The mediating role of behavioral intention on factors influencing user behavior in the E-government state financial application system at the Indonesian Ministry of Finance

Abstract

SAKTI is a treasury information system and financial report preparation newly implemented by the Ministry of Finance in 2022. The piloting phase has been conducted before official implementation to identify various challenges and the development of the application. The implementation of SAKTI heavily depends on the intention behavior and operator behavior to use the system. Therefore, the Unified Theory of Acceptance and Use of Technology 2 (UTAUT2) framework is used to examine the factors influencing behavioral intention and user behavior of the SAKTI system. The respondents involved in the study were 271 SAKTI users at the State Treasury Service Office. Hypothesis testing was conducted on the partial least squares-structural model equation (PLS-SEM) with SmartPLS 3.29 professional software. The research results provide empirical evidence that social influence and hedonic motivation do not affect the behavioral intention to use SAKTI. However, performance expectancy, effort expectancy, facilitating conditions, and habit impact the behavioral intention to use SAKTI. Facilitating conditions, habits, and behavioral intention influences the use of SAKTI. Behavioral intention to use SAKTI mediates the relationship between facilitating conditions and habit with the behavior of using the SAKTI system, with partial complementary mediation. The practical implication is that KPPN needs to provide special technical guidance for SAKTI operators to make the system easier for them to use. Enhancing security in the SAKTI application and improving internet capacity are important considerations to maximize the use of SAKTI.

Keywords:

• E-government financial application
• accounting information system
• behavioral intention
• unified theory of acceptance and use of technology

REVIEWING EDITOR:

• Anbazhagan Neelamegam, Alagappa University, India

SUBJECTS:

• Finance
• Business, Management and Accounting
• Information Technology

1. Introduction

In line with the development of the times and the era of Industry 4.0, where companies integrate automation with cyber technology, information technology has rapidly grown. Computers are no longer just tools for processing transactions; they have evolved into integrated systems within a company. The Industrial Revolution 4.0 in the government is marked by the implementation of the e-government system, an electronic-based system that utilizes information and communication technology to enhance efficiency, effectiveness, and accountability in the pursuit of good governance. The Indonesian government has issued Presidential Regulation Number 95 of 2018 concerning the Electronic-Based Government System (SPBE), regulating governance, management, information and communication technology audits, organizers, acceleration, and monitoring and evaluation of the SPBE. The State Financial Application System at the Institutional Level known as SAKTI is an information system that modernizes the implementation of state financial management functions from the perspective of budget users by integrating various application systems in state financial management. SAKTI is an information system newly launched by the Ministry of Finance on January 27, 2022, complementing the modernization of the General Treasurer of the State (BUN) financial management through the implementation of the State Treasury and Budget System as known SPAN. The function of managing state finances, from the stage of preparation to financial accountability, starting from the Work Unit level (Satker) up to the State Ministry/Institutions (K/L), is carried out in one system through SAKTI, which is a continuation of the implementation of the Integrated Financial Management Integration System (IFMIS). The development of SAKTI goes through several stages namely, the feasibility study stage, needs analysis, application design, and application development (piloting stage) (www.djpb.kemenkeu.go.id).

Previous research has been conducted from the piloting stage to the implementation of SAKTI by the Ministry of Finance. According to Rahman et al. (2023), SAKTI provides convenience for the Semarang State Treasury Office (KPPN 1) to carry out treasury and financial management processes, but improvements are needed to fully utilize the application. Information quality, system quality, and perceived ease of use affect user satisfaction, but access speed, available features in the system, and user training need to be enhanced (Rahayuningtyas, 2022). SAKTI facilitates the process of preparing financial reports due to single entry, integration of data between modules, real-time data presentation, and automatic generation of financial report components (Pambudi et al., 2022). The advantages of SAKTI include a centralized database, high-security level, ease of application installation, and better application performance consistency. The future challenges for the use of SAKTI are human resource readiness, internet capacity, and continuous training (Nasution & Nasution, 2022). Positive individual perception as well as technology and system are factors influencing user satisfaction in the piloting stage (Sutiono & Taufiqurahman, 2020). Amriani and Iskandar (2019) stated that empirically using the Delone and Mclean model approach, the implementation of SAKTI has not been successful in the piloting stage, and provided recommendations for future research using the UTAUT framework.

Several studies using the UTAUT framework to test its influence on the acceptance and adoption of e-government systems, which are mandatory systems for government and public sector institutions, are described by Ofosu-Ampong et al. (2023), explaining two UTAUT variables which are performance expectation and hedonic motivation influencing behavioral intention. Dbesan et al. (2023) indicate that the UTAUT2 framework influences behavioral intention, while according to Addy et al. (2022), performance expectation and social influence affect behavioral intention. Price value and habit do not affect doctors’ behavioral intention to use e-consultation during the COVID-19 pandemic (Dash & Sahoo, 2021b), while the UTAUT2 framework influences patients’ behavioral intention to use e-consultation (Dash & Sahoo, 2021a).

Based on Amriani and Iskandar (2019) who recommended for future research using the UTAUT framework, this study will use the UTAUT2 framework. Using the UTAUT2 framework is expected to provide empirical evidence for theoretical and applicative development, especially recommendations for the Ministry of Finance, and factors influencing the behavioral intention of operators to use SAKTI. Price value, which is one of the variables in UTAUT2, is not used in this study, because SAKTI does not impose costs on operators using it. Hedonic motivation is one of the variables that will be used in this study, although SAKTI is a mandatory system of the Ministry of Finance, researchers will explore whether user operators happily use it or see it as an obligation to fulfill. Manrai et al. (2021) found empirical evidence that behavioral intention is able to mediate facilitating conditions and habits with behavior using mobile payment services, which is a supporting transaction banking service system for customers. The intention of behavior as a mediating variable in the acceptance of the mandatory system of the Ministry of Finance will be tested in this study. The research questions arising are (1) whether the UTAUT2 framework influences behavioral intention and the use of SAKTI, and (2) whether behavioral intention is able to mediate the relationship between facilitating conditions and habits with behavior usage.

