The Effect of Tax Audit, Tax Amnesty, and Tax Sanctions on Taxpayer Compliance at Pratama Bima Tax Services Office

ABSTRACT


INTRODUCTION
Tax is a source of income for the government, which plays a significant role in the continuity of government. When we look at our country Indonesia, we can understand it through the conditions around us, such as education, health, public facilities, and others, because our state taxes are distributed to these sectors. Judging from the decline in state revenues from year to year, the problem with tax collection is that many people often want to avoid fulfilling their tax obligations. Namely, there are still many tax arrears. Individual taxpayer compliance (WPOP) in Indonesia is still relatively low. Taxation obligations to guarantee state revenue from the tax sector. This will later become one of the goals of the Director General of Taxes in conducting tax audits. A tax audit aims to verify compliance with tax obligations and other purposes related to implementing tax laws and regulations. Tax  Procedures. Then in 2006 issued, an additional order of directors was. In the context of SAS, a tax audit carried out professionally by the Fiskus is a form of tax law enforcement. A tax audit oversees taxpayers' implementation of the SAS system (1)(1). Without investigations, tax audits, and the power of the tax office, taxpayer negligence can grow to paralyze the tax system. Keeping taxpayers within the corridors of tax regulations is carried out by examining taxpayers who meet the enforcement criteria (2). Tax amnesty is a government policy given to taxpayers for tax amnesty, and in exchange for that amnesty, taxpayers must pay a ransom. The first reason Indonesia imposes a tax amnesty is that citizens, both at home and abroad, have property rights that are not or are not entirely reflected in the Annual Income Tax Return. Tax amnesty is an attempt by the government to drain people's money parked in banks in other countries. Tax amnesty will be given to those who have not paid properly," Deputy Minister of Finance Mardiasmo said on the sidelines of the XII. Congress of the Indonesian Institute of Accountants in Jakarta. Taxpayers can be influenced by many factors other than tax audits and amnesty applications. As we know, tax enforcement tax laws also affect taxpayer compliance. Some laws regulate general rules and procedures for taxation to ensure tax compliance and tax penalties for violations (3).
Taxpayers will comply with their tax obligations when they see that applying tax sanctions will be more detrimental to them (4). The enforcement and imposition of these sanctions are administrative sanctions/fines and criminal sanctions. The word sanction in the Indonesian dictionary means responsibility (action and punishment) to force people to fulfill agreements or comply with legal regulations. Tax sanctions are a guarantee that the provisions of tax laws and regulations (tax standards) are met/observed. In other words, tax sanctions are a tool (preventive) to prevent taxpayers from violating tax norms (5).
Bima is one of the areas in West Nusa Tenggara that is of concern regarding tax management. KPP Pratama Bima is an official institution with a tax service office for Bima. Bima regional government spending from taxes will be realized if the taxpayers obey paying taxes. However, it will be hampered if the taxpayer violates the tax regulations. The control of tax violations against taxpayers can be dealt with quickly by the Bima government, namely by fines, tax amnesty (tax amnesty), and so on. However, until now, the government is still experiencing difficulties with violations by taxpayers. This violation is detrimental to the regional government of Bima because tax revenues should be higher than they are, so the Bima government is less able to provide maximum public facilities. BPS Kota Bima data shows the number of taxpayers and tax realization in 2017 -2021.  Table 1 shows that from 2017 to 2021, there was a weakening increase in the number of taxpayers. Although there has been an increase, the growth has remained relatively high. The same thing can be seen from the realization of tax revenue which has stayed the same. This indicates that there are problems that must be identified and solutions sought. Tax plans and completion must be clear every year.
To increase tax revenue in Kota Bima, there needs to be an appeal or direction for taxpayers to pay and report their taxes.

