The Whistleblowing System, Literacy of Big Data Analytics, and Tax Avoidance: Experimental Study

Awareness of the existence of a system that can monitor non-compliance will certainly affect the behavior of taxpayers. Testing taxpayers' intentions to avoid taxes because of the whistleblowing system and big data analytics literacy is very relevant due to being in the era of 4.0. The purpose of this research is to find out the differences in taxpayer decisions to avoid taxes based on the whistleblowing system implemented by the Directorate General of Taxes (DJP) and big data analytics literacy. The test is conducting an experimental method with a 2x2 factorial design. Researchers experimented with student respondents who consisted of several groups with certain treatments. To test the main effect and interaction effect hypotheses between the variables studied, used the Two Way ANOVA analysis technique. The results showed that the main influence and interaction proved significant or not. Taxpayers' actions in tax avoidance can be minimized when the DJP whistleblowing system runs effectively. However, there is no big data analytics literacy effect on tax avoidance, and there is no interaction effect between factors. The uneven literacy of big data analytics among taxpayers in Indonesia will certainly change shortly. So, the DJP should make more optimal use of advances in science and information technology to increase its tax revenue target.


Introduction
Taxes are the largest source of state revenue (Darma et al., 2019;Febri & Sulistiyani, 2018;Hutapea & Herawaty, 2020;Siringoringo, 2017;Sunarsih et al., 2019). However, the efforts made to collect taxes from the public do not run easily because the public or taxpayers do not obtain the reciprocity directly so that the taxpayer does not feel that they have received anything from the tax payments made. This can cause taxpayers to avoid taxes. Furthermore, tax officials or the tax authorities who commit corruption are a serious trigger for taxpayers not fulfilling their tax obligations. Therefore, taxpayer compliance is still an important issue in Indonesia (Darma et al., 2019;Sulistyowati & Pahlevi, 2018).
Several factors have been studied related to things that can affect taxpayer compliance, one of which is the quality of tax authorities (Safitri & Silalahi, 2020). The state is obliged to provide the best service so that the public can participate in carrying out tax activities (Ariani & Biettant, 2019). Tax officers can be said to be qualified if they provide accurate information about taxation, including procedures for calculating, depositing, and reporting it, and do not commit criminal acts that violate applicable rules and Standard Operating Procedures (SOP) (Safitri & Silalahi, 2020). Optimal service will increase the satisfaction of taxpayers (Ariani & Biettant, 2019).
However, there are times when tax officials do not carry out their duties properly, even committing criminal acts, such as accepting bribes from taxpayers. As was the case in 2019, where there were tax officials who committed violations, four suspects were involved in the PT Wahana Auto Ekamarga (PT WAE) tax refund bribery case. They were given bribes by the commissioners of PT WAE so that the submission of PT WAE tax refunds in the 2015 fiscal year amounting to Rp5.03 billion and the 2016 fiscal year amounting to Rp2.7 billion was approved by them (CNN Indonesia, 2019).
Tools are needed to prevent similar cases from occurring in the future (Naufal et al., 2020). One way that can be done is by creating media that can report acts that violate the rules that can hinder the progress of national development and damage the country's integrity. The states that every organization wants honesty from and among its employees. The existence of honesty allows full dedication to carry out the mission in achieving organizational success. To that end, the Directorate General of Taxes (DGT) has provided a unique violation reporting system within the Directorate General of Taxes called the whistleblowing system.
Various fraud cases are revealed through whistleblowing carried out by internal parties in an organization (insider) (Dhamija, 2014). Employees who work in an organization are often the first to report fraud or immoral acts committed by others in their place of work (Dhamija, 2014). Whistleblowers (fraud reporters) can report cheating cases to someone, both internally and externally, whom they trust to solve the fraud (Dhamija, 2014;Lee & Xiao, 2018). By encouraging a whistle-blowing culture, organizations build transparent structures and effective and clear communication. What is even more important is that this whistleblowing action can protect organizational clients.
