Antecedents of e-government perceived net benefits: a study of e-filing in Indonesia

ABSTRACT

being built. Past research has revealed that trust in e-government improves perceived system quality [50], [51]. As a result, when citizens have more trust in e-government websites, they will have a better perception of system quality. The level of trust in e-government websites also has an effect on the perceived quality of service. As a result of their confidence in e-government websites, the general public will expect them to provide higher-quality public services [52]- [54]. Therefore, we draw the following hypothesis. − H3: trust in e-government websites will influence quality of information positively and significantly. − H4: trust in e-government websites will influence system quality positively and significantly. − H5: trust in e-government websites will influence service quality positively and significantly. Better quality of information will determine perceptions of the usefulness of an e-government website [29], [49], [52]. Face-to-face services are being phased out in favor of the high-quality information available on the internet, which is representative of government services. As a result, the perceived value of an e-government system will be impacted by the information's quality. Additionally, the level of information quality has an effect on user happiness. The quality of the information provided is crucial for increasing customer happiness [26], [49]. Therefore, we draw the following hypothesis. − H6: the quality of information has a positive and significant impact on the perceived usefulness. − H7: the quality of information has a positive and significant impact customer satisfaction. A good quality system is what the public and the government want because it is a medium for the government to deliver its services. In its website, the government must ensure good quality information, such as security and reliability because this affects the user's perception of the usefulness of an e-government website [23], [24], [44]. In previous research, system quality has also been confirmed to affect community satisfaction. This happens because the better the system built as seen from the speed of access, process and response, the higher the community's satisfaction [53], [55], [56]. Therefore, we draw the following hypothesis. − H8: system quality will influence perceived usefulness positively and significantly. − H9: system quality will influence customer satisfaction positively and significantly.
Prior studies have shown that the perceived utility of an e-government website is influenced by perceptions of service quality. As e-government websites are supposed to provide improved and streamlined services to satisfy users' issues, service quality is a determinant of satisfaction [23], [24], [44]. Therefore, we draw the following hypothesis. − H10: service quality will influence perceived usefulness positively and significantl. − H11: service quality will influence customer satisfaction positively and significantly. Online tax systems are anticipated to increase the performance of tax-related tasks and hence encourage user happiness, as systems that accomplish critical tasks are deemed more helpful. Users are more likely to be satisfied with a system if it is more useful. As long as the system is useful in carrying out its functions, it will benefit users by boosting their output [23], [24], [44]. People are more likely to enjoy and justify total value in an online system if it is regarded to be useful, so that investment pays off more quickly. It is projected that customer happiness will have an impact on how much value people derive from the services they receive. If users are happy, the e-government system's benefits will be felt more strongly as a result. − H12: perceived usefulness has a positive and significant effect on the perceived net benefit of e-government systems. − H13: user satisfaction has a positive and significant effect on the perceived net benefit of the e-government system.

METHOD
The purpose of this research is to discover the antecedents of service users' perceived net benefit of e-government system. It was quantitative study that was utilized to reach this end. Research designs that use quantitative approaches examine the link between several variables to see if they are valid [57].

Sample
The questionnaire was developed based on the previous research of [44] and distributed to 195 individual taxpayers, a five times of 39 questionnaire items as recommended by Hair et al. [58]. The research procedure begins by submitting letter of permition to the taxation agency. After obtaining permission, research assistants distributed questionnaires to the individual taxpayers. Purposive sampling, a nonprobability sampling technique, was employed in this investigation. Purposive sampling is a sampling technique that takes into account a number of factors before selecting a sample [59]. The criteria include an active taxpayer as an individual and is not representing a business or other sector entities; and since the system updated regularly, the other criterion is that the respondent was at least 6 months visiting the egovernment website (https://djponline.pajak.go.id/).

Measurements
This research employed nine variables including trust in government (TIG) [44], trust in technology (TIT) [44], trust in e-government website (TIEGW) [44], information quality (IQ), system quality (SQ), service quality (SERVQ) [44], perceived usefulness (PU) [44], user satisfaction (US), [44], and perceived net benefits (PB) [44]. Trust in government is quantified using four factors and the phrase "in my opinion, the government acts in the public interest". Trust in technology [43] is quantified using four factors and the statement "because the internet provides a high level of security, I am confident in using it to connect with government entities". Trust in e-government websites [43] is quantified using three criteria and the statement "a dependable online tax report submission system". Measurement of information quality [43] consists of six criteria and is illustrated by the statement "the information provided in the online submission system for tax returns is accurate".
Measurement of system quality [43] entails nine variables and uses the statement "the system includes a download form" as an example. Service quality measurement [43] entails four variables, as illustrated by the statement "the e-filing system's operation was exceptionally responsive to my requests". Perceived usefulness [43] is quantified using four factors and the statement "when it comes to preparing to submit my taxes, having an e-filing system would help me be more efficient". User satisfaction measurement [43] entails four variables, as illustrated by the statement "e-filing, in my opinion, fits the needs of my relationship with the tax office". Perceived net benefits [43] are quantified using five variables and the statement "e-filing saves time". All of the statements were given 7 answers ranging from strongly disagree to strongly agree.