Taxpayer
According to research by Fischer, CM, Wartick, M., & Mark, M. (1992), there are two basic tax compliance models, namely the financial self-interest model and an expanded model. The economic selfinterest model assumes that every taxpayer has a rational nature and tries to maximize the income earned by considering the possibility of being detected by the tax authorities (6). This theory identifies three variables affecting tax compliance: tax rates, probability of detection, and applicable tax penalties. An Expanded model (Jackson & Milliron, 1986) identifies 14 main variables that affect tax compliance. These variables are grouped into four sections, namely demographics (e.g., age and gender), probability of non-compliance (e.g., education and amount of income), attitudes (e.g., ethics and group influence), and structural (e.g., sanctions, audit probability, and tax rates). (7).
Taxpayer compliance is taxpayers' compliance in carrying out their tax obligations by applicable regulations. Generally, taxpayer compliance is measured by obedience in paying and reporting taxes, whether it has been done correctly by applicable regulations (8). Taxpayer compliance is promptly entering and reporting the required information, correctly filling in the amount of tax payable, and paying taxes on time without coercive action. Non-compliance arises when a defined condition is not met (9).

Tax audits
According to Priantara and Diaz (2012), a Tax audit is the obedience or compliance of the taxpayer in carrying out the self-assessment system, namely filling, calculating, calculating, collecting, deducting, and reporting all tax obligations according to the provisions of the applicable tax law (10). Oktaviani (2007) states that tax audits affect tax revenues. This means that a tax audit is an essential instrument for determining the level of taxpayer compliance, both formal and material, aiming to test and improve a taxpayer's tax compliance, where taxpayer compliance is a strategic position in increasing tax revenue (11).
A tax audit is carried out as a form of supervision and guidance by the Director General of Taxes so that the provisions of the tax law carry out tax collection. The expected result of conducting a tax audit is increasing taxpayer compliance in fulfilling its obligations in taxation. The taxpayer can use the findings of the audit results as a valuable experience and at the same time as learning so that in the next tax year, the taxpayer can immediately correct his mistakes and not repeat the same mistakes in carrying out his obligations in the field of taxation (12 ).

Tax amnesty
According to the Tax Amnesty Law Number 11 of 2016, "Tax Amnesty" is the elimination of tax payable through disclosure of assets and payment of ransom without being subject to tax administration sanctions or criminal sanctions in the field of taxation. as stipulated in this Law, tax amnesty is the exemption from tax payable through disclosure of assets and payment of redemption money, without being subject to tax administration sanctions or criminal sanctions in the field of taxation, according to PMK No. 118/PMK.03/2016. Tax Amnesty is defined as a Taxpayer disclosing his assets and paying a ransom for Tax Amnesty (13). Amnesty, or tax amnesty, is a government attempt to pay off taxes owed by paying a ransom to generate additional tax revenue and an opportunity for non-compliant taxpayers to comply. Applying the Tax Amnesty is expected to increase the voluntary compliance of taxpayers in the future (14).

Tax Sanctions
Sanctions are an additional burden for taxpayers who neglect to fulfill their tax obligations. Tax sanctions will be imposed if, based on the audit results, unpaid tax debt is found. In examining the level of tax compliance, Lars, P. and Bruno, S. (2002) used audit probability variables, imposition of fines, tax rates, and income levels as independent variables ( (15)15). This study found that the imposition of fines was positively correlated with the level of tax compliance. The results of this study are in line with research (16).

Research Approach
This study uses a quantitative approach (Quantitative Approach), which is an approach that emphasizes testing theories or concepts through variable measurements and data analysis procedures with statistical tools to test hypotheses. Based on its type, this research is classified as explanatory research (explanatory research). Explanatory research analyzes the relationships between a variable and other variables or how a variable affects other variables through hypothesis testing (17). Therefore, using this method provides a good understanding and reasoning for the results of this study as a whole.

Location and Time of Research
This research was conducted at the Bima Primary Tax Service Office, Jl. Soekarno Hatta No. 17 Raba-Bima 84115 NTB. The research time required is during January-March 2022

Types and Data Sources
The data used in this study, namely primary data. Primary data is a source of research data obtained directly from the source (not through intermediary media). Primary data is collected explicitly by researchers, namely data obtained directly from individual taxpayers registered at KPP Pratama Bima. This data is in the form of a questionnaire filled out by the taxpayers and the selected respondents.

Data collection technique
Researchers obtain the necessary data in the following ways: 1) Primary data tracking by doing documentation. Usually done to collect primary data from various sources. Both personally and institutionally. In this case, researchers used data collection techniques using documentation obtained from the Bima Pratama Tax Service Office. 2) Distributing questionnaires (questionnaires). Researchers' questions and respondents' answers can be written in writing through a questionnaire. This technique gives responsibility to the respondent to read and answer the questions.

