Determinants of Intention to Use Village Fund Information System

This study aims to examine factors that affect intention to use of village fund information system (SISKEUDES). Respondents in this research are village apparatus as operators of SISKEUDES. Data were collected using questionnaires. Sampling is conducted using the census sampling method, producing 212 respondents. This study uses the Structural Equation Model by using Partial Least Square (PLS) as a method of analysis. The result indicates that the quality of human resource and social factor influence the use of SISKEUDES. In addition, perceived system quality, perceived usefulness and perceived ease of use partially mediate the relationship between the quality of human resource and the intention to use SISKEUDES. In contrast, perceived information quality does not mediate the relationship between the quality of human resource and the intention to use of SISKEUDES.


INTRODUCTION
Regional autonomy is the government's effort in prospering people. Regional autonomy was born from a paradigm that the centralized government made some regions feel less attention that, eventually, led to conflict, both vertical and horizontal. Provision of regional autonomy allows each region to be able to optimize its potential resources and to encourage regional development according to its economic characteristics, social culture and geography.
Indonesia entered the decentralized government era characterized by the issuance of Law No. 32 and 33 of 2004 on Regional Government and Regional Finance. In the decentralization era, regions will have authority in the form of responsibility, regional development, politics, public service until the utilization and management of its financial resources. With the delegation of authority, the regions are expected to build its territory in accordance with their needs and potentials.
One of the government programs in the era of decentralization is village fund, which is regulated in Law No. 6 of 2014. The central government through the Village Ministry distributes village fund to all villages in Indonesia. The village fund gives authority to the village and stimulus to strengthen the village economy. It is expected to provide opportunities for village communities to participate in development at the village. To support the implementation of Law No. 6 of 2014, in 2015, the central government allocated Rp20.7 trillion transferred to 74.093 villages across Indonesia. The amount of the village fund increased to be Rp46.9 trillion in 2016 and it was transferred to 74.754 villages (Direktorat Jenderal Bina Pemerintah Desa 2017). The increasing of village fund shows the commitment of the central government to accelerate development of the village in accordance with Nawa Cita of President Joko Widodo. Infrastructure development in rural areas is the main target of village fund utilization.
However, the village fund program raises doubts and debates from various parties. Provision of funds to villages is considered effective in reducing the development gap between villages. It also provides opportunities for each village to manage its development program. On the other hand, some parties questioned the village's readiness to manage village funds, in particular, related to the lack of quality of human resources, lack of information systems, and lack of accounting skills. As a result, the head of the village administration has the potential to mismanage village funds which can be categorized as corruption. As evidence, the Corruption Eradication Commission (KPK) revealed that fourteen potential problems in village fund management were divided into four aspects, namely regulatory and institutional aspects, management aspects, supervision aspects and aspects of human resources. Meanwhile, according to Indonesian Corruption Watch (ICW), the potential for corruption in village funds is due to weak governance aspects, such as the APBDesa and village accountability reports that do not meet standards and are prone to manipulation. In addition, weak regulations and institutions are also factors that cause corruption in village funds. ICW noted that, during the 2016 period, there were 62 cases of corruption that occurred in the village government with a value of state losses of Rp18 billion.
To overcome the potential problems related to the village fund, the Ministry of Home Affairs through the Directorate General of Village Administration, in cooperation with the State Development Audit Agency (Badan Pengawasan Keuangan dan Pembangunan or BPKP), launched a village financial management-based system application/ software. Initially, the application developed by BPKP was named as SIMDA Desa. In a further development, the name was changed to the Sistem Keuangan Desa (SISKEUDES) and began to be socialized to all villages in Indonesia. According to the Ministry of Home Affairs Circular No. 143/8359/BPD, this application was expected to be used gradually by all villages in Indonesia starting in 2016. The SISKEUDES application is expected to facilitate the village apparatus in managing of village fund in accordance with the Permendagri No. 113 of 2014 related to management of village fund. The use of this application is a prerequisite for disbursing village fund.
The implementation of new technology in an organization is not easy, especially when it is related to how the user receives technology that will be adopted. This is because the application of new technologies will affect the entire organization, especially human resources. User factors must be the main concern to be considered in the implementation of the new system because the level of readiness of the user to accept the new system affects the determination of the success and failure of the system (Kustono 2011). That also happened in the application of SISKEUDES. System implementation has many obstacles. This application system is implemented when the village fund program has been running, and some villages have used software applications built by the private sector. In addition, many features in SISKEUDES are not functioning. This is further aggravated by the lack of training for village officials. These things, of course, can influence the intention of village officials to use the system.
Based on the discussion above, this study intends to examine the determinants of intention to use the SISKEUDES. This study extends existing literature in several ways. First, this study employs an extended model of technology acceptance model (TAM) by adding relevant factors in the context of the SISKEUDES. In the technology acceptance studies, TAM is empirically proven, and it is recognized as a flexible model (Alomary and Woollard 2015). There are some extended models of TAM that take into account several factors, such as TAM 2 (Venkatesh and Davis 2000), combined TAM and Theory of Planned Behavior (Taylor and Todd 1995), UTAUT and TAM 3 (Venkatesh and Bala 2008). However, we propose the new modified of TAM that integrates some relevant factors in the context of the SISKEUDES, namely human resources quality, perceived system quality, perceived information quality, and social factor. Second, to the best our knowledge, this study is the first study that examines determinants of intention to use the SISKEUDES in Indonesia by using modified TAM. TAM has proven to perform well in voluntary and compulsory environments (Lai 2017).
The aims of this study are: (1) to provide empirical evidence whether quality of human resource and social factor influence the interest in using of SISKEUES, and (2) to examine whether perceived information quality, perceived system quality, perceived usefulness and perceived of ease of use mediates the relationship between the quality of human resource and the intention to use of the SISKEUDES. The research was conducted in Ponorogo Regency because Ponorogo Regency was chosen by BPKP as a pilot project for implementation of the SISKEUDES in the East Java Province in 2016. Furthermore, Ponorogo Regency has never implemented any village fund application including applications provided by the private sector (Muhlis 2017). This implies that the village apparatus does not have the experience to handle financial applications. As a result, the study on their acceptance on the SISKEUDES might provide a pure and valuable respond that could be used to evaluate the implementation of the SISKEUDES.

