EDUCORETAX
Volume 6 No.
7, 2026
THE INFLUENCE OF SUPERVISORS AND DIFFERENCES IN
OWNERSHIP ON TAX AVOIDANCE BY INDONESIAN INSURANCE
COMPANIES
Cecilia Ancelin Feodora Anthony.
Ferry Irawan .
* 2210112094@mahasiswa.
Universitas Pembangunan Nasional AuVeteranAy Jakarta ferryirawan@upnvj.
Universitas Pembangunan Nasional AuVeteranAy Jakarta corresponding author Abstract Taxes are an important instrument for the sustainability of the state because taxes finance more than 82% of state However, there are still many companies that engage in tax avoidance practices to increase their These legal practices, which exploit grey areas in tax legislation, cost the state around $2 billion.
Tax avoidance factors can vary, some of which include the role of supervisors and owners who should be instrumental in curbing tax avoidance practices so as not to cause harm to either the company or the state.
This study aims to determine the effect of institutional ownership, independent commissioners, and gender diversity on the board of directors on tax avoidance.
The selected companies are insurance companies listed on the Indonesia Stock Exchange during the period of 2019 to 2024.
To test this, a quantitative method was used with secondary data sources, derived from company annual reports, with a total of 96 observations using purposive sampling.
Data analysis was conducted using panel data regression and t-tests using STATA v.
17 software.
The results showed that the moderating effect of profitability weakened the relationship between institutional ownership and tax avoidance, while institutional ownership, independent commissioners, gender diversity in the board of directors, and the role of profitability, as well as the relationship between independent commissioners and board gender diversity, did not affect tax avoidance in insurance companies listed on the IDX.
Keywords: Gender diversity.
Independent commissioners.
Institutional ownership.
Insurance Companies.
Tax
INTRODUCTION
In the operational activities carried out by the state, all costs incurred naturally come from state revenue.
This state revenue is divided into two categories, namely taxes and non-tax state revenue.
To date, more than 82% of state revenue comes from taxes (Badan Pusat Statistik.
This tax itself is an important instrument of state finance used to carry out development in order to improve the welfare of the people (Satriya et al.
, 2.
Figure 1.
Tax Revenue Chart in Indonesia Source: Organisation for Economic Co-Operation and Development .
Processed Data .
However, not all individuals or companies are willing to pay all their tax obligations to the state because they do not want to lose their income.
Therefore, many companies engage in tax avoidance practices (Sahrir et al.
, 2.
This tax avoidance practice is a common tactic used by companies to reduce their tax burden by exploiting grey areas or loopholes in tax legislation (Novianto, 2.
It is legal to do so, but it is immoral and will cause losses because not all tax revenues will be collected by the state (Nabilla et al.
, 2.
In detail, several Page | 419
EDUCORETAX
Volume 6 No.
7, 2026
government regulations related to the preparation of tax avoidance were made on Law No.
of 2008 on the Fourth Amendment to Law No.
7 of 1983 on Income Tax .
, the government imposes a minimum capital requirement of 50% or a capital-to-liability ratio (Debt to Equity Rati.
However, tax avoidance still occurs.
Indonesia loses around US$2 billion in taxes transferred to tax havens, contributing around 0.
13% to global losses.
(OECD, 2.
DDTC
News .
reported that Swiss Life, an insurance company of a US client, provided insurance policies to the client on behalf of the company to cover the value of the US citizen's portfolio in 2020 amounting to Rp 1 trillion.
There are various factors that can lead to tax avoidance.
Institutional ownership refers to the proportion of shares held by state institutions, insurance companies, banks, and foreign investors (Dewi, 2.
In general, the higher the shares owned by institutions, the more likely management will adopt policies that reduce tax payments, which should increase profitability (Phandi & Tjun, 2.
On the other hand, supervisors can also come from independent Independent commissioners are supervisors who are not from within and/or have no connection with the company (Andreas et al.
, 2.
These independent commissioners are believed to be able to directly influence the fate of the company because they must remain neutral, especially in complying with state taxation regulations.
Independent commissioners are supervisors who are not from within and/or have no connection with the company (Mishra & Jhunjhunwala, 2.
Profitability can play a role in tax avoidance.
