An Empirical Research on The Success of E-filing Tax Online Reporting During COVID-19 Pandemic in Indonesia

. This research aims to determine how the e-filing performance by taxpayers to complete their tax obligation by themselves which is more widely used during COVID-19 Pandemic with the TAM method. This research was conducted using a quantitative method use questionnaire that distributes to the taxpayers that used e-filing which will be managed using integrated TAM theory e-filing and processed using SEM-PLS with purposive sampling technique. The sample data obtained through the online questionnaires were 164 taxpayers. The result shows an interesting result where perceived usefulness and perceived ease of use have no remarkable relationship with behavioural interest due the obligatory used of e-filing in Indonesia. This research indicated that behavioural interest influences actual use because taxpayers continue to use e-filing as a tool in reporting their taxes. This research also shows that the quality of the system and service have no remarkable effect on other indicators. These results show that during the COVID-19 pandemic, the e-Filing Reporting can be considered successful where the e-filing is a handy application with updated system quality, an assurance security system that protects personal data and a good service quality that results in the taxpayers satisfied with the application.


Introduction
Since 2020, the COVID-19 pandemic has occurred all over the world, including Indonesia.The pandemic has resulted in major changes in people's daily lives where all activities are carried out by involving the use of information systems (Hambali, 2020a).The Enactment of these activity restrictions has an impact on all business and economic activities.These major changes have an impact on various sectors such as education, services, the economy, and food beverages sector.The changes that occur in these various sectors are no exception in the Taxation sector.One of the main sources of the Indonesia Revenue budget is Taxes.However, with the occurrence of this pandemic condition, there was a decrease in the tax revenue.Additionally, the services from the Tax Service Office can only be done online to prevent the spread of COVID-19.This results in taxpayers being required to be able to report and pay taxes independently without coming to the Tax Service Office (Sunitha et al., 2020; Jannah, 2020).E-Government is a strategy in implementing information technology by the government which aims to provide transparent, convenient, and easy services and information to the government regarding all matters relating to the government (Napitupulu, 2017).With E-Government, it can provide a better service to the community and give flexibility when providing service and information so Public can access it anywhere with minimum effort (BASU, 2004).In the Tax Sector, the Government implements E-Government through the efiling platform which aims to make its taxpayers easy to * Corresponding author: levana.prawati@binus.educarry out their tax obligation.According to the Directorate General of Taxation's regulation PER-02/PJ/2019 on Electronic SPT Procedures, "e-filing is an application that the taxpayer must be used to carry their tax obligation".The main function of this application is to give flexibility to the taxpayer so they can carry out their tax obligation anywhere and anytime.To support government's efforts to stop the pandemic COVID-19, the e-filing platform is being used as one of the most used tax services to avoid physical contact between taxpayers and the tax authorities.(Millenia, Kristianti, & Prawati, 2022) Despite the usefulness of the application, it has several issues such as server down, lags when loads many data, and poor UI that makes taxpayers confuse (Sijabat, 2020;Aliah, 2020;Widyari et al., 2021).One of the obstacles faced is changing the perception of taxpayers in reporting SPT by prioritizing the electronic system and not going to Tax Service Office.Before the pandemic, chat and telephone services are met by taxpayers who need assistance in reporting the Annual Tax.Although they can use electronic reporting, people prefer to come directly to the tax service office and ask for assistance.Therefore, the goal of this research is to determinate how the e-filing performance in Indonesia during the pandemic occurs because more widely taxpayers use e-filing to avoid contact happened.In the use of e-filing, the most important thing is how easy and useful it is to report taxes in anywhere and anytime.To find out, the model in this research was developed using the Technology Acceptance Model approach or more known as TAM.TAM was used because it's one of the leading model to explaining user behavior towards technology and measuring user experience as well as questions that applied in the studys are still valid today.The Researchers then develops and extends this model to explain perceived usefulness and intention to use in terms of System and service quality.TAM is a research models that used to analyzing the successful of the technology implementation.According to Davis, "there are two main variables in TAM are Perceived Usefulness and Perceived Ease of Use that will determinate the acceptance uses of the technology" (Al-Mamary et al., 2015; Surendran, 2012).E-filing affects perceptions of convenience and usability and analyzes behavioral attitudes as a factor for taxpayer to use of the technology application (Andika & Yasa, 2020).It is supported by the results of previous research that used TAM method which shows, and measure perceptions of convenience and usefulness affect the willingness of taxpayers which will affects the actual use of e-filing (Chang et al., 2005

