than smaller enterprises is
on the related party’s country is higher that the long-run business conducted by
than the host country), taxpayers tend companies as operating subsidiaries in
to avoid less taxes. Conversely, if the tax Indonesia. Operating company is a part of
rate gap is negative (tax rate on the multinational company which operates in
related party’s country is lower than the a resource country to exploit inputs (raw
host country), taxpayers tend to avoid materials, labors, etc.). Mostly, the operat-
taxes in the host country. However, the ing companies are fully fledged manufac-
result on Regression 1 may be suffered turers14. OECD (2010) explained that fully
by omitted variable bias. To overcome fledged manufacturers will assume a
this problem and to ensure the robust- larger range of business functions and
ness of the model, this study gradually risks, including production, R&D and
adds the entity fixed-effects and time intangible management, so they will
fixed-effects with clustered standard develop a well-managed company and
error for Regression (2) - (4). conduct a long-run business. Therefore,
big companies mitigate their risks by
4.1.2 Size of Enterprise and taking long-term strategies including
Non-Conforming Tax tax-compliance strategies. In the opposite
Aggressiveness way, non-operating company such as a
paper company mostly has zero assets,
Providing more rigorous result and min- which conducts no business but is proper-
imizing omitted variable bias, Regres- ly constituted and incorporated in one
sion 4 reveals that size of company has country only for registration certificate to
a positive significant correlation with access tax benefits.
ETR. According to previous research
conducted by Gupta & Newberry
14 Fully fledge manufacturing is responsible for sourcing materials, undertaking production and potentially selling to third parties at its own risk as well
as to related party distributors. Moreover, they also possible to establish intangible properties by Research and Development (R&D) activities.
14
Agung Endika Satyadini, Restu Rea Erlangga, Brigitta Steffi VDKG / Who Avoids Taxes? (2019) 1-26
4.1.3 Audit Penalties and Non- positive correlation between audit proba-
Conforming Tax Aggressiveness bility and compliance.
However, this paper assumes that
Regression 4 also demonstrates audit probability in all samples area is
that penalty has a positive significant equal since the sample is taken from rela-
correlation with ETR. It implies that tively homogenous taxpayers. This
higher penalty will stimulate taxpayers assumption is relevant with Hasseldine
to be less tax aggressive. This result is (1993), who suggested that targeted tax
consistent with Slemrod’s (2007) audit for homogeneous taxpayers seems
emphasiing on Becker’s (1968) study to be more effective in increasing tax
about how an enterprise established its compliance rather than random audit.
strategy related to tax aggressiveness This explanation is supported by Witte &
decision to maximize its utility by con- Woodbury’s (1985) argument that
sidering possible penalties (as contin- explained that the influence of audit prob-
gent costs). This result, to some extent, abilities on tax compliance varied depend-
is also consistent with Beck et al. (1991) ing on the group of taxpayers: strong
and Park &Hyun (2003), who provided influences demonstrated by sole propri-
evidence of the positive correlation etors (large taxpayers), and weak influenc-
between size of penalty and tax compli- es demonstrated by salaried taxpayers
ance. Investigating the size of enterprise (small taxpayers).
or individual as one of the taxpayer’s
characteristic and taking into account 4.1.4 Tax Rate Differential and
that the sample is taken from medi- Non-Conforming Tax
um-large taxpayers, the result is also Aggressiveness
consistent with Witte & Woodbury
(1985) who acknowledged the positive Regression 1 demonstrates that tax
correlation between penalty and tax rate differential has a significant negative
compliance for large taxpayers. correlation with ETR. It implies that if the
Prior studies provided evidence gap between Indonesia’s statutory tax
of a correlation between audit probabil- rates and counterpart’s tax rate is negative
ity and tax aggressiveness. Generally, (meaning that counterpart’s tax rate is
audit probability is also considered as higher), the reported ETR in Indonesia
one of the deterrent factors that chang- should be higher, and vice versa. This
es taxpayer’s compliance behavior. result, to some extent, can be analogized
Some studies demonstrated a negative with the profit or cost-pooling strategy to
correlation between audit probabilities minimize worldwide tax burden. Profits will
and tax aggressiveness. In accordance be shifted to the lower tax rate countries,
with this explanation, to some extent, and costs will be dumped to higher tax
Fischer, Wartick, & Mark (1992) as cited rate countries. This scheme will result in
by Chau & Leung (2009) revealed a lower profitability for high tax rate coun-
tries and higher profitability for lower tax
rate countries.