2. Theoretical framework and hypothesis development

2.1. State-level institution application system (SAKTI)

Minister of Finance Regulation of the Republic of Indonesia Number 171/PMK.05/2021 regarding the Implementation of the SAKTI System states that SAKTI is a system that integrates the processes of planning and budgeting, implementation, as well as accountability for the state revenue and expenditure budget in government institutions. It is part of the national financial management system. The State Treasury and Budget System (SPAN) is an integrated system covering all budget management processes, including budget preparation, budget document management, supplier management, procurement commitment management, payment management, state revenue management, cash management, accounting, and reporting. SAKTI consists of (1) Administration Module, (2) Budgeting Module, (3) Commitment Module, (4) Treasurer Module, (5) Payment Module, (6) Inventory Module, (7) Fixed Asset Module, (8) Receivables Module, and (9) Accounting and Reporting. SAKTI is used by the Budget Section (BA) of State Ministries/Institutions, the General Treasurer’s Budget Section (BA BUN) with User Access rights, the General Treasurer, and other units granted User Access rights (Muhtaromin, 2018; Nasution & Nasution, 2022; Nugroho & Lestyowati, 2020).

Rahayuningtyas (2022) provides empirical evidence that system quality, information quality, and perceived usefulness influence SAKTI user satisfaction. Nugroho and Lestyowati (2020), using the PIECES framework (Performance, Information, Economics, Control, Efficiency, and Services), provide empirical evidence that control plays a crucial role in user satisfaction, while performance is the least influential factor. Sutiono and Taufiqurahman (2020) provide empirical evidence that individual perception and technology/system are two factors influencing user satisfaction with SAKTI during the piloting phase. Amriani and Iskandar (2019) provide empirical evidence that system quality affects user satisfaction with SAKTI, while information quality and user satisfaction affect net benefits. According to Muhtaromin (2018), SAKTI users experience satisfaction with several advantages, namely real-time data, integrated applications, single database, and integrated information. Perceived usefulness and perceived ease of use influence the acceptance of SAKTI in the piloting stage (Prabowo, 2017).

2.2. Unified theory of acceptance and use of technology (UTAUT)

UTAUT is a framework developed by Venkatesh et al. (2003) based on eight models: (1) Theory of Reason Action (TRA), (2) Theory of Planned Behavior (TPB), (3) Technology Acceptance Models (TAM), (4) Model of PC Utilization (MPCU), (5) Diffusion of Innovation Theory (DOI), (6) Motivation Models, (7) Social Cognitive Theory, and (8) Combined TAM-TPB. It includes four main variables influencing behavioral intention: performance expectancy, effort expectancy, social influence, and facilitating conditions. Venkatesh et al. (2012) extended UTAUT to UTAUT2 by adding three variables influencing the behavioral intention to use information technology: habit, hedonic motivation, and price value.

Ofosu-Ampong et al. (2023) found empirical evidence that hedonic motivation and technology readiness are the strongest factors influencing the behavioral intention to use the government digital census. Dbesan et al. (2023) used the UTAUT2 framework to test its influence on doctors’ behavioral intention to use blockchain technology in government hospitals, where performance expectancy, effort expectancy, social influence, facilitating conditions, and price value affect it. Alkhwaldi et al. (2022) stated that performance expectancy, social influence, facilitating conditions, and technological task fit influence the behavioral intention to use the Human Resource Information System (HRIS) in public sector companies in Jordan. Addy et al. (2022) explained that the UTAUT2 model can provide strong and good predictions for electronic procurement acceptance, where performance expectancy, social influence, and habit are variables influencing the behavioral intention to adopt electronic procurement.

Dash and Sahoo (2021a) stated that UTAUT2 can influence patient adoption of digital health consultation services, and similarly, Dash and Sahoo (2021b) provide empirical evidence that the UTAUT2 framework can influence doctors’ adoption of e-consultation, where price value and habit have no influence. Kirat Rai et al. (2020) explained that the UTAUT framework can influence attitudes and behavioral intentions to use government-to-government (G2G) systems. Rinjany (2020) explained that effort expectancy and facilitating conditions influence the individual behavioral intention to use e-government. Wiafe et al. (2019) explained that the UTAUT model can contribute to the use of information systems in the maritime industry (INTTRA), where performance expectancy and facilitating conditions influence individual behavioral intention. Wu and Wu (2019) provided empirical evidence that performance expectancy, effort expectancy, and facilitating conditions influence individuals to continue using library information systems in the context of public services. Naranjo-Zolotov et al. (2019) provided empirical evidence that performance expectancy and facilitating conditions influence the individual behavioral intention to use e-participation systems. Mansoori et al. (2018) stated that performance expectancy, effort expectancy, and facilitating conditions affect citizens’ behavioral intention to use e-government systems. Saxena and Janssen (2017) provided empirical evidence that the UTAUT framework can influence individual behavioral intention to use open government data in India.