Whistleblowing system or WiSe is a medium in the form of an application provided by the DGT for everyone, both employees and taxpayers, who have information related to violations or misappropriations and corruption crimes committed by humans Resources (HR) in the DGT environment (Siringoringo, 2017). The DGT Whistleblowing System was implemented in 2012. With this reporting channel, it is expected that violations that occur within the DGT environment can be detected early and can form a corrective culture within the DGT environment and increase taxpayer compliance and scrutiny of the DGT so that the process maximum tax revenue can run (Siringoringo, 2017;Sulistyowati & Pahlevi, 2018). Besides, any reports received can be followed up according to procedures if the whistleblowing system is implemented effectively.
Indonesia has started to enter the era of the 4.0 industrial revolution, so to face this era, besides being needed old literacy, another literacy is also needed, namely new literacy. Literacy is the ability to read and write. Literacy is an initial ability that must be possessed by individuals to face life in the future. Therefore, literacy is very important to do. Old literacy includes competence in reading, writing, and arithmetic. Meanwhile, new literacy includes data literacy. The scope of data literacy is related to the ability to read, analyze, and make thinking conclusions based on data and information (big data) obtained (Fitriani, 2019). In this era of industrial revolution 4.0, of course, much data is scattered in the digital world. This abundance of data is called big data.
Data needs to be refined to realize its value. Data refinement can be done by turning the data into analytical materials so that it allows organizations to extract usable information from the resulting very large, diverse, and complex data sets. By performing big data analysis (BDA), an organization can identify the different types of data they produce. When the big data collected is put together to reveal a pattern and trend, the organization can expect higher efficiency and effectiveness levels related to the organization's performance (Kibe, 2018).
Big data has benefits in many fields. One is that the application of big data can be useful in forensic accounting education and practice as both academics and practitioners agree that "descriptive analysis," "understanding data," and "collecting or integrating data" are very useful in forensic accounting practice and show that they believe that techniques big data is important in advancing forensic accounting education and practice (Rezaee & Wang, 2019).
Besides, at the Technical University of Kenya, there has been a digital revolution regarding the development of new technologies such as universal computing devices, flexible classrooms, and online courses that are open en masse. This helps to provide more space for learning and provides organizations with an understanding of big data analytics regarding online learning methods (Kibe, 2018). BDA also has a significant positive relationship with innovating supply chain services and supply chain service performance (Fernando et al., 2018). Besides, BDA has benefits for e-commerce companies. This BDA can further provide value for companies by using the dynamics of people, processes, and technology to turn data into insights for effective decision-making and solutions to business problems (Akter & Wamba, 2016).
BDA can be useful in determining hidden patterns. If it is a taxpayer, the DGT will see data related to the Taxpayer, such as data on the stage, where the Taxpayer works, who is paying him, who has a business relationship with him, so that the DGT can see the transactions carried out by the Taxpayer, either Individuals and Bodies. Therefore, if the Taxpayer recommends tax evasion or commits errors in taxation regulations, the DGT can identify the patterns for taking these actions.
Based on the introduction above, the formulation of the problem in this study is how the difference in decisions to do tax avoidance based on the DGT whistleblowing system and big data analytics literacy partially, simultaneously (interactions) and the effect of these interactions. Ajzen (1991) states that the Theory of Planned Behavior (TPB) is a theory developed from the Theory of Reasoned Action (TRA) or Reasoned Action Theory. As in the concept in the Theory of Reasoned Action, this theory's main factor is a person's intention to behave. This intention is assumed to describe the factors that can influence a person's behavior. The stronger someone intends to do something, the more likely the intention will be realized in the form of behavior (Ajzen, 1991).  Ajzen, 1991) According to Ajzen (1991), the TPB concept can show the factors influencing a person to do something. These factors are summarized in three beliefs, namely as follows:

Theory of Planned Behavior (TPB)
1. Behavioral Beliefs and Attitude towards Behaviors 2. Normative Beliefs and Subjective Norms 3. Control Beliefs and Perceived Behavioral Control