15
Agung Endika Satyadini, Restu Rea Erlangga, Brigitta Steffi VDKG / Who Avoids Taxes? (2019) 1-26
Adopting fixed effects and clustered 4.1.5 Assistance and Non-Conforming Tax
standard error, Regression 4 demon- Aggressiveness
strates insufficient evidence to conclude
significant correlation between the tax Using similar model with the previ-
rate differential and ETR. Empirically, ous variables, Regression 4 displays a pos-
possible reason why adoption of fixed itive significant correlation between assis-
effects results in insignificant correlation tance and ETR. Logically, higher effort of
is because of tax rates in many countries tax assistance in current period will stimu-
are almost time invariant. It means that late taxpayers to be less tax aggressive.
tax rates are relatively constant, so that Rather than deterrent effect as produced
fixed effects cannot capture the data by audit penalties, tax assistance provides
variation over the observation period. preventive and educative approach for
Conceptually, the insufficient taxpayers. This result is consistent with Ola
evidence of correlation between tax (2001) as cited by Ebimobowei & Peter
differentials and ETR also can be (2012) who provided evidence that tax
explained by using the same analogy of assistance was strongly related to taxpay-
capital inflow or outflow as an impact of er’s compliance, also Torgler & Schneider
tax differentials as suggested by (2006) who revealed that majority of their
Bénassy-Quéré et al. (2005). They respondent confirmed that tax knowledge
explained asymmetry in the impact of assistance would influence the willingness
tax differentials on investment: lower to pay taxes.
tax rates in the recipient countries fail to DGT (2016) explained two major
significantly attract foreign investment, strategies to inflate the taxpayers’ compli-
while higher taxes tend to discourage ance. First, prevention strategy by adopt-
new FDI inflows. They also suggested ing counseling and tax education for
that the impact of positive tax differen- taxpayers. Second, reaction strategy by
tials is not homogeneous regarding the conducting tax examination. Comparing
tax treaty arrangement in countries. the responsiveness of those two variables
Another possible reason can be in this model, Regression 4 presents an
explained by using Hybrid Mismatch evidence that based on empirical analysis,
Arrangement concept as suggested by tax assistance has higher responsiveness
OECD. Due to hybrid entity arrange- (coefficient value: 0.0402) rather than
ment, it is possible if profits shifted from audit penalty (coefficient value: 0.0103).
a country is not subject to tax in another Intuitively, it demonstrates that the appli-
country. Therefore, reduction in profit- cation of tax authority’s strategy to hike tax
ability in higher tax rate country is not compliance should be more on preven-
always related to increase profitability in tion rather than reaction.
lower tax rate country. However, to
examine this phenomenon, individual
level country-by-country data is
required.
16
Agung Endika Satyadini, Restu Rea Erlangga, Brigitta Steffi VDKG / Who Avoids Taxes? (2019) 1-26
The positive significant correlation with related party’s country is lower than the
TAXOCF is also demonstrated by the host country), taxpayers tend to avoid
age variable. Intuitively, it shows that more taxes in the host country.
older companies tend to avoid less The result on Regression 1 may be
taxes. However, the tax rate difference suffered by omitted variable bias. To over-
between the host country and the relat- come this problem and to ensure the
ed party’s country has a negative signif- robustness of the model, this study gradu-
icant correlation with TAXOCF, it means ally adds the entity fixed-effects and time
that if tax rate difference is positive (tax fixed-effects with clustered standard error
rate on the related party’s country is in Regression 2 to 4. The result is relatively
higher than host country), taxpayers consistent except for TAXATEGAP and
tend to avoid less taxes in the host LCSTOCK.
country. Conversely, if the tax rate
difference is negative (tax rate on the
18
Agung Endika Satyadini, Restu Rea Erlangga, Brigitta Steffi VDKG / Who Avoids Taxes? (2019) 1-26
4.2.2 Size of Enterprise and Con- certificate to access tax benefits such as
forming Tax Aggressiveness treaty shopping and profit shifting, usually
has zero asset.