The link itself with tax avoidance is that if a company has high profitability, it will incur higher tax burdens, therefore tax avoidance behaviour will increase in order to maintain profitability (Norisa et al.
, 2.
Due to the differences in research results and the continuing phenomenon of tax avoidance, a study was conducted entitled.
AuThe Influence of Supervisors and Differences in Ownership on Tax Avoidance by Indonesian Insurance CompaniesAy using profitability as a moderator and company size and capital intensity as control variables.
The purpose of this study is to examine the effect of institutional ownership, independent commissioners, and gender diversity in the board of directors on tax avoidance practices in insurance companies in Indonesia.
It is hoped that this study will contribute to the development of knowledge in the field of corporate governance, providing insight to accounting or other economics students on the factors that influence tax avoidance practices in Indonesia.
It is also hoped that this study will provide general knowledge about the impact of tax avoidance decision-making in Indonesia.
LITERATURE REVIEW
Agency Theory The agency theory itself was popularised by Jensen and Meckling, which divides a company into agents and principals.
Agents here are actors who carry out the company's operational tasks as delegated by the principals .
(Jensen & Meckling, 1976.
Unfortunately, this agency theory will result in information asymmetry.
Information asymmetry is an imbalance of information within a company.
Agents who carry out operational work have more knowledge than principals who do not supervise the day-to-day running of the company (Muuna et al.
, 2.
Managers can also take advantage of this asymmetry.
managers make decisions that benefit themselves, this will harm the company and jeopardise its stability, thus requiring the role of a supervisor to deal with the impact of this information asymmetry (Alawiyah & Wulandari, 2.
Tax Avoidance One of the effects of information asymmetry is tax avoidance.
Tax avoidance is the method of exploiting loopholes in legislation that does not regulate such efforts (Ghozali, 2.
Tax avoidance is included in tax planning or more aggressive tax planning, but cannot be Page | 420
EDUCORETAX
Volume 6 No.
7, 2026
penalised because there are no regulations governing tax savings (Hoffman, 1961.
Tax avoidance can be measured by BTD or Book Tax Difference.
This ratio shows the difference between fiscal tax income and financial statement tax income (Nukman et al.
, 2.
The higher the BTD, the greater the tax avoidance measures taken by the company (Hanlon & Heitzman, 2010.
Current Tax Expense .
rofit before tax Oe ( Tax Rate
BTD =
Total Asset Institutional Ownership Institutional ownership will mediate agency problems that arise.
Institutional companies that are shareholders can influence managers' behavior by controlling their decisions (Vemberain & Triyani, 2.
A calculation ratio that divides the shares owned by institutions by the number of outstanding shares (Azwar & Fitrijanti, 2.
Oc Shares owned by institutions Institutional Ownersip = y 100% Oc Outstanding shares Independent Comissioners Independent commissioners are positions within a company that play a supervisory role over the board of directors and must have no connection to the company, whether through share ownership or relationships with other employees (Doho & Santoso, 2.
This study uses the ratio of independent commissioner competence to determine how many independent commissioners are competent, namely by dividing the number of independent commissioners who have an education in economics by the total number of independent commissioners (Zein et al.
, 2.
Oc Independent commissioner with an economic background
IC =
y 100% Oc Independent comissioner Board Gender Diversity Gender diversity is one of the factors that can balance decision-making (Hindasah & Harsono, 2.
Women are considered to be more sensitive and more careful in making In calculating board diversity, the diversity ratio can be used as a representation.
The formula for determining gender diversity is to compare the total number of women serving on the company's board of directors with the total number of directors in the company (Rahman.
Oc Female Directors Board Gender Diversity = y 100% Oc Directors Profitability When considering investing in a company's shares, investors will use ROA as a tool to measure the efficiency of the company's use of assets for investment (Jefriyanto, 2.
The formula is (Putra & Zahroh, 2.
Profit after tax
ROA =
y 100% Total Asset Firm Size With this company size, the company's ability to conduct operations and the stability of its performance can be measured (Ariska et al.
, 2.
The size of a company can be measured by using log natural total asset.
This is because it has a better level of stability compared to using other proxies and is more likely to be maintained continuously between periods (Saraswati & Bernawati, 2.
Firm Size = ln(Total Asse.