Village Financial Information System (SISKEUDES)
To oversee village financial accountability, the village apparatus is expected to implement village financial information systems to manage village fund. Village Financial Information System (SISKEUDES) is a computer application developed by the BPKP in cooperation with the Directorate General of Government and Village Development. The aim is to assist the village apparatus in implementing village fund management. This application is expected to help budget users to minimize data recording errors related to the use of funds in the village. With this application, village fund management can be carried out in a transparent, accountable and participatory manner.
The SISKEUDES application manages the village fund cycle that consists of planning and budgeting, administration, bookkeeping and reporting phases. At each stage, the application produces documents and reports in accordance with the provisions of the legislation. Compared to similar applications used to manage village finances, the SISKEUDES application has several advantages: (1) in accordance with the laws and regulations, (2) facilitating village apparatus in village finance governance, (3) simple and user-friendly, (4) equipped with the internal control system (built in internal controls), and (5) supported by manual user (BPKP 2015).

Technology Acceptance Model (TAM)
The technology acceptance model (TAM) is widely used by some prior studies in technology acceptance studies. Davis, in 1986, developed TAM in his doctoral study and it was an adaptation of the Theory of Reasoned Action (TRA) (Davis 1989). TAM provides a framework to measure users' perceptions of and intentions to use new technology within and throughout organizations (Alomary and Woollard 2015). The basic of TAM model includes only two specific beliefs, namely perceived usefulness and perceived ease of use. Unlike TRA, this model does not include subjective norms because of the psychometric problem (Wu et al. 2011

Human Resource Quality, Social Factor, Perceived System Quality, and Perceived Information Quality
In prior studies on the modified TAM, there were several user characteristics proposed by some researchers such as demography variables (i.e. gender and experience), personal traits, and self-efficacy (Wixom and Todd 2005). The human resource quality, social factor, perceived system quality, and perceived information quality are contextual variables added in our modified TAM. In this study, we assume human resource quality as an important determinant in using the SISKEUDES. The human resource quality in this study closes to self-efficacy. This refers to the user's confidence in his ability to perform certain tasks. Venkatesh and Davis (2000) add social factor or social influence in TAM 2. They use subjective norms as a proxy of such variables. They argue that, in a mandatory context, subjective norms have a direct effect on intention through the mechanism of compliance. In contrast, in voluntary context, social influences can influence intention indirectly through the mechanism of internalization and identification (Li 2010). Perceived system quality or output quality refers to individual perceptions of how well the system performs the tasks. Venkatesh and Davis (2000) suggest that output quality has a positive effect on perceived usefulness. Meanwhile, Venkatesh and Davis (2000) and Venkatesh and Bala (2008) argue that information of quality is an important factor for the intention to use information technology.