Providing more rigorous result
and minimizing omitted variable bias, 4.2.3 Audit Penalties and Conforming
this model also employs both entity Tax Aggressiveness
fixed-effects and time fixed-effects with
clustered standard error as demonstrat- Rigorous result using both entity
ed by Regression 4. Examining the fixed-effects and time fixed-effects with
result on Regression 4, it is rigorously clustered standard error as summarized in
demonstrated that size of company has Regression 4 demonstrates that penalty as
a positive significant correlation with a product of audit on previous period also
TAXOCF. It suggests that the bigger has a positive significant correlation with
company’s size (in terms of total assets), TAXOCF in current period. It can be inter-
the lower tendency of avoiding taxes. It preted that higher penalty in previous
is difficult to compare and analyze the period will stimulate taxpayers to be less
consistency of this result with prior stud- tax aggressive in terms of both tax and
ies, since to the author’s best knowl- accounting reporting.
edge, conforming tax avoidance is rela- Different with ETR, TAXOCF model
tively unexplored rather than non-con- captures reduction of both accounting
forming tax avoidance and only a few and tax incomes. Comparing the coeffi-
literatures explore the measurement of cient value of AUDITRESULT between
conforming tax-avoidance. However, to non-conforming and conforming tax
some extent, this result is consistent avoidance, generally non-conforming tax
with Slemrod (2007), who relied on the avoidance (ETR) provides higher value
U.S General Accounting Office data, coefficient rather than conforming tax
estimating that big enterprises tend to avoidance (TAXOCF). Intuitively, it means
more comply than the smaller one. that in the context of tax avoidance mea-
Similar explanation with surement, audit penalty is more respon-
non-conforming tax avoidance subsec- sive to non-conforming tax avoidance,
tion, the operating enterprise will which is reduction of tax income relative to
assume a larger range of business func- accounting income. Applying similar logic
tions and risks, and in the long-run with non-conforming tax avoidance, tax
business cycle, they will develop a penalty can reduce the tax aggressiveness
well-managed company. Therefore, big in terms of tax and accounting incomes.
companiesmitigate their risks by taking However, as suggested by Becker (1968)
long-term strategy including tax-com- and emphasized by Allingham and
pliance strategy. In opposite way, Sandmo (1972), the taxpayer’s compliance
non-operating enterprise such as a also depends on audit probability and tax
paper company which conducts no rate.
business but is properly constituted and
13
19
Agung Endika Satyadini, Restu Rea Erlangga, Brigitta Steffi VDKG / Who Avoids Taxes? (2019) 1-26
4.2.4 Tax Rate Differential and DGT (2016) explained two major
Conforming Tax Aggressiveness strategies to increase taxpayer’s compli-
ance. The first one is prevention strategies
Using similar estimation of both by adopting assistance and tax education
entities fixed-effects and time fixed-ef- for taxpayers. The second one is reaction
fects, Regression 4 provides insufficient strategies by conducting tax audit. This
evidence to conclude a significant study finds that both tax assistance and tax
correlation between tax rate differential audit have effective impact in reducing tax
and TAXOCF. As discussed in the previ- aggressiveness. DGT can formulate
ous subsection, a possible reason why prevention or reaction strategies prior to
adoption of fixed effects resulted in addressing specific cases to minimize
insignificant correlation is because of losses from tax avoidance. However,
tax rates in many countries are almost further examination on individual cases
time invariant, so that fixed effects should be considered to determine the
cannot capture the variation of data comparative effectiveness of both strate-
over the observation period. As gies15.
explained before, the insufficient
evidence of correlation between tax 4.2.6 Age of Enterprise and
rate differential and TAXOCF also might Conforming Tax Aggressiveness
be caused by several reasons including
enterprises’ model and typologies, Broadly like ETR analysis as shown in Table
nature of investment in Indonesia and 2, Regression 4 on Table 3 provides an
hybrid mismatch arrangement. evidence of negative significant correla-
tion between enterprise’s age and
4.2.5 Assistance and Conforming Tax TAXOCF. Intuitively, it means that older
Aggressiveness enterprises tend to have lower TAXOCF or
more tax aggressive in terms of both
As showed in Regression 4 on Table 3, reduction of tax and accounting incomes.
conforming tax avoidance measure- The previous analysis also suggested that
ment captures a positive significant the older company, the more efficient
correlation between assistance and management will be, in this case, man-
TAXOCF. This result is broadly consistent agement efficiency including tax manage-
with non-conforming tax avoidance ment efficiency (tax planning). In the simi-
measurement as shown in Table 2. The lar vein, Arrow (1962), Jovanovic (1982)
positive significant correlation implies and Ericson and Pakes (1995) suggested
that higher intensity of assistances or tax that company’s age could affect efficient
education efforts will stimulate taxpay- management, with discovery and
ers to be less aggressive. improvement of management including
tax management.