2.2.1. Performance expectancy (PE)

Performance expectancy is defined as the level of an individual’s belief that using information technology will help them gain benefits in their job performance (Venkatesh et al., 2003). Performance expectation is defined as the extent to which the use of digital census tools enables officials to perform their tasks effectively (Ofosu-Ampong et al., 2023). In the context of SAKTI usage, performance expectation refers to the level of belief that operators are capable of effectively, efficiently, and timely completing treasury tasks and preparing financial reports for work units. This is consistent with findings by Ofosu-Ampong et al. (2023), Dbesan et al. (2023), Alkhwaldi et al. (2022), Addy et al. (2022), Sanjeev et al. (2021), Wiafe et al. (2019), Naranjo-Zolotov et al. (2019), Mansoori et al. (2018) who found empirical evidence that performance expectation influences behavioral intention to use e-government.

H1. Performance expectancy positively influences individuals’ behavioral intention to use SAKTI

2.2.2. Effort expectancy (EE)

Effort expectancy is defined the level of ease an individual perceives in using a particular information technology system (Venkatesh et al., 2003). Effort expectation reflects users’ perception of how easy it is to use the Human Resources Information System (HRIS) in Human Resources Management (HRM) (Alkhwaldi et al., 2022). SAKTI users perceive the system to be easy to understand and use, thus requiring minimal effort to learn. This is consistent with the findings of Dbesan et al. (2023), which state that effort expectation influences doctors’ behavioral intention to use blockchain technology in healthcare, as well as with Rinjany (2020), Wu and Wu (2019), and Mansoori et al. (2018), who found empirical evidence that effort expectation influences individuals’ behavioral intention to use e-government.

H2. Effort expectancy positively influences individuals’ behavioral intention to use SAKTI

2.2.3. Social influence (SI)

Social influence is the extent to which an individual perceives that important others believe they should use a new system (Venkatesh et al., 2003). The effort of SAKTI users to learn the system is through peer-to-peer learning provided by colleagues, as well as through training conducted by the workplace. This is consistent with the research conducted Dbesan et al. (2023), which state that social influence affects doctors’ behavioral intention to use blockchain technology in healthcare services. Alkhwaldi et al. (2022) and Addy et al. (2022) provide empirical evidence that social influence affects the behavioral intention to use e-government. Dash and Sahoo (2021a) and Dash and Sahoo (2021b) give empirical evidence that social influence affects the behavioral intention of doctors and patients to use e-consultation. While, Saxena and Janssen (2017) explaining that respondents’ perceptions of family, coworkers, and superiors influence their behavioral intention to use open government data (OGD).

H3. Social influence positively influences individuals’ behavioral intention to use SAKTI

2.2.4. Facilitating conditions (FC)

Facilitating conditions are the level of belief or the extent to which an individual believes that organizational and technical infrastructure can support a new system (Venkatesh et al., 2003). Facilitating conditions can be defined as someone having easy access to electronic resources, such as computers, smartphones, internet connection, chat rooms, and other supportive conditions (Naranjo-Zolotov et al., 2019). Facilitating conditions such as computers, laptops, smartphones, internet, and dedicated staff are able to run the Jakarta Smart City (JSC) application program (Rinjany, 2020). Mansoori et al. (2018) stated that adequate facilitating conditions can encourage the use of e-government services. Facilitating conditions influence the behavioral intention to use blockchain healthcare services (Dbesan et al., 2023), as well as human resource systems Alkhwaldi et al. (2022). According to Dash and Sahoo (2021a) and Dash and Sahoo (2021b), facilitating conditions influence the behavioral intention of doctors and patients to use e-consultation.

Adequate facilitating conditions are able to influence the use of the INNTRA information system (Wiafe et al., 2019). Supportive facilities also contribute to increased usage of mobile banking in Malaysia (Ahmad & Yahaya, 2022). According to Mohd Thas Thaker et al. (2021), technical support and basic infrastructure needs can support the usage of internet banking. Manrai et al. (2021) explained that with a little basic training, rural women’s habits indirectly increase their usage of digital payment systems through their behavioral intentions. Based on evidence of Manrai et al. (2021), this research will test whether behavioral intention becomes a mediating variable in the context of SAKTI usage.

H4a. Facilitating conditions positively influence individuals’ behavioral intention to use SAKTI

H4b. Facilitating conditions positively influence the behavior of using SAKTI

H4c. Behavioral intention to use SAKTI mediates the relationship between facilitating conditions and behavior of using SAKTI

2.2.5. Habit (H)

Habit is the extent to which someone tends to perform behavior automatically due to learning (Limayem et al., 2007). Hsu et al. (2019) explained that habit is a perception construct that can be influenced by the environment and experiences that may be unconscious. The behavioral intention of SAKTI users arises from experiences using systems prior to SAKTI as well as other software, as well as the ability to adopt information technology. According to Ahmad and Yahaya (2022), the behavioral intention of Asnaf to use mobile banking is influenced by previous experiences using Internet banking and other financial applications. Habit also influences the behavioral intention of Sharia bank consumers to use m-banking (Iqbal et al., 2022), as well as internet banking (Mohd Thas Thaker et al., 2021). The habits of rural women bring about their behavioral intention to use digital payment systems (Manrai et al., 2021).

Consumers of Islamic banks who are accustomed to adopting information technology systems will use m-banking in their financial transaction activities (Iqbal et al., 2022), and consumers accustomed to digital systems will use ride-hailing applications for their daily activities (Chakraborty et al., 2021). Çera et al. (2020) and Owusu Kwateng et al. (2019) state that someone who is accustomed to digital systems will use m-banking in every financial transaction activity. Manrai et al. (2021) found empirical evidence that the behavioral intention to use digital payment systems mediates the relationship between habit and usage behavior. Like facilitating conditions, the behavioral intention to use SAKTI will be tested as a variable mediating the relationship between habit and usage behavior.