Literature Review
There are eight models and concepts in the subject of individual adoption of technology system, according to (V.Venkatesh et al., 2009) User involvement in system development has been seen as a factor in acceptance.As a result, many technology acceptance theories and models have been established in the field to evaluate the use of specific produced and deployed technology (Taherdoost, 2018).The COVID-19 pandemic has made the government undergo a digital transformation in the distribution of its tax services, such as e-Filing.The online tax reporting system, often known as Efiling, exists since the conventional tax reporting system is inconvenient and inefficient.As result, the system needs to be more efficient and effective to facilitating manual and online interaction in the tax reporting service system while also assisting in the prevention of Covid-19 (Rahman, 2015).The e-Filing system allows taxpayers to plan, examine, and pay their taxes online.As a result, if the taxpayer files the SPT electronically, the risk of late submission is reduced (Hussein, 2011) (Bandiyono, 2020).According to Islam et al. (2012), the implementation of an e-Filing system positively affects taxpayers in a variety of ways.The Study also offers the potential benefit of increasing administrative commitment to service delivery efficiency and quality.Furthermore, they stated in another study that e-filing is simple, time saving in tax reporting, and ensures the privacy of taxpayer information, influencing taxpayer attitudes and behavior in the usage of e-filing (Dewi, 2018).The Technology Acceptance Model, D&M IS Success Model, Unified Theory of Acceptance and Use of Technology, and other models and frameworks have been developed to explain the factors that influence Efiling users' acceptance.The Technology Acceptance Model theory is used in this study to examine the supporting elements that influence E-filing acceptance.Davis (1986) introduced the Technology Acceptance Model Theory, which has six constructs: extended variables, perceived ease of use, perceived usefulness, attitude toward using system, behavioral intention to use, and actual system use.The key theory used to characterize the use of e-filing in this study is TAM (Al-Mamary et al., 2015) (Surendran, 2012).Several studies that look at specific technology acceptance behavior in various information system constructs, focusing on variables of perceived usefulness and ease of use system e-Filing that have extensively explored and validated by TAM.By integrating latent variables like service quality and system quality, this research have an objective for finding the more relevant view of acceptance e-Filing system in Indonesia.Previous research has never combined those two variables, perceived ease of use and usefulness.That's why, according to the study model developed it's to identify the successful factor of system e-Filing implementation in Jakarta (Figure 1), the system quality and service quality under this pandemic scenario affect perceived usefulness and ease of use.Both perceived will influence the behavioral intention to use e-filing, and behavioral intention affect the actual e-Filing system use.2.1 System Quality, Perceived Usefulness, and Perceived Ease of Use The system's quality is a variable that assesses the system's success from the perspective of the information system's technology.According to Davis, is described as a system's capacity to give information targeted to its users' demands.To analyze the quality of online tax reporting system information, variable model like service quality and system quality are used.In other research (Al-Mamary et al., 2015) also said that user of system information are more likely to be happy and use the system again if the information shown is accurate, relevant, simple to use, reliable, consistent, and offers the services needed.Several factors, such as ease of use, response speed, functionality, data security, and flexibility can be used to assess the quality of e-filing.When an e-filing system has a good quality system, the user will be more interested and more likely to use it.Several studies have discovered a strong relation beetween system quality, service quality and effect to the usefulness and ease of system use (Tam & Oliveira, 2016; Veeramootoo, 2018; Pramanita & Rasmini, 2020; Hambali, 2020a).As a result, the research hypothesis is formulated: H1.The quality of the e-filing system affects the perceived usefulness.H2.The quality of the e-filing system affects the perceived ease of use.

Service Quality, Perceived Usefulness, and Perceived Ease of Use
Service quality is useful for measuring the effectiveness of several technology systems, including e-government, according to Beheshti & Beheshti (2010).Service quality is a variable that assesses user satisfaction with the quality of services given by the helpdesk information system.Good service quality can increase the user of the system and satisfaction while meeting system performance and objectives (Shahzad et al., 2020;Weerakkody, 2013;Rana et al., 2015).Otherwise users will migrate to another system if the support system is inadequate.
As a result, this study updated the service quality variable using the TAM model, thus if the service quality given by tax authorities toward the use of e-filing improves, taxpayers will be happier and more satisfied with the system's use and expect to continue using it.Several studies have discovered that the quality of service has a beneficial effect on the used of such systems.Another study (Farizi, 2020; Masunga, Mapesa, & Nyalle, 2020; Zaidi, 2017) also discovered that the ease and use of information systems are influenced by service quality.Thus, the hypothesis are: H3.The effect of service quality in pandemic conditions on perceived usefulness .H4.The effect of service quality in pandemic conditions on perceived ease of use.