Page | 421
EDUCORETAX
Volume 6 No.
7, 2026
Capital Intensity This capital intensity will describe the ratio between total fixed assets and total assets owned by the company (Rusmanto et al.
, 2.
In describing the extent to which a company's fixed assets compare to its total assets, the fixed asset intensity ratio is used (Pertiwi & Purwasih, 2.
Total Fix Asset Capital Intensity = y 100% Total Asset Institutional ownership explains how many shares are owned by other institutions or the government or legal entities whose existence is expected to mediate conflicts caused by information asymmetry between management and stakeholders (Pratomo & Rana, 2.
Institutions will exercise greater control over the behaviour of management who can make decisions on tax avoidance (Dewi & Oktaviani, 2.
Several studies support this negative influence, including: Sumekar et.
Yuni & Setiawan .
, and Hendrianto .
H1 = Institutional Ownership Variables Have a Negative Effect on Tax Avoidance In addition to differences in ownership, in establishing good corporate governance, owners must reduce information asymmetry by conducting supervision (Jensen & Meckling.
The presence of independent commissioners within a company is important and mandatory because they are expected to control the behaviour of management within the It is these independent commissioners who protect the company's reputation, which could be damaged by non-compliant actions of management (Dewi, 2.
Independent commissioners must have adequate skills or knowledge.
This includes education or training in economics and company conditions so that they can make better decisions (Purwaningsih & Mardiana, 2.
This research is in line with research conducted by Dewi & Oktaviani .
and Deviansyah, et.
H2 = Independent Commissioner Variables Have a Negative Effect on Tax Avoidance Within supervision, female directors can also encourage collaborative decision-making, strengthen cooperation, and boost company performance (Sohdi, 2.
The presence of more female directors can be expected to increase the number of decisions made.
Their more meticulous and sensitive nature, as well as their avoidance of challenges, will lead to decisions that reduce tax avoidance practices (Zulfa et al.
, 2.
This statement is consistent with research conducted by Jarboui, et.
and Zhang, et.
H3 = Board Gender Diversity Variable Has Negative Impact on Tax Avoidance The higher the return on total assets of a company, the better its performance (Gitman & Zutter, 2015.
Therefore, when income is high, this will increase the tax burden, which management seeks to avoid.
This conflicts with institutional ownership, which expects companies to do everything in accordance with the law.
H4 = Profitability Variables Weaken the Negative Relationship between Institutional Ownership and Tax Avoidance High profits mean that the majority of company agents avoid paying tax because it would reduce their profits (Grace & Sihotang, 2.
Independent commissioners only serve as supervisors and are unable to make decisions because they oppose decisions and only serve as supervisors (Nurjanah & Aligarh, 2.
H5 = Profitability Variables Weaken the Negative Relationship between Independent Commissioners and Tax Avoidance When faced with the choice of maintaining profits and not taking risks, female directors will allow tax avoidance (Prasetyo, 2.
H6 = Profitability Variables Weaken the Negative Relationship between Board Gender
Diversity and Tax Avoidance Page | 422
EDUCORETAX
Volume 6 No.
7, 2026
METHODS
In this study, there were 16 insurance companies in Indonesia listed on the IDX in 20192024.
A quantitative approach was used to examine the data, and secondary data sources were used as the basis.
These sources were obtained from the IDX official website .
and other information was obtained from the websites of each banking company.
To test the data, this study used the panel data regression method with the help of STATA version 17, while the results of the hypothesis testing were used to draw conclusions from the study.
Panel Data Regression Model The panel data regression research model used is as follows:
Model 1 .
yaAycNyaycnyc = yu yu1 yaycCycnyc yuycayca yayayaycnyc yu3 yaAyayaycnyc yu1 yaycIycnyc yu2 yayaycnyc yceycnyc Model 2 .
ith moderatio.
yaAycNyaycnyc = yu yu1 yaycCycnyc yuycayca yayayaycnyc yu3 yaAyayaycnyc ycs1 ycEycIycCyaycnyc ycs2 yaycC O ycEycIycCyaycnyc ycs3 yayaya O ycEycIycCyaycnyc ycs4 yaAyaya O ycEycIycCyaycnyc yu1 yaycIycnyc yu2 yayaycnyc yceycnyc Book-Tax Difference (BTD), measurement of Tax Avoidance as dependent variable, is constant, 1, 2, 3 is the partial regression coefficient of the independent variable.