Research Model
This study intends to examine the determinants of intention to use the SISKEUDES by using modified TAM (see Figure 1). There are two main independent variables (i.e. human resource quality, social factor), and four mediator variables (i.e. perceived system quality, perceived information quality, perceived usefulness, and perceived ease of use). Perceived usefulness and perception ease of use are the main variables in the original model of TAM. Meanwhile, perceived system quality and perceived information quality are contextual variables that have been used by prior studies (i.e. Venkatesh and Davis 2000; Venkatesh and Bala 2008). The main difference of our model compared to others is the use of human resource quality as the main determinant. Human resource quality is a critical problem in village fund (Basri 2014). Thus, we argue that human resource quality is a root factor that might also influence other determinants. For example, low human resource quality might cause a negative perception on system quality, information quality, perceived usefulness and perceived ease of use. Consequently, it might cause a low intention to use the system.

Hypotheses Development
Based on the theoretical components of the TAM, we propose the following hypotheses regarding the intention to use the SISKEUDES.

Quality of Human Resource
Human resource is the basic capital that must be owned by the organization to perform activities to achieve organizational goals. Therefore, the ability of employees, as a human resource in an organization, is very important meaning and its existence is needed to increase work productivity in the organization. The quality of human resource is defined as a human resource that meets criteria of physical and health quality, intellectual quality (knowledge and skills), and spiritual mental qualities (Danim 1996). According to Matutina (2001), the quality of human resource refers to knowledge, skills and ability. Knowledge is the ability of employees, which are oriented to intelligence, and power of thought, and mastery of knowledge that is widely owned by employees. Skills are the ability and technical operational mastery in certain areas owned by employees. Ability is capabilities formed from several competencies possessed by an employee that includes loyalty, discipline, cooperation and responsibility. Therefore, the quality of human resource is people, who have physical and health quality, and have knowledge and skills and abilities. The level of user understanding of information technology will determine the success in utilizing the information system; otherwise, a low understanding of the user about the new system can lead to failure in using information technology. Setianingsih and Indriantoro (1998) find that increasing of user understanding of information systems will affect success in utilizing information technology. Hargo (2001) states that the level of understanding of information technology affects the implementation of information technology. Thus, we propose the following hypothesis: H1: The quality of human resource positively influences the intention to use the SISKEUDES. Thompson et al. (1991) state that social factor affects the use of information technology including number of colleagues, who use information technology in carrying out the task, senior managers or superiors, who help in introducing and in utilizing information technology, and the company itself that is very helpful in the use of information technology. Meanwhile, according to Davis et al. (1989), social factor is operated as the degree to which individual assumes that another important person convinces him to use or not to use a new information system. If an individual perceives that an important social actor can punish nonbehavior or reward behavior, the social influence of the compliance effect will occur. Venkatesh and Davis (2000), in TAM 2, argue that social factor was likely to have a significant influence on behavioral intention to be used in a mandatory environment.

Social Factor
The implementation of the SISKEUDES for management of village fund is the policy of the local government to overcome problems related to the amount of funds managed by the village and demands of transparency and accountability. The SISKEUDES application is used by village officials because leaders and organizations have instructed to use the system. Although they are not ready to use the application, since organizations and leaders have advocated their use, the village apparatus must use the system. Therefore, the hypothesis in this research is as follows: H2: Social factor positively influences the intention to use the SISKEUDES.