H5a. Habit positively influences individuals’ behavioral intention to use SAKTI

H5b. Habit positively influences the behavior of using SAKTI

H5c. Behavioral intention to use SAKTI mediates the relationship between habit and behavior of using SAKTI

2.2.6. Hedonic motivation (HM)

Hedonic motivation is the pleasure and comfort that arises when someone uses information technology (Venkatesh et al., 2012). When SAKTI users feel comfortable and happy with the system, they will intend to use SAKTI. This state aligns with Ofosu-Ampong et al. (2023), explaining that the pleasure and comfort of using the government’s digital census system influence users’ behavioral intention to use the system. Alkhwaldi (2023), Chakraborty et al. (2021), Addy et al. (2018) found empirical evidence that comfort and happiness when using information systems influence users’ behavioral intention to use the information system.

H6. Hedonic motivation has a positive impact on individuals’ behavioral intention to use SAKTI

2.2.7. Behavioral intention (BI) and behavior (B)

Behavioral intention is the willingness and effort of an individual to engage in underlying behavior (Upadhyay et al., 2022). When SAKTI users intend to use it, they will behave sustainably. This is supported by Ofosu-Ampong et al. (2023), Dash and Sahoo (2021a) and Dash and Sahoo (2021b), who explain that the behavioral intention of e-government users influences their usage behavior of the system.

H7. The individual’s behavioral intention to use SAKTI has a positive impact on the individual’s behavior of using SAKTI

Figure 1 is a research framework illustrating the relationship between independent and dependent variables as explained in the hypothesis development above.

Figure 1. Research model.

3. Research methodology

3.1. Data collection

The population in this study consists of users and operators of SAKTI at the level of the National Treasury Office (KPPN) work units. Respondents involved in the sample were selected using the purposive sampling method (Gray et al., 2007; Ranjit, 2012; Sugiyono, 2013) with the criteria that (1) respondents are employees of KPPN type A1 Semarang II, (2) respondents are users and operators of SAKTI at KPPN type A1 Semarang II. The selection of KPPN Semarang II as the sampling location is because it had the highest value for budget implementation realization compared to other KPPNs in 2022. The research sample size is determined based on Hair, Black, et al. (2019), stating that the minimum research sample size can be obtained by multiplying the number of question indicators in the research by 5 or 10. The research data was obtained from respondents who filled out a 1–5 Likert scale questionnaire in Google Forms (Appendix) via the S.id link. The informed written consent has been obtained in this study for the respondents who are involved as research samples. Ethical approval was obtained from ethics committee of Faculty of Business and Economics, Universitas Diponegoro with reg number 29/UN.7F2.6.2/AK/V, and head of KPPN A1 Semarang II with the letter number S-467/KPN.1402.

Research data were collected during the period of September 2023, and 271 out of a total of 486 respondents who were given the questionnaire provided complete answers. The response rate for this study was 55,76%, which consists of 150 males and 121 females, with 178 individuals aged between 21–40 years and 93 individuals aged between 40–59 years. There are 3 types of operators, namely payment operators consisting of 96 people, budget operators consisting of 100 people, and commitment operators consisting of 75 people. Regarding SAKTI usage experience, 55 respondents had less than 1 year of experience, 105 had 1–2 years, and 111 had 3–5 years of experience. Table 1 summarizes the demographic characteristics of the respondents involved in this study.

Table 1. Respondent demographic.

Respondent demographic	Category	Total	Percentage
Gender	Male	150	55,35%
	Female	121	44.,65%
Age	21–40	178	63,69%
	40–59	93	34,31%
Operators	Payments operators	96	35,8%
	Budgeting operators	100	36,6%
	Commitment operators	75	27,6%
Experience	<1 year	55	20.29%
	1–2 year	105	38.74%
	3–5 year	111	40.97%

3.2. Data analysis method

Data analysis in this study used the partial least squares structural equation modeling (PLS-SEM) method analyzed with SmartPLS 3.2.9 professional software. PLS-SEM is used in this study because it has stronger statistical power compared to CB-SEM, allowing for certain relationships to be more significant (Hair et al., 2022). Non-response bias testing was conducted because there were respondents who answered late using an independent sample t-test. Descriptive analysis presented minimum and maximum values, mean, and standard deviation in this study. Hair et al. (2022) explained that there are two evaluations, (1) evaluation for the measurement model and (2) evaluation for the structural model. The evaluation for the measurement model consists of four steps, indicator reliability, internal consistency reliability, convergent validity, and discriminant validity. Evaluation for the structural model also consists of four steps, assessing the collinearity of the structural model, assessing the explanatory power of a model, assessing the predictive capability of a model, and evaluating the significance and relevance of the relationships in the structural model.

4. Analysis and result

4.1. Non-response bias test

There were 30 respondents out of the 271 who answered the questionnaire late. Therefore, a non-response bias test was conducted on the responses of those who answered late to investigate whether there were differences in responses between those who answered and returned the questionnaire on time. If the significance value (2-tailed) > 0.05, then there is no difference in responses between respondents who answered on time and those who did not. Table 2, concerning the non-response bias test, shows significance values (2-tailed) for respondents who answered on time and those who answered late, 0.495 and 0.362, respectively. The test results indicate no difference in responses among respondents, allowing the late responses to be utilized as research data for hypothesis testing.

Table 2. Non-response bias test.