Z1.
Z2.
Z3 is the partial regression coefficient of the moderating variable, 1, 2 is the partial regression coefficient of the control variable, i is showing the order of temporary companies while t symbolizing the period, and A representing error.
RESULTS AND DISCUSSION
Descriptive Statistical Table 1.
Descriptive Statistical Analysis Test Results Variable BTD
ICC
BGD
ROA
Firm Size
Capital Intensity Description BTD
ICC
BGD
Obs Mean Tax Avoidance Institutional Ownership Competencies of Independent Commissioners Board Gender Diversity Std.
Dev.
ROA
Firm Size Capital Intensity Max Min Profitability Control Variable Control Variable Source: Data processed by researchers .
Output STATAv.
Based on the data, if a variable has a lower average value than its standard deviation, then the data is certainly not evenly distributed, and vice versa.
For example, which describes tax avoidance has a mean of 0.
This means that, on average, companies only engage in tax avoidance of 0.
Meanwhile, the standard deviation of 0.
2260161 is well above the This means that the distribution of tax avoidance is uneven.
some companies do not engage in tax avoidance, while others take full advantage of ways to avoid paying taxes.
Compared to the independent commissioner variable, which has a mean of 0.
8454866 and a standard deviation of 0.
On average, companies have 84.
5% competent independent The distribution in each company is average because the standard deviation has a value below the mean.
Data Panel Model Test Chow test examines which model is more suitable, common effect or fixed effect.
For the first regression data model, it has F probability value for Fixed Effects is 0.
5267 > prob
Page | 423
EDUCORETAX
Volume 6 No.
7, 2026
Lagrange Multiplier test that test examines which model is more suitable, common effect or random.
It shows that prob 0.
Based on both tests, it was found that Common Effect was actually more suitable for this research model.
Thus, the Hausman test was not necessary.
Based on both tests, it was found that the Common Effect was actually more suitable for this research Thus, the Hausman test was not necessary.
The first research model will be tested using the Common Effect Model.
Second regression data model chow test result was an F probability value for Fixed Effects is 0.
7032, larger than prob .
To test the Random Effects and OLS models, the Lagrange Multiplier was used.
The result was prob 0.
0650, which is larger than prob 0.
OLS .
r knownly as Common Effec.
was considered the best.
Similar to the first model, based on both tests, it was found that the Common Effect was actually more suitable for this research Thus, the Hausman test was not necessary.
The second research model will be tested with the Common Effect Model.
Classical Assumption Test Normality testing, to see if there are any outliers in this study, can be done by observing the skewness distribution between -3 and 3 and the kurtosis between -10 and 10.
The normality test results were above the approved skewness and kurtosis values, so a 5% winsorising treatment was required to obtain more uniform research results.
For multicollinearity, it can be determined whether or not there is a strong relationship between the independent variables This is proven by assessing that the variance influent factor (VIF) should not be more than 10 and below 0.
The first research model did not have multicollinearity issues.
Meanwhile, for the second research model, the variables showed VIF values greater than 10.
Therefore, treatment in the form of centring the data with multicollinearity was used (Kutner et , 2005.
The autocorrelation test examined the presence or absence of confounding variables in a given period and found that both research models were free from autocorrelation.
Shown by the value of Prob > F that is larger than 0.
For the first model is 0.
9551 and for the second model is 0.
The heteroscedasticity test showed variations in residuals from one observation period to another using the Breush-Pagan-Godfery test method.
The data results show that the first both of research model exhibits heteroscedasticity issues, it shows by the value of Prob > chi2 is smaller than 0.
05, that is 0.
Due to the heteroscedasticity issue, a treatment was applied by using robust additions to the regression testing model.
Panel Data Regression Testing the Coefficient of Determination (R.
The following are the results of the coefficient of determination test (R.
done using STATA v.
Table 2.
Results of the Coefficient of Determination Test (R.
First Model Number of Obs Prob>F R-Squared Source: Data processed by researchers.
STATA v.
17 output .