Perceived Information Quality
A good system, of course, must be supported with the quality of information. The quality of information is the output of information generated by the information system used (DeLone and McLean 1992;Rai et al. 2002). Quality of information refers to the output of the information system regarding the value, benefits, relevance, and urgency of the resulting information (Pitt et al. 1997). Liu and Arnett (2000) argue that quality information will improve the perception of user usability and will, of course, increase the intensity of the use of information systems.
Information quality is a crucial factor for intention to use information technology (Venkatesh and Davis 2000; Venkatesh and Bala 2008). However, in the context of the village fund, the quality of human resource in the village is relatively low and becomes a root of the problem. We argue that user perceptions on information quality are influenced by the quality of human resource. For example, the low quality of the village apparatus will result in a low perception of the information quality. Consequently, the village apparatus might be less interested in using the system. Therefore, the hypothesis in this research is as follows: H3: Perceived information quality mediates the relation between the quality of human resource and the intention to use the SISKEUDES.

Perceived System Quality
Quality of system is used to measure the quality of the information technology system itself. Quality of the system means the quality of a combination of hardware and software in an information system (DeLone and McLean 1992). The focus of system quality is on the performance of the system that refers to how well the hardware, software, policies, and information systems procedures provide information to users (DeLone and McLean 1992). The better quality of the system and the given quality of the system output, such as the speed of time to access and the usefulness of the system output, will cause the user to use, so that, the intensity of the system usage will increase.
Some researchers suggest that system quality can affect user satisfaction (DeLone and McLean 1992;Seddon and Kiew 1994;Iivari 2005;Almutairi and Subramanian 2005;Roldan and Leal 2003). In addition, the quality of the system also affects the use of the system (DeLone and McLean 1992; Wahyuni 2011). In contrast, Iivari (2005) and Roldan and Leal (2003) state that the quality of the system does not affect the user intensity.
System quality can also be measured based on overall system performance. For example, if there are many bugs in the system, users will tend not to use the system and the system cannot perform tasks according to the needs of its users. Understanding system user also influences the interest in using the system. The higher the level of understanding of users of a system will increase the desire to use the system. Therefore, the quality of human resource can influence the acceptance of the use of information systems through the quality of the system. For example, the low quality of the village apparatus might create a low perception of the system quality that causes a low intention to use the system. Based on the description above, it can be hypothesized as follows: H4: Perceived quality of system mediates the relation between the quality of human resource and the intention to use the SISKEUDES.

Perceived Usefulness
Perceived usefulness is defined as a measure using technology is believed to bring benefits to people who use it (Davis 1989), making the job easier, useful, increasing productivity, improving effectiveness, and developing job performance (Chin and Todd 1995). Venkatesh and Davis (2000) argue that perceived usefulness and perceived ease of use have a positive effect on the acceptance of the technology. When someone understands and benefits from using the information system, he will positively accept and want to use the technology.
The perceived usefulness will be influenced by the perception of ease of use of technology because an easier-to-use technology becomes more useful (Venkatesh 2000). However, in this research, two TAM variables, namely perceived usefulness and perceived ease of use, are not connected, because the focus of this research is to examine whether perceived usefulness variable and perceived ease of use mediate the relation of quality of human resource and interest in using information system. A consistent argument with the mediation role of system quality and information quality, we argue that village apparatus quality might affect perceived usefulness of the SISKEUDES. Therefore, the hypothesis in this study is as follows: H5: Perceived usefulness mediates the relation between the quality of human resource and the intention to use the SISKEUDES.

Perceived Ease of Use
Perceptions of ease of use of technology are defined as the degree to which a person believes that information technology can be easily understood and used (Davis 1989), as well as barrier-free . Perception of ease of use is believed to be someone if working with technology will be easier, when compared with someone who does work activities without using technology. The construct of perception of ease of use of technology influences perceptions of technological usability, attitude using technology, interest in using technology and actual use. The feelings of easy use of information technology will lead to a feeling that the system has its uses and, hence, sense a comfort when working with an information technology system (Venkatesh and Davis 2000). Some research results indicate that, when a person feels the ease of technology, the person is willing to use the technology (Davis 1989;Malhotra and Galleta 1999;Tangke 2004). However, consistent with the prior argument, in the context of the village fund, we argue that the quality of the village apparatus might influence perceived ease of use. Therefore, the hypothesis in this study can be formulated as follows: H6: Perceived ease of use mediates the relation between the quality of human resource and the intention to use the SISKEUDES.