Data analysis method
Partial least square-structural equation model (PLS-SEM) is is used to predict a model for theory development [58], [60]. This study utilized PLS-SEM to develop a model which can predict perceived net benefits of e-government service. Data analysis was conducted with the help of WarpPLS 6.0.

RESULTS AND DISCUSSION 4.1. Respondent demographic
This part includes information about the respondents' demographic characteristics. There are five things we asked respondents which are gender, age, latest education and length of time using e-filing. Table 1 shows the gender profiles of respondents divided into male and female. Most of our respondents were women with a total of 146 people and the rest were male with a total of 109 people.  Table 2 draws the age of the respondents. Most of the taxpayers who filled out our questionnaire were those aged 17−20 years. Then, this was followed by those aged 31−40 years, namely 19.6%. Respondents with the lowest age range were those aged 51−60 years or about 1.8%.  Table 3 summarizes the respondents' most recent educational profile. The majority of our respondents had completed senior secondary education, with 58.5% having completed senior high school/vocational high school and 31.9% having completed a bachelor's degree. With 0.4% and 0.7%, respectively, respondents with a graduate degree and a junior high school diploma had the least education. Table 4 contains the duration profile of e-filing usage. Most respondents were new users with a usage period of less than one year, which is 67.5%. Then, the data is followed by those who had used e-filing between 1−3 years, which is 15.2%. The number of respondents using this system for more than six years is 6.9%.

Validity and reliability test 4.2.1. Convergent validity
To find out whether the constructs and indicators meet the criteria of convergent validity, the authors present the factor loading value of each indicator. The data anlysis indicated the value is greater than 0.70 based on the results of the measurement of all indicators, indicating that the indicators are preserved. Indicators with a loading value of more than 0.70 indicate that the latent variable can explain the variance of each indicator by 50% [58]. By paying attention to the loading value of each indicator, the authors conclude that each indicator in the study meets the criteria for convergent validity.

Discriminant validity
The data analysis indicated the outer loading value is greater than the cross-loading value, based on the measurement performance. This means that each indicator can only represent a single variable and cannot represent any other variables [58]. The second method of determining discriminant validity is to compare the square root of average variance extracted (AVE) to the correlations between the latent variables in Table 5.  Table 5 shows the AVE square root and the relationship between latent variables. Previously, a construct was stated to meet the requirement for discriminant validity if the square root of AVE was greater than the correlation between latent variables [58]. Since the square root of AVE is greater than the correlation between latent variables for and variable, all variables in this analysis meet these criteria for discriminant validity. Table 6 shows the results of the calculation of composite reliability and Cronbach's alpha for each research indicator. We provide both scores for each variables. It is intended to measure the reliability to each variable employed in this study. Using Cronbach's alpha, the dependability of each research variable is summarized in Table 6. Cronbach alpha value of greater than 0.6 indicates that a variable is accurate when its composite reliability score is higher than 0.70 [58]. Based on the calculation results, the composite reliability value for each variable is above 0.70. The Cronbach's alpha value is then greater than or equal to 0.60. Each variable can be labeled dependable by focusing on these two tests.