Respondent answer	N	Means	Levene’s test		t-test
			F	Sig	t	Sig (2-tailed)
On time	241	123,88	5,716	0,018	−0,684	0,495
Late	30	125,94			−0,920	0,362

4.2. Descriptive statistics

Descriptive analysis was conducted to determine the minimum, maximum, mean, and standard deviation of the variables used in the study. The average value of performance expectation falls within a high range at 17.04, approaching the maximum value of 20. Other variables, such as effort expectation, social influence, facilitating conditions, habit, and hedonic motivation, have mean values within a moderate range above the midpoint. Similarly, for behavioral intention and behavior variables, the mean values are within a moderate range, above the midpoint. The standard deviation values for each variable are below the mean, indicating that respondents’ answers fall within the range of the mean, with no abnormal data. Table 3 shows descriptive statistic results.

Table 3. Descriptive statistics.

Variable	N	Minimum	Maximum	Mean	Std. deviation
PE	271	9	20	17.04	2.602
EE	271	9	20	16.10	2.755
SI	271	10	20	16.54	2.298
FC	271	8	20	16.85	2.322
HM	271	8	20	16.21	2.604
H	271	9	20	16.61	2.349
BI	271	7	15	12.57	1.921
B	271	7	15	12.20	1.979

4.3. Measurement model

4.3.1. Indicator reliability

The assessment of the outer model includes indicator reliability, internal consistency reliability, convergent validity, and discriminant validity. Hair, Risher, et al. (2019) and Hair et al. (2022) state that a good outer loading value is above 0.7, while values between 0.4–0.7 are still acceptable, considering internal consistency reliability and convergent validity. The results of the outer loading test show that one indicator has a value below 0.7, specifically 0.646 for the first question indicator of the social influence variable (Table 4 and Figure 2). The values of CR and AVE are considered to retain the indicator aforementioned.

Figure 2. Results of the outer model assessment.

Table 4. Evaluation of outer loading, CR and AVE.

Variable		Outer loading	CR	AVE
Performance expectancy	PE1	0,847	0,890	0,669
	PE2	0,834
	PE3	0,847
	PE4	0,740
Effort expectancy	EE1	0,830	0,892	0,673
	EE2	0,815
	EE3	0,816
	EE4	0,821
Social influence	SI1	0,646	0,837	0,563
	SI2	0,799
	SI3	0,787
	SI4	0,759
Facilitating condition	FC1	0,747	0,861	0,607
	FC2	0,783
	FC3	0,830
	FC4	0,755
Habit	H1	0,749	0,876	0,638
	H2	0,821
	H3	0,805
	H4	0,817
Hedonic motivation	HM1	0,769	0,889	0,668
	HM2	0,842
	HM3	0,845
	HM4	0,811
Behavioral intention	BI1	0,892	0,884	0,719
	BI2	0,858
	BI3	0,790
Behavior	B1	0,796	0,875	0,700
	B2	0,847
	B3	0,866

4.3.2. Internal consistency reliability and convergent validity

The value of internal consistency reliability refers to the composite reliability (CR), and a good CR value is in the range of 0.7–0.9. Convergent validity value refers to the average variance extracted (AVE), and a good AVE value is equal to or greater than 0.5 (≥0.5). The composite reliability (CR) for each variable falls within the range of 0.7–0.9, indicating that the variables used in the study are reliable. The AVE values for each variable are ≥ 0.5, confirming the validity of the variables in this study (Table 4).

4.3.3. Discriminant validity

Discriminant validity assessment involves the Fornell-Larcker criterion and the Heterotrait-Monotrait Ratio (HTMT) criterion. The Fornell-Larcker criterion compares the square root of AVE for each construct with the correlation values between constructs and other constructs. A good Fornell-Larcker value is achieved when the square root of AVE is greater than the correlation values between constructs and other constructs. The HTMT criterion is the ratio of the correlation between traits to the correlation within traits. HTMT is also the average of all indicator correlations across all constructs measuring different constructs relative to the average of correlations of indicators measuring the same construct.

A good HTMT value is below 0.90 for a 95% confidence level. Discriminant validity testing with the Fornell-Larcker criterion shows that the square root of AVE for each construct is greater than the correlation values between constructs and other constructs. Discriminant validity assessment with the HTMT criterion shows that the correlation values between constructs are below 0.90 (Tables 5 and 6), concluding that the study variables are valid.

Table 5. Fornell-Larcker.

	B	BI	EE	FC	H	HM	PE	SI
B	0,837
BI	0,688	0,848
EE	0,628	0,718	0,821
FC	0,622	0,696	0,697	0,779
H	0,617	0,708	0,675	0,673	0,799
HM	0,612	0,702	0,708	0,744	0,749	0,817
PE	0,615	0,691	0,751	0,650	0,621	0,689	0,818
SI	0,549	0,629	0,601	0,633	0,661	0,643	0,573	0,750

Bold is the value of square root of AVE for each construct.

Table 6. Heterotrait-monotrait ratio (HTMT).

	B	BI	EE	FC	H	HM	PE	SI
B
BI	0,863
EE	0,767	0,870
FC	0,788	0,872	0,858
H	0,766	0,867	0,818	0,842
HM	0,747	0,847	0,843	0,830	0,852
PE	0,755	0,843	0,899	0,802	0,751	0,820
SI	0,711	0,802	0,762	0,832	0,852	0,825	0,726

4.4. Structural model

4.4.1. Assessment of structural model collinearity

The assessment of the inner model consists of evaluating the collinearity of the structural model, assessing the explanatory power of a model, evaluating the predictive power of a model, and assessing the significance and relevance of the relationships in the structural model. Hair et al. (2022) and Hair, Risher, et al. (2019) state that the structural model’s collinearity must be assessed to ensure unbiased regression results. The assessment of collinearity in the structural model is done through the calculation of the variance inflation factor (VIF). If the VIF value is above 5, it indicates collinearity among constructs. VIF values between 3–5 are acceptable, although there is a possibility of collinearity among constructs, and the ideal VIF value is less than three (VIF <3).