RESEARCH METHOD
This study examines the influence of the quality of human resource on the intention of using the SISKEUDES by using the modified TAM consisting main variables (i.e. perception of usability and perception ease of use) and contextual variables, namely perceived system quality, perceived information quality, and social factor. Thus, this research is categorized as explanatory research, because it aims to test the relationship between variables through hypothesis testing.
This study uses primary data. Data were collected by using questionnaire method. The questionnaire used in this study was adopted from previous research (i.e. Davis et al. 1989;Iivari 2005;Venkatesh et al. 2003;Marwoto 2012). The variable of human resource quality in this research is defined as the level of understanding and ability of the village apparatus to manage and administer village fund using the SISKEUDES. The measurement of this variable was adapted from Marwoto (2012). The social factor is defined by how the influence of the important people in the organization environment both at the local government and village level affects users to use the SISKEUDES. The indicators used to measure social factor were adapted from Venkatesh et al. (2003). In terms of mediator variables, the perceived quality of information is defined by the perception of village apparatus on the usefulness of output or report generated by the SISKEUDES. Perceived system quality or output quality refers to the perception village apparatus about how well and reliable the quality of the SISKEUDES is used to manage and administer village fund. Indicators to measure the perceived quality information and the perceived system quality were adapted from Iivari (2005). Furthermore, the perceived usefulness is defined as a level or state where the village apparatus believes that using the SISKEUDES application will improve performance, both individual performance and organizational performance. Perception of ease of use is defined as a level or state where the village apparatus believes that using the SISKEUDES application does not require any effort (free of effort). Indicators to measure such variables were adapted from Davis (1989). In terms of dependent variable, the interest to use information system is interpreted by how much interest from village apparatus to use SISKEUDES in managing of village fund. We adapted instruments used by Davis et al. (1989) to measure this variable. Table 1 presents a summary of variable measurements.
Before being distributed to respondents, the questionnaires were examined in the pilot test. The purpose of a pilot test is to find out whether the statements that have been compiled in the questionnaire had really measured what had to be measured, and whether the statement would be consistent or stable over time (Ghozali and Latan 2014). Based on the results of the pilot test, all the questions were declared valid and reliable, thus it can be used in research. The detail of the questionnaire is in Table A1 and A2 in the Appendix.
The object of this study is the village in Ponorogo Regency because Ponorogo Regency was appointed by BPKP as a pilot project for implementation of the SISKEUDES in East Java Province in 2006. Ponorogo Regency was a pioneer of implementation of the SISKEUDES in East Java Province because the head of regency ordered all villages to implement the SISKEUDES and reject other system offered by private parties (Muhlis 2017). The population in this research were all operators of the SISKEUDES totalling 281 operators from 281 villages spread over 20 districts in Ponorogo Regency. The sampling method is the census sampling method. From April 11, 2017 to May 23, 2017, a total of 281 questionnaires were distributed to respondents (all members of the population). Of these, 221 questionnaires from 20 districts (78.65 per cent) were returned. Then, the returned questionnaires were screened and, finally, 212 complete data were obtained and can be processed for further analysis. In this research, data processing and analysis were carried out by using Structural Equation Modeling -Partial Least Square (SEM-PLS) with SMARTPLS 3.0 program.  Davis et al. (1989) Notes: BI = behavioral intention to use; QHR = quality of human resource; SF = social factor; SQ = perceived system quality; IQ = perceived information quality; PU = perceived usefulness; PEOU = perceived ease of use.

RESULTS AND DISCUSSION
The data analysis was conducted in several stages. In the early stage, we present demographic respondents that include gender, educational level, age and duration of computer use. These data were used in an additional test to examine the effect of moderation. There are two sub-models in the structural equation model. The inner model (structural model) specifies the relationship between independent and dependent latent variables, whereas the outer model (measurement model) specifies the relation between latent variables and observed indicators. In Smart PLS, it needs to assess the measurement models before evaluating the structural model. As this study employed reflective measurement model, the outer model was evaluated by the validity of convergent (outer loading, cross loading), discriminant validity (AVE, and communality), and reliability test (Cronbach's alpha and composite reliability). The inner model was evaluated by using R 2 and Q 2 . In the final stage, the hypotheses testing was done by using a t-test to examine the significance of the structural path parameter coefficients. Furthermore, testing the mediation effect used the bootstrapping method as suggested by Hair et al. (2016).