Reliability and Validity Test
To test whether the constructs (variables that cannot be measured directly but are formed through observed dimensions or indicators) that have been formulated are reliable and valid, it is necessary to test reliability.
A reliability test is a test to measure a questionnaire which is an indicator of a variable or construct. A questionnaire is reliable if one's answers are consistent or stable over time. Reliability is the extent to which the results of a measurement can be trusted and can provide results that are relatively the same when repeated on the same subject. A construct or variable is said to be reliable if it gives an Alpha > 0.60 (18).
A validity test is used to measure the validity or validity of a questionnaire. To determine whether an item is valid, a comparison is made between the questionnaire r count and the coefficient r table. The item is valid if the r count is more significant than the r table. Conversely, if the r count is more minor than the r table, the item is invalid.

Classic assumption test
The normality test aims to test whether, in the regression model, there are confounding or residual variables that have a normal distribution. To detect the normality of the data, this study will perform a nonparametric statistical test Kolmogorov-SmirnovTest (KS), if the probability value is significant (19).
The multicollinearity test is used to determine whether there is multicollinearity by investigating the magnitude of the intercorrelation between the independent variables. Whether multicollinearity exists can be seen from the Tolerance Value and VarianceInflationFactor (VIF) size. If the Tolerance Value ≥ 0.10 or the same as the VIF value ≤ 10.
The heteroscedasticity test was carried out to test whether the residual inequality occurs in the regression model between one observation and another. Suppose the variance of the residuals from one observation to another is constant. In that case, it is called homoscedasticity, whereas if the variance of the residuals between one observations is different, it is called heteroscedasticity. To detect the presence or absence of heteroscedasticity, analysis is used with the Glesjese test with the regression equation as follows: Ut = α + βXt + vt If the independent variable is statistically significant, it does not affect the dependent variable. Then there is no indication of heteroscedasticity occurring. This can be seen if the significance probability is above the 5% confidence level.

Regression Models
The analysis used in this study is multiple regression analysis, namely the regression model, to analyze more than one independent variable. The regression equation formulated based on the hypothesis developed is as follows: Y=α + β 1 X 1 + β 2 X 2 + β 3 X 3 + € Information: Y = Taxpayer Compliance α =Constant β 1 = Tax Audit variable regression coefficient β 2 = TaxAmnesty variable regression coefficient β 3 = Regression coefficient of the variable perception of tax sanctions X 1 = Tax Audit X 2 = Tax Amnesty X 3 = Tax Sanctions B = Regression Coefficient e = Regression Error

Hypothesis test
Hypothesis testing is used to measure the strength of the relationship between two or more variables and show the direction of the relationship between the dependent and independent variables. The accuracy of the sample regression function in estimating the actual value statistically can be measured from the value of the coefficient of determination, the value of the F statistic, and the value of t. Statistical calculations are statistically significant if the statistical test values are in the critical area (H0 is rejected). Conversely, it is called insignificant if the statistical test value is in the area where H0 is accepted.
The coefficient of determination (R 2 ) measures how far the model can explain the dependent variable's variation. The coefficient of determination is between zero and one. The small R 2 value means that the ability of the independent variables to explain the variation in the dependent variable is minimal. A value that is close to one means that the independent variables provide almost all of the information that is formed to predict the dependent variable. The fundamental weakness of using the coefficient of determination is the bias toward the number of independent variables included in the model. For each additional independent variable, R 2 must increase regardless of whether the variable significantly affects the dependent variable. Therefore, many researchers recommend using the adjusted R2 value when evaluating the best regression model (18).
Simultaneous Parameter Significance Test (Statistical F Test), after obtaining the estimated regression equation model, the next step is to test the significance of the event regression coefficient together (Testing the Overall Significance of Regression The t-statistical test shows how far the influence of one explanatory or independent variable individually explains the variation of the dependent variable. The t-test is used to partially test the effect of each independent variable used in this study. In the t-test, arithmetic is more significant than the t table, then Ha is accepted, and Ho is rejected. However, if the calculated t value is less than the t table, Ha is rejected, and Ho is accepted.