The assessment of collinearity in the structural model shows that the outer VIF is below 3, and the inner VIF for the relationships of performance expectations, social influence, facilitating conditions, and habits have values <3. The relationships between business expectations, hedonic motivation, and behavioral intention variables have VIF values between 3–5, namely 3.056 and 3.373, indicating good quality and no multicollinearity. The inner VIF value for the relationship between behavior and facilitating conditions, habits, and behavioral intention is <3, indicating no multicollinearity. Table 7 show VIF outer and inner value.

Table 7. VIF outer and inner value.

Variable		Outer VIF	Inner VIF values
			Behavioral intention	Behavior
Performance expectancy	PE1	2,111	2,654	–
	PE2	2,059
	PE3	1,480
	PE4	2,111
Effort expectancy	EE1	1,876	3,056	–
	EE2	1,713
	EE3	1,919
	EE4	1,868
Social influence	SI1	1,151	2,102	–
	SI2	1,716
	SI3	1,733
	SI4	1,454
Facilitating condition	FC1	1,471	2,774	2,217
	FC2	1,779
	FC3	1,987
	FC4	1,494
Habit	H1	1,552	2,790	2,295
	H2	1,892
	H3	1,748
	H4	1,927
Hedonic motivation	HM1	1,546	3,373	–
	HM2	2,005
	HM3	2,264
	HM4	1,954
Behavioral intention	BI1	2,137	–	2,436
	BI2	1,830
	BI3	1,558
Behavior	B1	1,552	–	–
	B2	1,631
	B3	1,883

4.4.2. Assessment of the strength of explanation and predictive power of the structural model

The assessment of the strength of the explanation of a model measures how much independent variables (endogenous constructs) can influence dependent variables (exogenous constructs). The strength of the explanation of a model is assessed through the coefficient of determination (R2), ranging from 0 to 1. The higher the value, the higher the explanatory power of the endogenous construct on the exogenous construct. The assessment results show a value of 0.654 for the influence on behavioral intention and 0.522 for the influence on behavior (Table 8).

Table 8. R2 value and Q2 value.

	R^2	Adjusted R^2	Q^2
Behavioral intention	0,661	0,654	0,462
Behavioral	0,528	0,522	0,356

The UTAUT2 framework can influence behavioral intention by 65.4% and behavior by 52.2%. The explanatory power of the UTAUT2 framework for behavioral intention and behavior is a moderate explanatory strength above 50%. The predictive assessment of a model is conducted to determine the accuracy of the model’s predictions. The predictive assessment of a model is done by looking at the Q2 value, and the assessment results show Q2 values of 0.462 for behavioral intention and 0.356 for behavior (Table 8).

4.4.3. Assessment of significance and relevance of structural model relationships

The significance and relevance of structural model relationships are assessed to provide empirical evidence for accepting or rejecting research hypotheses. Hair et al. (2022) explain that the assessment of the significance and relevance of structural model relationships is carried out to provide empirical evidence that independent variables significantly influence dependent variables. This assessment examines the t-values and p-values for all coefficients of the structural path, with t-values greater than the critical values of 1.28 (10% significance), >1.65 (5% significance), and >2.33 (1% significance) for one-tailed assessment. In addition to using t-values, the assessment of significance should consider p-values for the significance of relationships between variables. At a significance level of 5%, the p-value should be below 0.05, while for a significance level of 1%, the p-value should be below 0.01 as shown in Figure 3 (the result of inner model assessment) and Table 9 (the result of hypothesis testing).

Figure 3. Results of inner model assessment.

Table 9. Results of hypothesis testing.

Hypothesis	Path	t-statistic	p value	Decision	Mediation
H1	PE →BI	2,348	0,019	Accepted	–
H2	EE →BI	2,812	0,005	Accepted	–
H3	SI →BI	1,769	0,077	Rejected	–
H4a	FC →BI	2,364	0,018	Accepted	–
H4b	FC →B	2,987	0,003	Accepted	–
H4c	FC →BI B	2,112	0,035	Accepted	Partial mediation
H5a	H →BI	3,114	0,002	Accepted	–
H5b	H →B	2,652	0,008	Accepted	–
H5c	H →BI →B	2,645	0,008	Accepted	Partial mediation
H6	HM →BI	1,151	0,250	Rejected	–
H7	BI →B	5,848	0,000	Accepted	–

5. Discussion and conclusion

The UTAUT2 framework effectively influences behavioral intention and behavior in the context of e-government usage, specifically the State Financial Application System (SAKTI). The explanatory power of the UTAUT2 framework is 66.1% for individual behavioral intention and 52.8% for the behavior of using SAKTI. Users of SAKTI feel that the system can efficiently and effectively accomplish treasury tasks (budgeting, disbursement, and inventory mechanisms) and the process of preparing financial reports. Therefore, SAKTI users intend to use the system, based on a t-value of 2.348 and a p-value below 0.05 (H1 accepted). Male SAKTI users feel that their performance increases more than females through the system, thus enhancing their behavioral intention. Performance expectations positively influence behavioral intentions, consistent with the research findings of Ofosu-Ampong et al. (2023), Dbesan et al. (2023), Alkhwaldi et al. (2022), and Addy et al. (2022). The ease of use factor is also perceived by SAKTI operators, which aligns with the training provided by KPPN II Semarang, thereby increasing operators’ behavioral intention to use the system (t-value of 2.812 and p-value below 0.01, H2 accepted). Female SAKTI operators perceive it easier to use the system compared to male operators. Effort expectancy affects behavioral intentions, in line with the research findings of Dbesan et al. (2023) and Rinjany (2020).