Demographic of Respondent
The results of data collection show that, of 281 distributed questionnaires, 221 questionnaires were returned. However, nine questionnaires were incomplete, so that, completed questionnaires that can be analyzed were 212 pieces. Thus, the response rate in this study was 75.55 per cent. Table 2

Convergent Validity
Convergent validity is the degree to which multiple items measuring the same concept. As suggested by Hair et al. (2016), factor loadings and average variance extracted (AVE) are used to assess convergence validity. In the first analysis, some indicators did not meet the criteria of convergent validity, because these indicators had values of factor loading below 0.5, namely PEOU1 and PEOU2. Thus, the PEOU1 and PEOU2 indicators were eliminated from the model. In the second analysis, all the indicators had values of factor loading above 0.5 (see Table  A3 in Appendix). The factor loading values have a range from 0.634 to 0.866, which exceed the recommended value of 0.5 (Hair et al. 2016). The average variance extracted is range from 0.530 to 0.660, which exceed the recommended value of 0.5 (Hair et al. 2016).

Discriminant Validity
The discriminant validity is the degree to which items differentiate among constructs or measure distinct concepts (Cheung and Lee 2010). The results indicate that all crossloadings are higher than the correlating values in the row and the column indicating adequate discriminant validity (see Table A4 in Appendix).

Reliability Analysis
The reliability is a test of how consistently an instrument measures a concept (Sekaran and Bougie 2013). The reliability testing was conducted by looking at the value of composite reliability and Cronbach's alpha. The recommended composite reliability and Cronbach's alpha score should be greater than 0.7, although a score of 0.6 is still acceptable for exploratory study (Hair et al. 2016). As can be seen in Table A3 in the Appendix, the composite reliability and Cronbach's alpha scores for all constructs are above the acceptable level of 0.6.

Structural Model (Inner Model)
We employed in a step-by-step analysis of the structural model to provide a detailed description of our results and to comprehensibly test H1 to H6. As suggested by Hair et al. (2016), mediation analysis was carried out using the bootstrapping method. The advantage of this method is that it does not rely on the assumption of normality, thus, it is also fit for small sample sizes (Hair et al. 2016). In this approach, bootstrapping must be done twice. In the first step, bootstrapping was done without the presence of mediation (direct path). In this step, we focused on the effect of the quality of human resource and social factor on behavior intention (H1 and H2). Subsequently, in step 2, we introduced the mediators (H3 to H6). Table A5 presents the results of step 1 that is the results of structural model estimation and evaluation of the relationship between the quality of human resource, social factor and behavior intention (H1 and H2). The central criterion for the structural model assessment, namely the coefficient of determination R 2 (Hair et al. 2016). The R-square parameter (R²) is used to measure the variation level of the independent variable changes to the dependent variable and the path coefficients for the significance test between the constructs in the structural model shown by the t-statistics value. The higher of the R² means the better the predicted model. As can be seen in Table A5, the behavioral intention has a high value of 0.758, which means that 75.8% change variation of the behavioral intention to use can be explained by the construct of quality of human resource and social factor. This value is included in the strong category. The Q 2 value of predictive relevance also supports this finding. After running the blindfolding procedure, we obtained the Q 2 value of behavioral intention (0.421), which is well above zero, indicating the predictive relevance of the PLS path model. Furthermore, in terms of the fit model, the value of SRMR is 0.046, which is below than 0.05. It means that the model is fit. In step 1, the direct effect of human resource quality on behavior intention has a negative and significant (p < 0.01) value of 0.649 (see Table A5 in Appendix), rejecting H1. In contrast, the effect of the social factor on behavior intention has a positive and significant (p < 0.01) value of 0.330. Thus, H2 has been empirically substantiated. The significant of the relationship between human resource quality and behavior intention provide evidence that we could move to step 2 by adding the mediation variables on the structural model.
In step 2, we run PLS-SEM analysis by adding all the mediator variables (perceived system quality, perceived information quality, perceived ease of use, and perceived usefulness) on the structural model. Table A5 in the Appendix shows the result of the model assessment with all mediator variables. As can be seen that, except perceived quality information, all mediators have relatively high of R 2 indicating predictive power of the model. These results are also supported by the values of Q 2 that are well above zero. Figure 2 shows estimates of the path PLS model.
After including the mediators construct, we found that human resource quality has a high and significant effect on all the mediators construct, which in turn has a strong and significant relationship with behavioral intention. The indirect effects of human resource quality via all mediators are significant (see Table 3). At the same time, the relationship between human resource quality and behavior intention remains significant (Figure 2; path coefficient of -0.493, p < 0.01), but has a negative sign and is significantly lower than when the mediator constructs are not present. Hence, the level of mediation is determined based on the variance account for (VAF). The Variance Account For (VAF) is calculated by the indirect effect/total effect. The total effect is the direct influence plus indirect influence. The VAF values above 80% are categorized as full mediation, VAF values between 20%-80% are categorized as partial mediation, and VAF values less than 20% are concluded that there is almost no mediation effect (Hair et al. 2016).
Based on Table 3, the VAF values of perceived system quality, perceived usefulness, perceived ease of use are 28%, 21%, and 62% respectively. It means that those variables partially mediate the relation of quality of human resource on interest in using of the SISKEUDES, confirming H4, H5, and H6. In contrast, in terms of the perceived information quality variable, it can be seen that VAF is equal to -7 percent. It indicates that perceived quality of information does not mediate the relation of the quality of human resource to interest in using of the SISKEUDES, rejecting H3.