15 It is still difficult to confidently suggest the comparative effectiveness of both strategies in this model, because of different specification data between
audit penalties and assistance.
20
Agung Endika Satyadini, Restu Rea Erlangga, Brigitta Steffi VDKG / Who Avoids Taxes? (2019) 1-26
Both Regression Ext.1 and Ext.2 on Table The second reason is related to risk aver-
4 demonstrate that size of company, sion. The foreign-controlled enterprises
age, and capital stocks are more may consider that audit penalty is less
responsive to tax avoidance for PE. risk-significant to their tax avoidance deci-
Another evidence shows that tax assis- sion so that the tax aggressiveness deci-
tance is more responsive to tax avoid- sion is relatively irrelevant to the amount
ance for foreign-invested companies, it of audit penalty.
suggests that the impact of tax assis-
tance is relatively higher for foreign-in- 4.2.5 Assistance and Conforming Tax
vested companies than full sample. Aggressiveness
Analyzing the second variable,
audit penalty has a less impact on tax Providing evidences of tax avoid-
avoidance for foreign-controlled enter- ance responsiveness for specific sectors,
prises. There are two possible reasons this model stratifies two groups of panel
for this phenomenon, the first one is data: (1) Financial Sector and (2) Mining
related to marginal cost and benefit of Sector. This stratification is paramount to
tax avoidance. If the taxpayer considers demonstrate tax avoidance risks for spe-
marginal penalty as marginal cost of cific sectors in Indonesia.
detection, the marginal cost of detec-
tion may be much higher than the mar-
ginal benefit.
21
Agung Endika Satyadini, Restu Rea Erlangga, Brigitta Steffi VDKG / Who Avoids Taxes? (2019) 1-26
22
Agung Endika Satyadini, Restu Rea Erlangga, Brigitta Steffi VDKG / Who Avoids Taxes? (2019) 1-26
Taken together, these results are para- 5.1 Limitations and Future Research
mount as the empirical approach in tax
policy formulation. Tax avoidance pro- This paper does not distinguish between
files as presented in this research in legal tax avoidance and illegal tax evasion,
relevant with the risk engine core of because of practical complication to cate-
Compliance Risk Management (CRM) gorize them on empirical data. This paper
adopted by DGT. Another important also uses limited variables and employs
practical implication is the relevance of relatively homogeneous upper-mid-
these results to Risk-Based Audit to dle-taxpayers due to data access limita-
pursue the efficient audit coverage as tions. Related to penalties, this paper
depicted in Figure 1. Ensuring the assumes that probability of audit is con-
best-fit policy formulation, these results stant for all taxpayers due to the difficulty
are also pertinent with prevention or to measure the audit rate during observa-
reaction strategies to minimize losses tion period. Future study may investigate
from tax avoidance. In this sense, tax avoidance behavior for a larger range
revealed that application of tax authori- of taxpayers and employ more relevant
ty’s strategy to hike tax compliance variables. Furthermore, using various
should be more likely to prevention measurements for both non-conforming
rather than reaction. Furthermore, in and conforming tax avoidance is also
the brain area of academic research, the beneficial for the development of tax liter-
findings also contribute to the field of ature in the future.
tax literature by providing simultaneous
empirical models including conforming
and non-conforming tax avoidance,
which have been relatively unexplored
in prior studies.
24
Agung Endika Satyadini, Restu Rea Erlangga, Brigitta Steffi VDKG / Who Avoids Taxes? (2019) 1-26
[27] Lisowsky, P. (2010). "Seeking Shelter”: Empir- Empirical Evidence from An Emerging Econo-
ically Modeling Tax Shelters and Examining my. International Journal of Business and
Their Link to the Contingent Tax Liability Society, Vol. 14 No. 3, 412 - 427.