SAKTI users also feel the influence of colleagues, but it does not affect their behavioral intention to use the system. This is consistent with the intensive training provided by KPPN II, so individual influence does not matter. Social influence does not affect behavioral intentions, indicated by a t-value of 1.769 and a p-value of 0.077 (H3 rejected), in line with the research findings Wiafe et al. (2019), which stated that the behavioral intention of INTTRA users is not influenced by social influence, and Naranjo-Zolotov et al. (2019), provided empirical evidence that social influence does not affect individual behavioral intentions to use e-government. SAKTI users feel that the available supporting facilities greatly assist in using the system; computer facilities, large bandwidth capacity, system facilities, and system security to prevent hacking, thus leading to an intention to use. Facilitating conditions affect behavioral intentions, indicated by a t-value of 2.364 and a p-value of 0.018 below 5%, thus H4a is accepted. Male operators feel that the facilities are adequate, thus affecting their behavioral intention to use the system compared to female operators. These research findings are consistent with Dbesan et al. (2023), Alkhwaldi et al. (2022), Dash and Sahoo (2021a), Dash and Sahoo (2021b), Rinjany (2020) and (Mansoori et al., 2018).

Adequate supporting facilities can influence the behavior of using SAKTI in daily financial report preparation and treasury tasks. Hypothesis testing results in a t-value of 2.987 and a p-value of 0.003 below 1%, thus H4b is accepted. Male SAKTI users are more inclined to use it compared to female users. Facilitating conditions directly affect the behavior of use, consistent with the research conducted by Ahmad and Yahaya (2022), Mohd Thas Thaker et al. (2021) and Wiafe et al. (2019). Male SAKTI users feel that the provided facilities are adequate, thus influencing their behavioral intention and behavior in using the system in their daily activities compared to females. Behavioral intention partially mediates the relationship between facilitating conditions and behavior, indicated by a t-value of 2.112 and a p-value of 0.035 below 5%, thus H4c is accepted. The mediating nature of behavioral intention is a complementary partial mediation, where the direct and indirect effects of facilitating conditions on behavior are significant in the same direction. The behavioral intention, serving as a mediating variable, aligns with the findings Manrai et al. (2021), demonstrating that the behavioral intention of rural women can mediate the relationship between facilitating conditions and the use of digital payment systems.

SAKTI users who are accustomed to using computers, IT advancements, and other systems are influenced by their behavioral intentions to use the system. Female operators show more intention than male operators to use SAKTI, despite both being accustomed to using systems and computers. The statistical t-value of 3.114 and p-value of 0.002, below 1%, thus H5a is accepted, consistent with the research of Ahmad and Yahaya (2022), Mohd Thas Thaker et al. (2021) and Manrai et al. (2021). Conversely, SAKTI usage behavior is influenced by habits in using computers and IT applications (t-value = 2.645, p = 0.008, H5b accepted). In contrast to behavioral intentions, male operators use SAKTI in their daily activities compared to females. Habit influences usage behavior, consistent with the findings of Iqbal et al. (2022), Çera et al. (2020), Chakraborty et al. (2021) and Owusu Kwateng et al. (2019). The behavioral intention of operators to use SAKTI mediates the relationship between facilitating conditions and habits with SAKTI usage behavior. Hypothesis testing yields a t-value of 2.645 and a p-value of 0.008, below 1%, thus H5c is accepted. Manrai et al. (2021) found empirical evidence that behavioral intention mediates the relationship between habits and behavior in the context of digital payment system usage. The mediating nature of behavioral intention is partially complementary, where the direct and indirect effects of habits on behavior are significant in the same direction.

SAKTI users feel comfortable, happy, and satisfied when using the system; however, it does not influence individual behavioral intentions to use SAKTI (t-value = 1.151, p = 0.250, H6 rejected). Both male and female operators’ behavioral intentions are not influenced by hedonic motivations. According to Alkhwaldi et al. (2022), hedonic motivation does not affect individual behavioral intentions to use e-government. The behavioral intention to use SAKTI influences usage behavior (t-value = 5.848 and p-value = 0.000, H7 accepted). Both male and female operators intend to use SAKTI, thus affecting their usage behavior. This is consistent with the research conducted by Ofosu-Ampong et al. (2023), Dash and Sahoo (2021a), and Dash and Sahoo (2021b).

The UTAUT2 framework can influence behavioral intentions and technology usage behavior in the context of e-government in Indonesia. Performance expectations, effort expectations, facilitating conditions, and habits are variables influencing the behavioral intentions of operators to use SAKTI. The behavior of use operators SAKTI is influenced by facilitating conditions, habits, and behavioral intention. Behavioral intention can partially complementarily mediate the relationship between facilitating conditions and habits with the operator’s SAKTI use behavior. Social influence and hedonic motivation do not affect the operator’s behavioral intentions to use.

6. Implications

This research’s theoretical contribution lies in developing the UTAUT2 framework application and testing its influence on the behavioral intention and behavior of e-government users, particularly SAKTI users. Performance expectations, effort expectations, facilitating conditions, and habits are identified as robust variables influencing individual behavioral intention and behavior. SAKTI users expect that the system will enhance task completion and improve performance, coupled with ease of system use, impacting their behavioral intention. Adequate supporting infrastructure facilities, including internet access, hardware used to operate the SAKTI system, and features within the system, are identified as factors influencing behavioral intention and behavior. Users’ habits, developed through exposure to information technology, computer operation, and familiarity with digital systems, also affect their behavioral intention and system usage. This research supports empirical evidence of Dbesan et al. (2023) as well as Alkhwaldi et al. (2022), who used the UTAUT framework to influence behavior intention to use e-government. This study also supporting (Addy et al., 2022), Dash and Sahoo (2021a), and (Dash & Sahoo, 2021b), who were use the UTAUT2 framework to give empirical evidence of the factors influencing the behavioral intention to use e-government.