Whistleblowing system of the directorate general of taxes
National Committee on Governance Policy (KNKG) (2008), through a report entitled "Whistleblowing System Guidelines (WBS)," states that whistleblowing is the disclosure of violations or acts against the law, unethical acts. Alternatively, immoral or other actions that can harm the organization or stakeholders are committed by employees or leaders of the organization to the leaders of other organizations or institutions that are believed to follow up on reports of violations.
To prevent violations occurring within the DGT environment, in 2012, DGT provided a whistleblowing system. Whistleblowing system or WiSe is a medium in the form of an application provided by the DGT for everyone, both employees and taxpayers, who have information related to violations or misappropriations as well as acts of corruption committed by Human Resources (HR) in the DGT environment (Siringoringo, 2017).
On August 19, 2011, DGT issued the Director-General of Taxes Regulation Number PER-22 / PJ / 2011 concerning the Obligation to Report Violations and Handling Whistleblowing in the Directorate General of Taxes (Perdirjen Number PER22 / PJ / 2011) (Siringoringo, 2017;Sulistyowati & Pahlevi, 2018). However, this regulation is no longer in effect. It has been replaced by the Directorate General of Taxes Regulation Number PER-07 / PJ / 2019 concerning Procedures for Submit-ting Complaints on Taxation Services at the Directorate General of Taxes and comes into effect since the date of the stipulation of the regulation, namely on April 4, 2019 (DJP, 2019).
As for the Regulation of the Directorate General of Taxes Number PER-07 / PJ / 2019 concerning Procedures for Submitting Tax Service Complaints of the Directorate General of Taxation Article 1 Paragraph (1) it is stated that "Taxation Service Complaints after this referred to as Com-plaints, are information submitted by the Reporting Party regarding alleged services. Taxation that is not following the provisions of laws and regulations". Verma & Agrawal (2016) states that BDA refers to the process of collecting, organizing, analyzing large data sets to find different patterns and other useful information. BDA is a set of technologies and techniques that require new integrated forms to reveal the big hidden patterns of big data sets that differ from the usual, more complex, and at different very large scales. This analysis focuses on solving new problems or old problems in a better and more effective way.

Big data analytics literacy
BDA's main purpose is to help organizations make better decisions, predict future business trends, analyze transactions made, and update the form of data used (Verma & Agrawal, 2016).

Tax avoidance
According to Febri and Sulistiyani (2018) states that tax avoidance is generally carried out by taxpayers who aim to minimize the company's operating profit so that later the amount of Income Tax is reduced, and this will affect the amount of Taxable Income. Meanwhile, according to Lee et al., (2015), tax avoidance reflects companies' efforts to reduce their payable taxes, either with legal or illegal objectives or strategies. Febri and Sulistiyani (2018) state that tax avoidance is one-way taxpayers do avoid taxes le-gally or not violate taxation and take advantage of loopholes in taxation regulations. Tax avoidance is still a complicated issue because, on the one hand, this action is allowed, but on the other hand, it is undesirable and even has the potential to harm the state because state income is reduced.
According to Darmawansyah (2019), there are two types of tax avoidance in several countries: acceptable tax avoidance and unacceptable tax avoidance. If the transaction carried out does not have a good business purpose by engineering transactions, this action is unacceptable tax avoidance. However, if the transaction carried out has a good business objective without manipulating the transaction and without violating the applicable laws, this action is considered acceptable tax avoidance.
Because state income in Indonesia still relies on the taxation sector, and many Taxpayers do tax avoidance, Taxpayer compliance must also be improved. Therefore, the DGT also makes rules to improve Taxpayer compliance and so that Taxpayers do not evade taxes, especially through illegal means (Apandi, 2019).

Framework of thinking
The whistleblowing system provided by DGT can provide an opportunity for anyone to report violations committed by DGT employees. Implementing the whistleblowing system also aims to build public trust and change the DGT environment's culture from permissive to corrective. With this corrective culture's construction, DGT will not forgive or tolerate acts of violations that occur in their environment (Sulistyowati & Pahlevi, 2018).
By having the ability of BDA, when the Taxpayer evades tax or violates tax regulations, the DGT will immediately find out through the patterns generated from the BDA.
In this study, the DGT whistleblowing system as an external factor and big data analytics (BDAL) literacy as an internal factor in determining the Taxpayer's decision to do Tax Avoidance can interact with each other to form a Taxpayer's attitude, namely in the form of a decision to do tax avoidance. Also, to minimize tax avoidance, the tax must increase. Taxpayer compliance is also closely related to the Taxpayers' morals Feld & Frey (2002). The Taxpayer's morale in determining the decision can be determined through the Theory of Planned Behavior.
With this frame of mind, the form of this image below describing the research:

Hypothesis
The formulation of the hypothesis which is used as a temporary answer using ANAVA is as follows: 1. Ha1: Respondents who are faced with a good DGT whistleblowing system have a lower level of decision to do tax avoidance than respondents who are faced with bad DGT whistleblowing system conditions. 2. Ha2: Respondents who were given BDAL had a lower level of decision to do tax avoidance than respondents who were not given BDAL. 3. Ha3: There is an interaction effect between the DGT whistleblowing system and BDAL on tax avoidance.
Furthermore, if there is an interaction between the Directorate General of Taxation Whistling System (X1) and BDAL (X2) on Tax Avoidance (Y), it is necessary to carry out further tests. The followup test is also referred to as testing the simple effect hypothesis on each cell using the SPSS syntax. Based on a 2x2 factorial design, this study will test 4 (four) simple effect hypotheses, namely: 1. Ha4: Respondents who are faced with a good condition of the DGT whistleblowing system in the condition of obtaining BDAL have a lower level of decision to do tax avoidance compared to respondents who are faced with DGT's bad whistleblowing system. 2. Ha5: Respondents who are faced with the condition of the DGT whistleblowing system that is good in the condition that they do not get BDAL have a lower level of decision to do tax avoidance compared to respondents who are faced with bad conditions of the DGT whistleblowing system.

Preparation of termite nest sample
Actinomycetes isolate growth A sample of termite nests was collected in September 2016 from Pananjung Pangandaran Nature Reserve, West Java, Indonesia. Termite nest was obtained by cruise method. The type of collected nest termite was carton nest. Termite nest samples (200 g per sample) were placed in polyethylene bags and immediately transported to the Microbiology Laboratorium of Research Center for Biomaterials-LIPI. Termites samples (worker and soldier) inhabiting the nest were collected and preserved in a 70% alcohol tube. The termites were identified based on key identification. Photographs were taken with a digital microscope with 40 -80 × magnification. Nest samples were ground into fine particles and air-dried at an ambient temperature for 7 days before the isolation of Actinomycetes.

Research strategy
This research is an experimental study with a 2x2 factorial design, in which there are 2 (two) factors, namely the whistleblowing system of the DGT and BDAL, which consists of 2 (two) levels. This 2x2 factorial design model has also been used in the research of Cahyonowati et al. (2020) and Wahl et al. (2010). If the manipulations are factorial combined, then this experimental research will have 4 (four) cells represented in the form of a matrix and the following notation:

Measurement
In this study, there are differences in the treatment of groups that are positioned in the good whistleblowing system of the DGT and those positioned in the flawed whistleblowing system of the DGT. In terms of measuring taxpayer decisions in tax avoidance, this study uses a questionnaire as one of the research instruments. The questionnaire is a data collection method in which the form of a questionnaire can be in the form of several written questions, and the aim is to obtain information from respondents about what they experienced and know (Siyoto and Sodik, 2015, p. 79). Measurement of the questionnaire scale uses a Likert scale. The questionnaire is given to respondents via an online form, where respondents must answer the percentage of income to be reported on a scale of 0-100 as follows: 1. 0-20%: Very Low 2. 21-40%: Low 3. 41-60%: Quite High 4. 61-80%: High 5. 81-100%: Very high Population and sample The population in this study was all students at the Indonesia University of Education. Furthermore, the sample used was 82 students (after deducting the manipulation check). The reason for choosing students as subjects to be studied is because students are expected to become tax-compliant taxpayers. Therefore, big data analytics literacy (BDAL) knowledge is needed and to increase their trust in the DGT because the DGT whistleblowing system is effective and is expected to encourage students as prospective taxpayers to obey to pay taxes according to the nominal charged so that the taxation sector revenue in Indonesia can run optimally. Also, given the increasingly rapid development of technology and the importance of big data analytics literacy, and the absence of general basic courses related to big data analytics, which are applied to all majors in the Indonesia University of Education.

Sampling method
This study's sampling technique is probability sampling, namely by using a simple random sampling technique where all students at the Indonesia University of Education have the same opportunity to be sampled.

Research object
The objects in this research are the factors that influence the taxpayer's decision to do tax avoidance. The factors in question are the Whistleblowing System (X1) and BDAL (X2), where both of these factors are independent variables that get manipulation or treatment and their effect on the dependent variable Tax Avoidance (Y) on the response to the manipulation.