Perceived Usefulness, Perceived Ease of Use, and Behavioral Intention
The degree to which a person believes that information technology is simple to understand or simple to use is measured by perceived ease of use.According to Davis (1986), One of the most essential aspects in system adoption of a system is the users.To put it another way, the more users who find a system simple to use, the more likely they are to use it.Similarly, if the tax payers believes that e-Filing is simple to use, the level of behavioral interest in utilizing it rises.Several studies have also indicated that perceived ease of use influences system use intention (Novindra, 2017;Lie, 2013) It may be assumed, based on the findings of this study it is because the more user-friendly an e-Filing system is, the more likely people are to utilize it, and conversely.Therefore, the research hypothesis as follows.H5.Perceived ease of use affects perceived usefulness.H6.Perceived ease of use toward behavioral Intention to use e-Filing.

Perceived Usefulness, Behavioral Intention, and Actual Usage
According to the TAM model established by Davis (1986), Perceived usefulness have an important factor that will affect the user acceptance of technology system.The perceived usefulness of the system is related to the system's productivity and efficacy, as well as its advantages for better user performance (Surendran, 2012).As a result, the higher the taxpayer's behavioral intention in using e-Filing, the more likely the system used.Behavioral interest also influences actual system use where taxpayers are interest with the e-filing system determined by the perceived benefits and convenience.Taxpayers are comfortable with the electronic filing system because of the perceived convenience and benefits, according to behavioral interest.As a result, taxpayers will continue to use e-filing as a method of filing their taxes.Several research (Chang et al., 2005; Rakhmawati, 2020; Azmi & Bee, 2010; Lie, 2013) look at the impact of perceived benefits on e-Filing behavioral interest.The study's findings show that perceived usefulness influence people's willingness to try out the e-Filing system.If the taxpayer feels the usefulness, they will more likely to using the system, and vice versa.Therefore, the research hypothesis as follows.H7.Perceived usefulness affects behavioral intention to use e-filing.H8.Behavioral intention to use e-filing affects the actual use of e-filing.

Methods
The type used in this research is quantitative.The data collection technique uses online questionnaire and sampling technique used in the field research is purposive sampling where the sample is carefully selected so that it is relevant to the research program.In purposive sampling, the researcher determines the condition of the sample to be consistent with the research objectives.In this study, the population means (1) Respondents are private taxpayers who are registered and have tax ID number, (2) Respondents are personal taxpayers who use e-filing.Respondents that were sampled in this research are individual taxpayers in a total of 164 respondents processed.Data collection was obtained through a questionnaire whose questions included all research variables in TAM-2 which were measured using a Likert scale with 6 answer options, namely Strongly Agree (SS), Agree (S), Somewhat Agree (AS), Somewhat Disagree (ATS), Disagree (TS), and Strongly Disagree.(STS).The public who used the E-filing program during the COVID-19 outbreak were the subjects of this study.This research data was collected using a questionnaire which was then processed by Structural Equation Model Partial Least Square (SEM-PLS) using software SMART PLS-3.PLS is used to analyze data using the TAM model and aims to explain the effect of the relationship between exogenous and endogenous variables (Vinzi, 2010).
Processing of data using descriptive analysis was then carried out to test the validity, test reliability and test the research hypothesis using the partial least squares (PLS) method with an evaluation of the structural model (inner model) and measurement model (outer model).The authors used this method because the authors believe that the result of the analysis is more predictive and explains latent variables than testing a theory and the number of samples in the study is not large.

Numerical Results
Descriptive analysis was used to describe basic data in this research and help summarize the value of the outer loadings of the data.At this stage, the author analyzes respondents' answers, especially to questions on the profile of respondents using the e-filing system in the questionnaire (See Table 1 for Reference) to generate demographic information related to the characteristics of respondents on the success rate of implementing the e-filing system.The respondent data that was obtained by the researcher was 164 respondents.The results shows that the average value of the respondents is greater than the standard deviation it means that the data has low variability The demographic information includes the respondent's area, gender, age, last education, occupation, length of use of the system and level of acceptance of the e-filing system during this pandemic.