Reliability Test
A reliability test is conducted to show the extent to which measurement tools can be trusted. Generally, a variable is reliable if it gives a Cornbrash alpha value of > 0.6, so the research questionnaire is declared reliable. The results of the reliability test can be seen in Table 2. Where the Cornbrash alpha values are 0.920 (x1), 0.908 (x2), 0.916 (x3), and 0.916 (y), this indicates that the Cornbrash alpha coefficient is more significant than 0.6. Thus the measurement items on the variables are declared reliable and can be used in research. The reliability test results based on Table 2 show that the variable has a Cornbrash alpha X1 coefficient of 0.920, X2 of 0.908, X3 of 0.867, and Y of 0.916, this indicates that the Cornbrash alpha coefficient is more significant than 0.6. Thus the measurement items on the variables are declared reliable and can be used in this study.

Validity test
This test was carried out using the corrected item-total correlation, namely by correlating the score of each item with the total score, with several respondents 92 and a significance level of 5%. The questionnaire questions were declared valid if rcount>rtable and vice versa. From this study, it is known that the table is equal to. The results of testing the validity of the dependent and independent variables from the 99 sample respondents can be seen in Table 3.

Classic assumption test Normality test
The data normality test is conducted to see whether data is normally distributed. Data normality test was performed using standardized histogram residual with PP plosstandardized residual. The normality test aims to test the dependent and independent variables, namely tax audits (X1), tax amnesty (X2), and tax sanctions (X3) on taxpayer compliance (Y) both of which have a normal distribution or not. The normality test can be seen in Figure 2 below.
In the standard graph plot of Figure 1, it can be seen that the dots spread around the diagonal line, and the distribution follows the direction of the diagonal line. Based on this, the data is usually distributed.

Multicollinearity test
The multicollinearity test aims to test whether the regression model finds a correlation between the independent variables. If the independent variables are correlated with each other, then the variables are not orthogonal. A good regression model is free from multicollinearity. The cutoff value commonly used to indicate the presence of multicollinearity is a tolerance value of > 0.10 or the same as a VIF value < 10. Based on data processing results, variance inhibition shows that the VIF value is < 10, so it is categorized as multicollinearity-free, meaning that the dependent variables are orthogonal.

Heteroscedasticity Test
The heteroscedasticity test aims to test whether, in the regression model, there is an inequality of variance from one residual observation to another. If the variance from the residual of one observation to other remains, it is called homoscedasticity; if it is different, it is called heteroscedasticity. A good regression model has homoscedasticity or does not have heteroscedasticity. Based on the results of data processing, the results of the scatterplot are seen in Figure 2. From the scatterplot line in Figure 3 above, it can be seen that the points spread randomly and are spread both above and below zero on the Y-axis. It can be concluded that there is no heteroscedasticity in the l regression model.

Hypothesis test Multiple Linear Regression Analysis
After testing the validity and reliability of the data, it can be concluded that the model can be used to test multiple linear regression analysis. Based on the results of the research, data processing that has been done is shown in the attachment (coefficients). Based on the results of multiple linear regression tests that have been carried out using SPSS, the results are as follows: Y= 1.042 + 0.332 X1 + 0.322 X2 + 0.140 X3 Based on the results of multiple linear regression tests, the following data are obtained: The tax audit variable (x1) gives a regression coefficient value of 0.332, meaning that if there is an increase in the tax audit by one unit while the value of other variables remains the same, this will increase personal taxpayer compliance by 0.332. So the higher the level of a tax audit, the higher the level of taxpayer compliance.
Tax amnesty variable (x2) gives a regression coefficient value of 0.322 meaning that if there is an increase in the tax amnesty of one unit while the value of other variables remains the same, it will increase individual taxpayer compliance by 0.322. So the higher the level of tax amnesty policy, the more taxpayer compliance will also increase.
The tax sanction variable (x3) gives a regression coefficient value of 0.140, meaning that if there is an increase in the tax sanction variable by one unit while the value of other variables remains the same, it will increase individual taxpayer compliance by 0.140. So the higher the level of tax sanctions, the higher the taxpayer compliance.

Test the coefficient of determination (R 2 )
Based on the multiple linear regression table (Model Summary b ), it is obtained that the coefficient of determination R 2 is 0.484 or 48.4% indicating that three independent variables (tax audit, tax amnesty, and tax sanctions) simultaneously affect the dependent variable (individual taxpayer compliance), with other variables influence a variation of the effect of 48.4% and the rest.