Structural model evaluation
Once it is established that each variable in the analysis is accurate and consistent, the structural model is evaluated. The structural model is evaluated by examining the model's fit, the path coefficient, and R2. The first stage of model validation is to examine the model fit indices (model suitability). Table 7 summarizes the model fit indices and their associated parameters.  As seen in Table 7, the APC, ARS, and AARS each have a value of around 0.730. The P-value is less than 0.05 in each of the three instances. As a result, the model developed by the authors satisfies the requirements [58]. Additionally, the author's AVIF and AFIV scores, both of which are greater than the optimal 3.3 on a 1 to 5 scale, meet the model's requirements [61].
Additionally, Tenenhaus goodness of fit (GoF) has a 0.623 value (GoF). According to [61], the Tenenhaus GoF should be greater than 0.36 to qualify as an optimal model. Additionally, the SPR score is 1, indicating that the product meets all ideal characteristics [61]. Due to the fact that these investigations established that the model constructed was faultless, the RSCR value was set to 1. With an SSR value of 1, it is possible to conclude that the model is suitable [61]. The value of the NLBCDR is 1, indicating that the model is adequate. It is capable of reliably forecasting future occurrences, based on the author's model. Table 5 summarizes the structural model's test results based on the R-square (goodness of fit test). The findings indicated a close relationship between the total R-square value and the R-square value. Additionally, the TIEGW defined by trust in government and trust in technology are 0.70 or 70% [58]. The IQ, SQ, and SERVQ described by TIEGW have each R-square value of 0.72; 0.66; and 0.64 [58]. IQ, SQ and SERVQ can explain PU at 0.83, while the three variables are also able to predict US at 83% with an R-square value of 0.83. Last, the PB which can be explained by PU and US is 83% with an R-square value of 0.83.
Once the prerequisites have been completed, the hypothesis will be put to the test. We then looked at the path coefficient and P-value. Each hypothesis tested in the study has a path coefficient and a corresponding significance level shown in Table 8. Table 8 summarizes the thirteen hypotheses advanced in this review. Twelve of the thirteen hypotheses we suggested are accepted, while two were dismissed. The IQ to US hypothesis is dismissed. Simultaneously, the overall result, as summarized in Table 8, is positive.

Discussion
The study's objective is to determine the antecedents of service users' perceived net benefit of e-government system. There were thirteen hypotheses, and each was tested. According to the study's findings, there was a positive association between public trust and public trust in e-government websites. Individuals who believe in government are more likely to believe in the government's actions, one of which is requiring people to use e-government. This confirms several previous studies [26]- [28]. The digital government allows the government to build another image, which is faster, responsive and efficient, which is different from offline services which are slow, unresponsive, and therefore inefficient [52], [62], [63].
Additionally, the researchers discovered that people's faith in technology influenced their trust in e-government portals positively. Given the public's confidence in government-developed technology, it's logical to assume that consumers who utilize e-government services share that sentiment. This means that users' faith in technology becomes the basis for their trust in a specific technology, which encourages users to transact with confidence on an e-government website [64]- [66].
Trust in e-government websites improves information quality, the system, and how well it works as a result of these findings. Both quantity and quality of knowledge are affected, as was traditional. As a result, customers who believe in the website's reliability trust it. In order to keep the system running smoothly, e-government websites are essential. According to prior research, improvements should entail more community involvement. You have to have faith in government systems that are run by computers. According to the findings of this study, citizens who have faith in e-government websites receive better service.
While the quality of the system and service have an effect on these variables, the quality of the information itself is unaffected. As indicated by this finding, there is a correlation between user impressions of e-government website usability and system quality. As a result, the information offered may not be sufficient for consumers to file their taxes, diminishing its perceived usefulness. The quality of the system, the quality of the information, and the quality of the service all have an effect on how satisfied customers are with their services. This demonstrates that the system was developed effectively from an information systems perspective and has an effect on customer satisfaction [23], [44]. The information provided is sufficient in terms of quality to guide visitors through e-government websites. Users who are satisfied with the services they receive express a high degree of satisfaction with the service's quality.
Utility perceptions in e-government and user pleasure have an effect on users' perceived utility. Users believe that e-government websites boost their performance by increasing the effectiveness and efficiency of tax services, indicating that they view the system as beneficial from a perceived utility standpoint [23], [44]. E-government website utility perceptions were shown to be affected by user satisfaction, demonstrating that users are content with e-government websites, which in turn influences how pleased they are with e-government websites.

CONCLUSION
The findings indicate that citizens' faith in government and technology has an effect on their confidence in e-government websites. Credibility of e-government websites has a good and significant effect on the quality of information, the system, and its performance. As a result, consumer pleasure is influenced by a variety of variables. The perceived utility of e-government websites was shown to be influenced by the quality of the material as well as the system and service. This is also influenced by user satisfaction and perceptions of the utility of e-government.
The purpose of this study is to determine the antecedents of service users' perceived net benefit of e-government system. Academically, this study fills a hole in e-government research by examining the acceptance and use of e-government systems, as well as perceptions of their utility, which have not been adequately examined in previous studies. Indeed, various areas for improvement exist, including the consistency of offline public services and the information supplied on e-government portals. Due to the fact that not all government services are available online and due to the digital divide, the government must also improve the quality of offline government services. Consistent information dissemination might be improved; in addition to being reasonably useable and accessible, the material could be made more appealing.