Variables Questionaire Mean Median Min Max StDev Loadings System Quality (SQ)
SQ1: E-FILING System updates the system regulary 4,793 5,000 2,000 6,000 0,894 0,919 SQ2: E-FILING System makes users not encounter difficulties or errors when accessing the system for SPT reporting 4,604 4,000 2,000 6,000 0,888 0,821 SQ3: E-FILING System responds to commands from users quickly and runs under normal conditions reporting 4,671 5,000 2,000 6,000 0,891 0,858 SQ4: E-FILING System is flexible so that it can be used by users for SPT reporting anywhere and anytime 4,994 5,000 2,000 6,000 0,953 0,934 SQ5: E-FILING provide information in accordance with the functions and needs of users, for the security of user data 4,811 5,000 2,000 6,000 0,934 0,909 SQ6: E-FILING maintain the confidentiality of taxpayer data 4,823 5,000 2,000 6,000 0,883 0,903 Service Quality (SEQ) SEQ1: It is simple to use the information in the E-FILING system to have effective communication and understand user needs.
4,628 5,000 3,000 6,000 1,025 0,892 SEQ5: Criticism and suggestion services in the e-filing system, can accommodate criticism and suggestions that come in from users, so that they can improve and improve system performance for the better.

Mean Median Min
Max StDev Loadings PU2: The E-FILING system provides a complete SPT filling module according to user needs, so as to increase the effectiveness of making SPT reports.4,805 5,000 2,000 6,000 0,930 0,951 PU3: Using E-FILING is beneficial in accelerating tax reporting.4,518 4,000 2,000 6,000 0,844 0,839 PU4: Using E-FILING makes it easier for me to control the process of making tax returns.4,665 5,000 3,000 6,000 0,878 0,840 PU5: The E-FILING system provides guidance on handling common errors, which can help users obtain information regarding SPT reporting issues 4,622 5,000 2,000 6,000 0,878 0,914 PU6: Using E-FILING as a taxpayer can increase my productivity at work.4,780 5,000 2,000 6,000 0,988 0,939 Perceived Ease of Use (PEoU) PEoU1: The use of the E-FILING system is flexible because it can be accessed anywhere and anytime.4,841 5,000 2,000 6,000 0,910 0,899 PEoU2: The E-FILING display is easy to read so that it is easy to understand it.4,555 4,000 2,000 6,000 0,864 0,856 PEoU3: The E-FILING system will send electronic receipts, so that users get timely information when the user has completed the SPT report.information results 4,744 5,000 2,000 6,000 0,895 0,911 PEoU4: The E-FILING system provides information on how to use the system, thereby helping users to learn about the system.

Proposed Improvements
TAM focuses more on the perceived usefulness of the user; therefore the authors see that there is a positive correlation between the quality of the system and what is perceived by the users of the system.By adopting these previous studies, the authors developed the integration of the TAM Model.So, in developing the model and hypothesis proposed here, the author replaces the external variables of the TAM Model with two new variables, namely, system quality and service, but does not replace the variables in the updated TAM model.

Validation
The validity test can be done by looking at the convergent validity and discriminant validity of the indicators.Convergent validity of the measurement model with reflexive indicators is assessed based on the correlation between the component scores calculated with PLS.The Data Indicator shows that the data convergent validity is greater than 0.7 it means the outer loading shows that all indicators have been able to measure the latent variables well and tend to have high validity (Table 1).Discriminant validity can be evaluated in two stages, namely looking at the value of cross loading between indicators and cross loading Fornell-Larcker.If the construct's correlation with the measurement item is greater than the size of the other constructs, it shows that the latent construct predicts the size of their block better than the size of the other blocks (Hair, Jr. et al., 2014).Based on the AVE Value, The Discriminant Validity value is greater than 0,5 which means the research variables can be considered Valid (Table 2).Examination of the Fornell-Larcker cross loading value was carried out by looking at the root value of the AVE.AVE root value must be higher than the correlation between the constructs with other constructs (Prawati, Setyawan, & Elsera, 2020) Based on the Fornell-Larcker Criterion, we can see that the correlation value of each indicator is greater than the other indicators.Therefore, we can determinate that the data has good discriminant validity (Table 3).In addition, Cronbach's Alpha Statistical test and Internal Consistency Reliability was used to calculate the reliability test in this study so we can see that through the test each indicator's value result is greater than 0,7 so we can say that the variable indicator is reliable (Table 2) (Fornell, 1981) Based on those result, the result shows us the data indicator have good convergent validity, good discriminant validity also considers to be reliable to proceed evaluate the structural model analysis and hypothesis result analysis.