The practical benefits of the research serve as considerations for policymakers, indicating that the conducted training has made SAKTI users adept at using the system. Structured technical guidance should be frequently conducted by the Ministry of Finance, particularly the KPPN. Adequate supporting facilities and system features to enhance SAKTI users’ performance should be further improved, especially security factors and internet capacity for system usage.

7. Limitation and future research

The limitations of this study include the low response rate of respondents to the provided questionnaires and the delayed responses collected beyond the researcher’s specified deadline. Future research should focus on maximizing the distribution and collection of questionnaires, with the assistance of an intermediary or liaison appointed to facilitate communication between respondents and researchers. Further development of additional variables that can be integrated into the UTAUT2 framework is still possible in future research. Concepts such as the D&M model and Task Technology Fit (TTF) model can be integrated with UTAUT2.

Ethical approval statement

The study was approved by ethics committee of Business and Economics Faculty, Universitas Diponegoro with reg number 29/UN.7F2.6.2/AK/V, and KPPN A1 Semarang II with the letter number S-467/KPN.1402.

Authors’ contributions

Conceptualization: Wahyu Meiranto, Fortunella Farlyagiza. Data collection: Wahyu Meiranto, Fortunella Farlyagiza, Etna Nur Yuyetta, Elen Puspitasari. Formal analysis: Wahyu Meiranto, Fortunella Farlyagiza, Etna Nur Yuyetta. Methodology: Wahyu Meiranto, Faisal Faisal, Etna Nur Yuyetta. Project Administration: Elen Puspitasari. Software: Fortunella Farlyagiza. Validation: Wahyu Meiranto, Faisal Faisal. Writing – original draft: Wahyu Meiranto. Writing – review and editing: Faisal Faisal, Elen Puspitasari. All authors agree to be accountable for all aspects of the work.

Informed consent

Informed written consent is applicable to tis research.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Additional information

Funding

This research did not receive any research grants or funding from other sources.

Notes on contributors

Wahyu Meiranto

Wahyu Meiranto is a lecturer at the Accounting Department, Faculty of Business and Economics, Universitas Diponegoro, Indonesia. He has published paper in international journals with interest research in accounting information system, behavioral accounting.

Fortunella Farlyagiza

Fortunella Farlyagiza is an alumni of Accounting and Business Department, School of Vocation, Universitas Diponegoro, Indonesia. She has expertise in accounting and taxation.

Faisal Faisal

Faisal Faisal is lecturer and Professor at the Accounting Department, Faculty of Business and Economics, Universitas Diponegoro, Indonesia. He has published many papers in International journals with interest research in green accounting, corporate responsibility, sustainability and corporate governance. He is now Dean of Faculty of Business and Economics.

Etna Nur Afri Yuyetta

Etna Nur Afri Yuyetta is a lecturer at the Accounting Department, Faculty of Business and Economics, Universitas Diponegoro, Indonesia. She has published paper in international journals with interest research in accounting, capital market, corporate governance.

Elen Puspitasari

Elen Puspitasari is a lecturer at the Accounting Department, Faculty of Business and Economy, Universitas Stikubank, Indonesia. She has published paper in international journals with interest research in accounting, finance, and capital market. She is now Vice Rector III.

Appendix.

Research questionnaire

Table

Variable		Indicators
Performance expectancy	1.	SAKTI is an application that is useful in my work.
	2.	SAKTI makes it easier for me to do my job.
	3.	SAKTI allows me to do my work faster than other financial applications.
	4.	My productivity can increase by using SAKTI.
Effort expectancy	1.	I will not have difficulty learning how to use the SAKTI application.
	2.	I feel comfortable doing my job using the SAKTI application.
	3.	It is easy for me to become proficient in using the SAKTI application.
	4.	Overall, I find SAKTI easy to use.
Social influence	1.	People I respect suggest that I use SAKTI.
	2.	People important to me think that I should use SAKTI.
	3.	People in my environment who use the SAKTI application have higher prestige than those who do not.
	4.	I use the SAKTI application for work because many people also use it.
Facilitating condition	1.	I have the necessary resources (computer/internet) to use the SAKTI application.
	2.	I have the necessary knowledge to use the SAKTI application.
	3.	SAKTI is compatible with other systems that I use.
	4.	SAKTI has guidelines for its usage.
Hedonic motivation	1.	Using the SAKTI application is enjoyable for me.
	2.	Using the SAKTI application is very entertaining for me.
	3.	Using the SAKTI application is very rewarding for me.
	4.	Using the SAKTI application is very interesting and enhances my prestige.
Habits	1.	Using the SAKTI application is a habit for every operator at the DJPb agency level.
	2.	I feel satisfied when using the SAKTI application.
	3.	Every operator at the DJPb agency level should use the SAKTI application.
	4.	Using the SAKTI application is something I do without thinking for my job.
Behavioral intention	1.	I intend to continue utilizing SAKTI in my future work.
	2.	I will use SAKTI in my daily life.
	3.	Considering that I have a computer capable of accessing the internet, I will continue to use the SAKTI application.
Use behavior	1.	Sometimes, I use SAKTI for my work.
	2.	I often use the SAKTI application to perform daily tasks.
	3.	I always use the SAKTI application in all of my work.