Data collection
The data collection method in this research is quantitative with a quantitative experiment type. According to (Siyoto & Sodik, 2015), the experimental method is a research method that aims to explain the cause-and-effect relationship (causality) between one variable and another (independent variable or X with the dependent variable or Y) by controlling and measuring very carefully on the research variables.
In this study, the data collection technique is that respondents are given an online link to the questionnaire according to their respective groups. The questionnaire was given by sending a message personally, and we asked the respondent to fill in the questions in the link according to the conditions given 1 st ICEMAC 2020 34 through the scenario. In this scenario, the respondent is asked to act as an individual taxpayer who will report his income tax to the DGT. The questions in the scenario consist of: 1. Manipulation check Respondents were asked to answer manipulation check questions after being asked to read and understand the obtained experimental scenarios.

Percentage of Income Tax reported by Taxpayers
Respondents are asked to answer questions in the form of the amount of Income Tax (in percentage form), which will be reported to the DGT.

Research instrument
The research instrument in this study was a questionnaire. There were four questionnaires distributed to each experimental group, and each questionnaire had its scenario. The scenario used in this study was adapted from the research of Cahyonowati et al. (2020) and Wahl et al. (2010) adjusted for the variables studied. In each group, each respondent acts as a taxpayer. The scenario in question is as follows: 1. GWBS_BDAL scenario: There is follow-up information on reports of misconduct, and the perpetrator is punished under statutory regulations as well as BDAL treatment in which DGT knows the amount of salary, which gives the salary, and when the salary is transferred to the respondent's account. After that, the respondent wrote down the amount of tax reported as a percentage. 2. GWBS_NBDAL scenario: There is follow-up information on fraud committed reports, and the perpetrator is punished under statutory regulations. After that, the respondent wrote down the amount of tax reported as a percentage. 3. BWBS_BDAL Scenario: There is information that there is no follow-up on reports of misconduct and the DGT whistleblowing system is only a formality, and there is a BDAL treatment in which DGT knows the amount of salary, who gave the salary, and when the salary is transferred to the respondent's account. After that, the respondent wrote down the amount of tax reported as a percentage. 4. BWBS_NBDAL Scenario: There is information that there is no follow-up on the report of fraud that was carried out, and the DGT whistleblowing system is only a formality. After that, the respondent wrote down the amount of tax reported as a percentage.

Data analysis 1. Normality Test
The data normality test was performed as a prerequisite for performing the Analysis of Variance (ANOVA) test. In this study, the data normality test used Kolmogorov-Smirnov with the Monte Carlo technique in the SPSS 24 application. The test criteria and data normality hypothesis were as follows: Ho: α > 0.05, meaning that the population's decision to do tax avoidance by Taxpayers is normally distributed Ha: α < 0.05, meaning that the population's decision to do tax avoidance by taxpayers is not normally distributed.

2.Variance Homogeneity Test
The variance homogeneity test in this study can be carried out simultaneously with the ANOVA hypothesis test using the help of the SPSS 24 application. The test criteria and the variant homogeneity hypothesis are as follows: Ho: α> 0.05, meaning that all data groups have a homogeneous variant Ha: α <0.05, meaning that all data groups have heterogeneous variants.

Respondents profile
The subjects of this study were students at the Indonesia University of Education who could be categorized as follows: The number of 82 participants was divided into four cells that received different treatments.
ANOVA assumption test Before testing the hypothesis, two prerequisites must be fulfilled, namely normal and homoge-neous data.

Normality Test
Based on the data obtained, the data normality test results were obtained are shown in Table 6. The number of data in this study is 82, where this number indicates > 50. Therefore, the normal-ity test used is the Kolmogorov-Smirnov with the Monte Carlo technique, where the Sig. 0.122 > 0.05, meaning that the data is normally distributed.