Cronbach' rho
Composit Average SQ 0,949 0,968 0,959 0,795 SEQ 0,947 0,960 0,960 0,827 PEo 0,939 0,957 0,952 0,769 PU 0,942 0,946 0,954 0,774 BI 0,871 0,899 0,913 0,725 AU 0,857 0,870 0,898 0,639  Effect Size test (f2) was used to calculate the knowing the proportion of variance of exogenous variables to endogenous variables to predict the effect of the variable certain where the suggested effect size f2 value are 0,02 (small effects), 0,15 (moderate effects) and 0,35 (large effects).In Table 4 it can be seen the results of the test where from a total of 8 hypotheses there are 2 hypotheses with large value, namely the influence of PEoU on PU, and BI's influence on AU, then 2 hypotheses with medium value and 4 hypotheses with small value.(Chin, 1988) The results of f2 show that the effect of system quality on perceived ease of use, system quality to perceived usefulness, perceived ease of use to behavioral intention, perceived usefulness to behavioral intention is small (f2 < 0.15).The effect of service quality on perceived ease of use and service quality to perceived usefulness are moderate (f2 = 0.15 -0.35).Moreover, the effect of perceived ease of use on perceived usefulness and behavioral intention to actual ble 4).
relationship is among the independent and dependent variables.(Hair, Jr. et al., 2014) The value of path coefficients is getting closer to the value above 1 symbolizing the relationship between the two constructs is getting stronger and are declared to have a significant effect (Table 4).A value greater than zero indicates a positive correlation, while a value smaller than zero indicates a negative correlation.Therefore, it can be concluded that the correlation between variables in the construct is positive.Thus, there is a significant positive effect of BI on AU (H8).From those results, we can conclude that the relationship between behavioral intention with actual usage is the strongest one and with the relationship between system quality with the perceived of usefulness is the weakest.

Conclusion
E-Filing is one of the most important applications for taxpayers to carry out their tax obligation.Based on the result, we found out that out 5 of 8 hypotheses are all supported and 3 of 8 are not supported.Out of 8 hypotheses, there are 2 hypotheses that gave interesting results where the perceived usefulness and perceived ease of use didn't affect the behavioral intentions taxpayers to use e-filing.This is interesting because the result is different with others study that already conducted (Novindra, 2017) (Lie, 2013).From the result, it can be concluded that service quality and system quality has only significant effect to other indicators such as perceived usefulness and perceived ease of use.Other than that, it shows that the usefulness and the convenient use of the application have a positive effect on one another statistically.In accordance with Hypothesis 8, it also shows that behavioral interest has large positive effect on actual us.This can be happened due to the fact during the pandemic in Indonesia, the efiling is mandatory.
From this research, it hinted that Directorate General of Taxation, the creator and developer of e-filing service can raise the public awareness especially in terms of tax reporting compliance.This is Based on the result that we find where the aspect can be seen to able increase taxpayers' intention, so they continue to use e-Filing in annual SPT reporting not only during pandemic but after the pandemic as well.This due the e-Filing is required to carry out their duty.In addition, the system quality, security quality and service quality should be improved so the taxpayers comfortable and satisfied when used efiling.Not only that, the system and service must keep being improved to provide a handy application and convenient service.The improvement of the tax service can be done such as gives information that help the taxpayers when using the application.Other than that, if the taxpayers are encounter problems in reporting the annual SPT then it must be handled properly so it will not disturb the user convenience.
In this research, researchers aware there are some limitations, such the taxpayers' respondents are limited to the city of Jakarta, Bogor, Bekasi, and Tangerang and we just collect taxpayers that had use e-Filing.Due to this, the result may not represent Indonesia as whole so it cannot be generalized.With the limitation that exist within this study, we have several suggestions for future research such as used a larger sample not only from Jakarta, Bogor, Bekasi, and Tangerang and add a larger sample so it will provide a more accurate data.In addition, further research would be better if using different analytical techniques and developing research models and involving interviews of respondents and professionals in their fields to find out more about what factors influence the use of e-filing.

Fig. 1 .
Fig. 1.The Original TAM Model (the black lines) and the modification proposed in this study (the blue lines)

Table 2 .
Validity and Reliability results

Table 4 .
Hypotheses results can be seen in Table4that the T-statistic value is 5.927 (> 1.96) with an original sample estimate of 0.372.It means that there is a significant positive effect of PEoU on PU (H5).The influence of PEoU on BI gives a T-statistic value of 1.826 (< 1.96) with an original sample estimate of 0.240.Thus, there is no significant effect of PEoU on BI (H6).Next is the hypothesis testing between PU and BI.Table4gives a T-statistic value of 1.023 (<1.96).It means that there is no significant effect of PU on BI (H7).The last hypothesis is between BI and AU giving an original sample estimate of 0.538, the T-statistic value is 11.464 (>1.96).