Simultaneous parameter significant test (Statistical test F)
Based on calculations using the F test, the calculated F value is 27.540 with a significant level of 0.000 a because the probability of significance is much smaller than 0.05, so the regression model can be used to predict the variable taxpayer compliance (Y) or it can be said that tax audit (X1), tax amnesty (X2), tax sanctions (X3), have a significant effect on individual taxpayer compliance (Y).

Test the significance of individual parameters.
The t-test shows how far the influence of one independent variable individually can play a role in the variation of the dependent variable. Based on the results of research examining the evaluation of tax audits, tax amnesty, and tax sanctions against taxpayer compliance at the Bima Pratama Tax Office described above, several things can be explained in this study, namely:

Tax audit has a positive effect on taxpayer compliance.
Based on the study's results, it was found that variable X1 has a positive effect on variable Y. Tax audits have a relatively strong and positive influence, meaning that the higher the tax audit, the higher the predictability of taxpayer compliance. From the data obtained above, the implementation of tax audits positively influences individual taxpayer compliance.
The data obtained on the frequency of respondents' answers to the tax audit variable shows a positive answer. Most respondents' answers are in the fourth range, namely high. However, the first indicator occupies the position of the highest frequency, with the sound of the question " Tax audits must first look at the SPT and the taxpayer's financial statements" is the indicator with the highest frequency, which is in the fourth range, namely high, meaning most of the respondents agreed that the inspection was carried out by checking the SPT and financial reports first. This is in line with the formal tax law that has been regulated in Article 29, namely, "the director general of taxes has the authority to carry out audits to test compliance with the fulfillment of taxpayers' tax obligations and for other purposes in order to implement the provisions of tax laws and regulations." Article 29 states that the audit is carried out to test the fulfillment of the taxpayer's tax obligations by tracing the truth of SPT, bookkeeping or recording, and fulfillment of other tax obligations compared to the actual conditions or business activities of the taxpayer. Under Article 29, an audit is the right of the Director General of Taxes, which is legally guaranteed. The indicator that is at the lowest frequency is the second indicator, which reads "Before carrying out an inspection, first study the files and data of the taxpayer", but it remains the same, which is in the fourth range. Before carrying out a tax audit, the tax authorities carry out activities to collect taxpayer files and data files, starting with borrowing files from related sections and utilizing internal data contained in the administrative system of the tax office concerned. After the data is collected, it is studied and confirmed on the Tax Payment Slip (SSP), proof of withholding PPH article 23, and input tax invoices.
Based on the research results and the literature review on which this research is based, the implementation of tax audits can positively contribute to increasing taxpayer compliance and has implications for optimal tax revenues. Audit findings can be used by taxpayers as valuable experience and at the same time, learning so that in the following tax years, taxpayers can immediately correct their mistakes.
A tax audit is also a means to supervise and guide taxpayers. Besides aiming to test the level of compliance of taxpayers in fulfilling their tax obligations, it also has other objectives to implement the provisions of the tax law (20 ). Tax audits are conducted not to find fault with taxpayers but to test compliance with tax obligations. This is consistent with research conducted by (Dwi Rahayu, 2015). Auditing affects compliance with formal tax obligations for PPh Article 21 deposits, and audits affect material taxpayer compliance (21). Hendro Saputro (2011) also found a significant relationship between tax audits and increased taxpayer compliance (22).