Homogeneity Test
Based on the data obtained, the results of data homogeneity testing are as follows:

Hypothesis testing
The output in hypothesis testing in research that compares between cells or treatments is in the form of a Test of Between-Subjects Effects. This output provides ANAVA results regarding the presence or 1 st ICEMAC 2020 37 absence of the main effect between groups, namely the DGT whistleblowing system and BDAL, and the interaction effect between the DGT whistleblowing system and BDAL.  Table 8 shows that the value of F-count = 5.822 > 3.96 with a significance of 0.018 < 0.05 so that Ha1 is accepted, where the taxpayer is positioned in the whistleblowing system of the good DGT whistleblowing system has a lower level of tax avoidance decisions with taxpayers who are positioned in the flawed whistleblowing system of the DGT.
This result means that the better the DGT whistleblowing system's condition, the lower the taxpayers' intention to do tax evasion because the taxpayers know that the DGT is seriously following up cases of fraud that occur in their environment. In another sense, the whistleblowing system was not created for formality. This hypothesis test results indicate that the DGT whistleblowing system that is implemented effectively can increase taxpayer confidence in the DGT in managing reported taxes. That way, taxpayers can be motivated to pay taxes according to the nominal charged to them. This result is in line with research conducted by Siringoringo (2017) and Sulistyowati & Pahlevi (2018), which stated that the whistleblowing system is useful for detecting abusive actions that occur within the DGT environment. Therefore, a corrective culture will be formed and increase the Taxpayers' compliance and trust in the DGT. This is also in line with the statements stated by Carver (2010) that whistleblowing also helps create a transparent culture and effective and clear communication. This result will maximize the taxation sector revenue process. In line with the research conducted by Feld & Frey (2002) that the Taxpayers' morale also takes part in making decisions to comply with taxes. The results of testing hypothesis 1 are also in line with 2 (two) beliefs in TPB, namely Behavioral Belief and Attitude

Towards behavior and control beliefs and perceived behavioral control
The intention of taxpayers to avoid taxes based on BDAL treatment (hypothesis 2) in Table 8 shows that the value of F-count = 3.591 < 3.96 with a significance of 0.062 > 0.05 so that Ha2 is rejected, meaning that the taxpayer is given BDAL has the same level of decision to do tax avoidance as taxpayers who are not given BDAL. The results of testing hypothesis 2 in this study indicate that BDAL cannot prove its usefulness in the field of taxation. For example, the DGT will find out data related to taxpayers, such as income data, where they work, who pays them, whoever has a business relationship. With it, the DGT can determine the transactions carried out by Taxpayers, either by individuals or entities. Also, the results of testing hypothesis 2 cannot support the results of research by Mukherjee & Shaw (2016) which states that the combination of data and BDA can combat fraud so that this can be a factor in the absence of differences in the influence of BDAL. This study also cannot prove the statement stated by Mikalef 1 st ICEMAC 2020 38 et al. (2020) that BDA is more than an investment in technology, a very large collection of data, and later BDA can provide value to the organization. This hypothesis 2 result can be caused by the lack of technological literacy, especially big data analytics, in the Indonesian people (Sulisworo & Suryani, 2014), so that there are still people who do not understand the basics of big data and BDA.
The intention of Taxpayers to avoid taxes that receive a combination of treatment, namely the whistleblowing system and big data analytics literacy (WBS * BDAL) in Table 5 shows that the value of Fcount = 0.115 < 3.96 with a significance of 0.375 > 0.005, so that Ha3 is rejected, meaning there is no interaction effect between the DGT whistleblowing system with BDAL on tax avoidance. Based on the third hypothesis test results, there is no need for further ANAVA tests to determine the difference in the dependent variable's average score between the two groups of data or samples.

Conclusion
Based on the results of experimental research and tax avoidance based on the DGT whistle-blowing system and BDAL using 82 student respondents, it can be concluded that tax avoidance can be minimized if DGT maximizes the whistleblowing system that has been built by selecting reports, which is included in the criteria for fraudulent actions and following up. This can make the taxpayer feel worthless in reporting the tax owed because he believes in the DGT that the reported tax has not been misused so that the tax revenue process can run optimally. Also, the TPB concept only applies to the DGT whistleblowing system, where the results of the hypothesis 1 test are in line with 2 (two) beliefs in the TPB concept, namely Behavioral Beliefs, and Attitude Towards Behavior and Control Beliefs and Perceived Behavioral Control.
The absence of differences in the effect of the treatment of whether or not BDAL is present can be caused by several factors, such as the presence of Taxpayers who do not understand the technology and its benefits, the low level of literacy in Indonesia so that this affects the first fac-tor, there are still not many Human Resources with sufficient competence to manage and analyze big data, and other factors such as the inadequate financial availability in the process of managing and analyzing big data.