Figure 2 Structural Model with Mediators
Notes: QHR = quality of human resource; SF = social factor; SQ = perceived system quality; IQ = perceived information quality; PU = perceived usefulness; PEOU = perceived ease of use; BI = behavioral intention to use.

Table 3 Summary of Results
Notes: QHR = quality of human resource; SF = social factor; SQ = perceived system quality; IQ = perceived information quality; PU = perceived usefulness; PEOU = perceived ease of use; BI = behavioral intention to use.

Discussion
The results of the hypotheses testing are summarized in Table 4. According to Table 3, unexpected, the quality of human resource has a negative and significant effect on the interest in using the SISKEUDES, not confirming H1. The negative path coefficient indicates that the higher quality of human resource, the lower interest in using the SISKEUDES. A possible explanation is the skeptical thinking of the village apparatus. The higher the quality of human resources, they have wider insight, are more creative, and can think more critically in making decisions. Thus, the low interest in using the SISKEUDES is probably caused by the idea that the SISKEUDES will be the same as other information systems, such as the SIMDA that has been used by local governments in financial management. The SIMDA, which is also developed by BPKP, has some technical and non-technical obstacles. Based on this experience, they think that SISKEUDES will experience the same thing. Therefore, they are skeptical and less enthusiastic in using the SISKEUDES. In contrast, social factor has a positive and significant effect on interest in the use of the SISKEUDES. This result supports H2 that social factor influences the interest in using the SISKEUDES. The positive path coefficient means that the higher the influence of social factor, the higher interest in using of the SISKEUDES. The social factor is indicated by the enormous support of colleagues, senior managers, leaders and organizations. In terms of the influence of social factor, the village apparatus is encouraged to master and to use the SISKEUDES. The local government expects that village fund management can be implemented in a transparent and accountable manner, therefore the preparation of village financial statements can be submitted on time. This result is consistent with prior studies (i.e. Thompson et al. 1991;Venkatesh et al. 2003) that social factor has a positive effect on the utilization of information system in a mandatory environment. Based on Table 4, the VAF of perceived information quality is -7 per cent, indicating that there is no mediation effect of perceived information quality on the relationship between the quality of human resource and interest in using of the SISKEUDES. This result does not support H3, indicating that human resource quality is still the main determinant of using the SISEKUDES. Interestingly, the path coefficient of perceived information quality and intention in using the SISKEUDES is negative. Based on the result of the interview with some operators of the SISKEUDES, reports generated from the SISKEUDES are not in accordance with what they expect. For example, the auxiliary activity reports are not shown in detail for each subactivity. This result indicates that operators might be confused in understanding the output generated by the application. As a result, operators might have a negative perception of the information quality of the SISKEUDES.
The VAF of the perceived system quality is 28 per cent. It means that perceived system quality partially mediates the relation of quality of human resource to interest in using of the SISKEUDES, supporting H4. System quality refers to the characteristics of the inherent information about the system itself. Davis et al. (1989) and Chin and Todd (1995) define perceived system quality as a convenience perception that measures the level of ease of computer technology to be understood and used. The SISKEUDES was developed in 2015 by BPKP to improve the quality of village financial governance. This application is made simple and user-friendly, making it easier for users to operate it. This village financial application uses the Microsoft Access database, so it is more portable and easier to apply by even ordinary users. The operators of SISKEUDES might feel that the system has good system quality characterized by secure data, easy system usage, fast and reliable access. Therefore, the quality of a good system, of course, affects user to use the system.
The VAF of the perceived usefulness is 21 per cent, which means that perceived usefulness mediates the relation between the quality of human resource to interest in using of the SISKEUDES. Therefore, H5 is supported. Perceived usefulness is defined as the subjective capabilities of users in the future by using specific application systems and can improve performance in an organizational context ). Perceived usefulness serves as a basis that can increase one's belief that information systems are useful in organizational activities, so they will use the information systems. The greater the perceived benefits of using an information system will increase interest in using the system. In this study, the greater the perception of the users on the usefulness of the SISKEUDES will increase the desire of the user to use it in the context of village fund management.
The VAF of the perception ease of use is 62 per cent, which means that perceived ease of use partially mediates the relation of quality of human resource to interest in using of the SISKEUDES. It supports H6. Perception of ease of use is defined as a level or state in which a person believes that using a system does not require strenuous effort (Davis 1989). The positive sign of perception of ease of use might be caused by the availability of manual operation of the SISKEUDES, so as it facilitates the user in operationalizing the system. With the perception that using a system does not require hard effort from users, it will increase the desire to use the system.