From the results of the tests that have been conducted, the coefficient of determination (R.
value is 0.
This means that the variables tested, namely institutional ownership (IO), independent commissioner competence (ICC), board gender diversity (BGD), as well as control variables in the form of firm size .
irm siz.
and capital intensity (CI) are able to influence tax avoidance (BTD) as the dependent variable with an effectiveness of 11.
This panel data regression model is considered valid because the Prob>F value is 0.
0011, which is smaller than However, there are 88.
5% factors outside the tested research model that explain the effect of tax avoidance.
Page | 424
EDUCORETAX
Volume 6 No.
7, 2026
Table 3.
Results of the Coefficient of Determination Test (R.
Second Model Number of Obs Prob>F R-Squared Source: Data processed by researchers.
STATA v.
17 output .
From the results of the tests that have been conducted, the coefficient of determination (R.
value is 0.
This means that the variables tested, namely institutional ownership (IO), independent commissioner competence (ICC), board gender diversity (BGD), moderation of profitability on institutional ownership (IOROA), independent commissioner competence (ICCROA), gender diversity of directors (BGDROA), as well as control variables in the form of firm size .
irm siz.
and capital intensity (CI) in the model explains 64.
41% of the variation in tax avoidance.
This panel data regression model is considered valid because the Prob>F value 000, which is smaller than 0.
There are 35.
59% factors outside the tested research model that explain the effect of tax avoidance.
t-Test The significance level is 5%.
Here are the results of the test using STATA v.
Table 4.
t-Test Results Variable _const ICC
BGD
FIRMSIZE
Regression Model Common Effect Model, with robust yaa>ya Coef.
P>t ya Hypothesis Prediction H1: H2: H3: - Conclusion Rejected Rejected Rejected Source: Data processed by researchers.
STATA v.
17 output .
With the common effect model, panel data regression was processed using STATA Based on the test results, several pieces of information were obtained, allowing us to conclude that the panel data regression model is as follows:
yaAycNyaycnyc = Oe0.
0027922yaycCycnyc 0.
0035079yayaya ycnyc 0.
0086086yaAyayaycnyc 0.
002779yaycIycnyc Oe 0.
0899956yayaycnyc The value of the constant coefficient () is approximately -0.
This means that if all variables, institutional ownership (IO), independent commissioner competence (IC-C), and board gender diversity (BGD), company size (Firm Siz.
, and capital intensity (CI) show a constant value of zero .
or remain unchanged, tax avoidance (BTD) will decrease by Table 5.
t-Test Results Second Model
Variable _const ICC
BGD
ROA_T
IOROA_C
ICCROA_C
BGDROA_W
Regression Model Common Effect Model, with robust yaa>ya Coef.
P>t ya Hypothesis Prediction Conclusion H4:
H5:
H6:
Accepted Rejected Rejected Page | 425
EDUCORETAX
Volume 6 No.
7, 2026
Variable _const FIRMSIZE Ci_W Regression Model Common Effect Model, with robust yaa>ya Coef.
P>t ya Hypothesis Prediction Conclusion Source: Data processed by researchers.
STATA v.
17 output .
With the common effect model, panel data regression was processed using STATA Based on the test results, several pieces of information were obtained, allowing us to conclude that the panel data regression model is as follows:
yaAycNyaycnyc = 0.
0195475 Oe 0.
0238341yaycCycnyc 0.
0002826yayaya ycnyc 0.
0142464yaAyayaycnyc Oe 0.
0007054ycEycIycCyaycnyc 1.
318485yaycC O ycEycIycCyaycnyc Oe 0.
0587691yayaya O ycEycIycCyaycnyc Oe 0.
1153048yaAyaya O ycEycIycCyaycnyc 0.
0003545yaycIycnyc Oe 0.
0359102yayaycnyc The value of the constant coefficient () is approximately 0.
This means that if all variables, institutional ownership (IO), independent commissioner competence (IC-C), board gender diversity (BGD), moderation of profitability on institutional ownership (IO*PROF), independent commissioners (ICC*PROF), board gender diversity (BGD*PROF), company size (FS), and capital intensity (CI) show a constant value of zero .
or remain unchanged, tax avoidance (BTD) it turned out that it actually will increase by 0.
Discussion