Reserve. The Accounting Review Vol. 85, [40] Sandmo, M. G. (1982). Income Tax Evasion: A
1693-1720. Theoretical Analysis. Journal of Public
[28] Loderer, C., & Waelchli, U. (2009). Firm Age Economics 1, 323-338.
and Performance. SSRN. [41] Sarker, T. K. (2003). Improving Tax Compliance
[28] Markle, K. S., & Shackelford, D. A. (2012). in Developing Countries via Self-Assessment
Cross-Country-Comparisons of Corporate Systems - What Could Bangladesh Learn from
Income Taxes. National Tax Journal, Japan? Asia-Pasific Tax Bulletin Vol. 9, No. 6.
493–528. [42] Satyadini, A. E. (2018). Empirical Model of Tax
[29] Mills, L., Erickson, M. M., & Maydew, E. L. Avoidance In Indonesia. Journal of Economy
(1998). Investment in Tax Planning. Journal and Finance Studies, 52-69.
of the American Taxation Association, 1. [43] Satyadini, A. E., Basir, A., & Barata, A. (2019).
[30] Mohdalia, R., Isab, K., & Yusoffb, S. H. Modern Customs Risk Management Frame-
(2014). The impact of threat of punishment work: Improvement towards Institutional
on tax compliance and noncompliance Reform. International Journal of Innovative
attitudes in Malaysia. Procedia - Social and Science and Research Technology , 60-69.
Behavioral Sciences 164, 291-297. [44] Shaxson, N. (2011, January 9). Explainer: what
[31] OECD. (2015). Reforms to the International is a tax haven? Retrieved from The Guardian:
Tax System for Curbing Avoidance by Multi- h t t p s : / / w w w. t h e g u a r d i a n . c o m / b u s i -
national Enterprises. OECD Publishing. ness/2011/jan/09/explainer-what-is-tax-haven
[32] Omer, T. C., Molloy, K. H., & Ziebart, D. A. Sikka, P., & Willmott, H. (2010). The Dark Side
(1993). An Investigation of The Firm Size - of Transfer Pricing: It’s Role in Tax Avoidance
Effective Tax Rate Relation In The 1980s. and Wealth Retentiveness. Critical Perspec-
Journal of accounting, auditing & finance.- tives on Accounting. Critical Perspectives on
Thousand Oaks Vol. 8., 167-182. Accounting, 42-56.
[33] Palan, R. (2009, October 01). History of tax [45] Slemrod, J. (2007). Cheating Ourselves: The
havens. Retrieved from historyandpolicy: Economics of Tax Evasion. Journal of Econom-
http://www.historyandpolicy.org/policy-pa- ic Perspectives Volume 21, Number 1, 25-48.
pers/papers/history-of-tax-havens The International Consortium of Investigative
[34] Park, C. G., & Hyun, J. K. (2003). Examining Journalists. (2016). The Panama Papers. The
the determinants of tax compliance by International Consortium of Investigative
experimental data: a case of Korea. Journal Journalists.
of Policy Modeling, 673-684. [46] Torgler, B., & Schneider, F. (2006). What
[35] Publish What You Pay Indonesia. (2015). Shapes Attitudes Toward Paying Taxes?
Illicit Financial Flows And Tax Crime In Evidence from Multicultural European Coun-
Mining Sector. tries. IZA Discussion Paper No. 2117.
[36] Rego, S. O. (2003). Tax-Avoidance Activities [47] Wilson, R. (2009). An Examination of Corpo-
of U.S. Multinational Corporations. rate Tax Shelter Participants. Accounting
Contemporary Accounting Research Vol. Review 84, 969-99.
20 Issue 4, 805. [48] Witte, A. D., & Woodbury, D. F. (1985). The
[37] Rego, S., & Wilson, R. (2012). Equity Risk Effect of Tax Laws and Tax Administration on
Incentives and Corporate Tax Aggressive- Tax Compliance: The Case of The U.S Individ-
ness. Journal of Accounting Research 50, ual Income Tax. National Tax Journal Vo. 38
775-810. No.1, 1-13.
[38] Robinson, J. R., Sikes, S. A., & Weaver, C. D. [49] Yitzhaki, S. (1972). A Note on Income Tax
(2010). Performance Measurement of Evasion: A Theoretical Analysis. Journal of
Corporate Tax Department. The Account- Public Economics, 323-338.
ing Review Vol 83 No.3, 1035–1064. [50] Zimmerman, J. L. (1983). Taxes and firm size.
[39] Salihu, I. A., Obid, S. N., & Annuar, H. A. Journal of Accounting and Economics, Vol. 5,
(2013). Measures of Corporate Tax Audit: Issue 1, 119-149.
26