CONCLUSION
This study provides empirical evidence that human resource quality negatively influences the intention in using the SISKEUDES. On the other hand, social factor is positively associated with the intention of using the SISKEUDES. The results analysis of mediation variables indicate that perceived system quality, perceived usefulness and perceived ease of use partially mediate the relation of human resource quality and the intention in using the SISKEUDES. In contrast, perceived information quality is not able to mediate the relationship between the quality of human resource and the intention of using the SISKEUDES. It implies that, compared to other factors, human resource quality remains the main determinant of the intention to use the SISKEUDES. In addition, the attributes of the system (quality, usability and ease of use) are also determinants of interest in using the system. Therefore, increasing the capacity of human resources must be accompanied by improvements in system quality, system usability, and ease of use of the system.
Based on the findings, it is reasonable to provide a suggestion for Ponorogo Regency and BPKP. For the successful implementation of the SISKEUDES, Ponorogo Regency and BPKP should continually improve the capacity of operators by providing a lot of technical guidance and training. However, the implementation of technical guidance and training should also be accompanied by improvements to the weaknesses that still exist in the SISKEUDES. Based on the results of this research, high-quality operator resources coupled with a good perception of system quality, usefulness and ease of use will improve the intention to use the system. Therefore, BPKP, as the developer of the system, should be willing to receive feedback from operators related problems in the field. This is to avoid skeptical thinking from the operators who follow the training.
This study has limitations that must be considered. First, this research is only conducted in one regency, so it is necessary to be careful in generalizing the results of research because each region has distinctive characteristics and different human resource quality. Second, this study only focuses on the determinants of using the SISKEUDES without testing whether the implementation of the system produces the expected outcomes (e.g. transparency and accountability). Based on such limitations, suggestions that can be proposed for future research are as follows: (1) to improve the generalizability of the result, future study might replicate this study on other regencies; (2) further research needs to evaluate the issue of success of the SISKEUDES by examining impact of the SISKEUDES on the village government financial quality.  Intention to use Intention to use BI1 Davis et al. (1989) 5 Likert scale Intention to use in future BI2 Recommend to others BI3 Notes: BI = behavioral intention to use; QHR = quality of human resource; SF = social factor; SQ = perceived system quality; IQ = perceived information quality; PU = perceived usefulness; PEOU = perceived ease of use. 0.688 Notes: BI = behavioral intention to use; QHR = quality of human resource; SF = social factor; SQ = perceived system quality; IQ = perceived information quality; PU = perceived usefulness; PEOU = perceived ease of use.   : QHR = quality of human resource; SF = social factor; SQ = perceived system quality; IQ = perceived information quality; PU = perceived usefulness; PEOU = perceived ease of use; BI = behavioral intention to use