Tax Amnesty has a positive effect on taxpayer compliance.
Based on the results of the above study, it was found that variable X2 had a positive effect on Y, whereas tax amnesty (X2) had a positive effect on taxpayer compliance (Y). The positive effect shown by the tax amnesty indicates that the better the implementation of the tax amnesty, the better the taxpayer compliance will be.
The data obtained on the frequency of respondents' answers to the tax amnesty variable shows a positive answer. Where the first indicator occupies the position of the highest frequency with the sound of the question "Tax Amnesty is a policy that has benefited you directly financially, due to the abolition of administrative sanctions and criminal sanctions," which is in the fifth range, which is very high, meaning that almost all respondents agree that taxpayers participate in tax amnesty get a policy of eliminating administrative sanctions and fines. One of the advantages obtained by taxpayers who participate in tax amnesty is the abolition of administrative sanctions and criminal sanctions owned by the taxpayer. This is by the tax amnesty decision letter in Article 8, paragraph 3 letter A, which reads the elimination of tax payable, tax administration sanctions, and criminal sanctions in the field of taxation for tax obligations before this law is enacted, which has not yet been issued a tax assessment letter. The indicator at the lowest frequency is the third indicator with the sound of the question "Tax Amnesty reduces your trust in the government in the field of taxation," which is in the fourth range, namely high. This means most respondents do not fully trust the government in taxation matters. Related to this, one of the objectives of tax amnesty is to restore confidence in the government and increase our tax base in the future.
Tax Amnesty is an opportunity for a limited time for certain groups of tax payments to pay a certain amount within a specific time in the form of amnesty of tax obligations (including interest and fines) relating to the previous tax period or a certain period without fear of criminal penalties. The severity of tax evasion. As a result of this policy, benefits can be obtained through funds, especially the return of funds stored abroad. However, this policy could be more robust in the long term because it can have dire consequences in the form of a decrease in voluntary compliance from obedient taxpayers. This is also said by Ngadiman and Daniel Huslin (2015) in their research that tax amnesty y has a positive and significant effect on taxpayer compliance (23).

Tax Sanctions have a positive effect on taxpayer compliance.
Based on the research results above, it was found that variable X3 hurt Y, where tax sanctions (X3) had a positive effect on taxpayer compliance (Y). This means that the taxpayer's perception of tax sanctions has a positive and significant effect on individual taxpayer compliance at the Bima Pratama tax service office. This means that if the level of taxpayer perception of tax sanctions is good, it will tend to increase taxpayer compliance and vice versa. This is also consistent with the proposed hypothesis.
The data obtained on the frequency of respondents' answers to the tax sanctions variable shows a positive answer. Most of the respondents' answers were in the fourth range, which is high. However, the fifth indicator is the highest frequency with the question, "WP will be subject to administrative sanctions if they do not pay/underpay the tax owed when it is due." This means that the majority of respondents agree that if the taxpayer does not pay/underpays the tax payable when due, he will be subject to administrative sanctions. Administrative sanctions are usually only in the form of fines. In general provisions and tax procedures, these administrative sanctions are divided into 3 points: interest, fines and increases. The indicator at the lowest frequency is the first indicator and is in the fourth range with the sound of the question, "Taxpayers (WP) will be sanctioned if they are late or not on time in fulfilling their tax obligations." This means that most respondents agree that if the taxpayer is late or not on time in fulfilling his obligations, he will be subject to sanctions. Individual and corporate taxpayers late in depositing/paying taxes by the provisions will be subject to an administrative penalty of 2% per month of unpaid taxes. The 2% penalty is calculated from the due date until the tax payment date, where part of the month is calculated as 1 (one) entire month.
Tax sanctions are made with the aim that taxpayers are afraid to violate tax laws. For these laws and regulations to be complied with, there must be sanctions for violators. Mardiasmo argues that tax sanctions guarantee that the provisions of tax laws and regulations (tax norms) will be obeyed, obeyed, and complied with. In other words, tax sanctions are a deterrent tool so that taxpayers do not violate tax norms (24). Tax sanctions are one of the factors that can affect taxpayer compliance. To increase compliance, it is necessary to measure tax sanctions through administrative sanctions in the form of (1) Fines for not reporting tax returns payable, (2) Fines for late reporting of SPT, and (3) Fines for having tax arrears.
The results of this study are consistent with research conducted by Sri Putri Tita Mutia (2008), namely that tax sanctions have a significant and positive effect on taxpayer compliance. Likewise to Ngadiman and Daniel Huslin (2015), tax sanctions positively and significantly affect taxpayer compliance (23).

CONCLUSION
This study concludes that tax audit partially positively and significantly affects individual taxpayer compliance. This means that the higher the tax audit level, the higher the predictability of taxpayer compliance.
Tax amnesty partially has a positive and significant effect on individual taxpayer compliance. This is due to the relief of tax fines, thereby increasing tax compliance. The tax amnesty also shows a positive effect, which identifies that if the tax amnesty increases, the taxpayer compliance rate will also increase.
Tax sanctions partially have a positive and significant effect on individual taxpayer compliance. This means that the better the application of sanctions, the level of taxpayer compliance will also be good.