Factors Determining Intention to Use Banking Technology in Indonesian Islamic Microfinance

Abstract

This study aims to determine the perceptions of Islamic Financial Cooperative (BMT) managers in the Indonesian BMT Association on the acceptance of core banking technology. The Technology Acceptance Model (TAM) is used because it has simple theoretical characteristics (parsimony) and is supported by data (verifiability). This study develops the TAM model by integrating new variables -perceptions of maqashid sharia, perceptions of economies of scale, perceptions of market structure, and perceptions of technology procurement costs. These new variables are used to measure intention in using technology and actual usage in BMT operations. This study used PLS-SEM with smartPLS 3. The study was conducted in Central Java in six ex-Residency at 35 BMT with 300 respondents consisting of six levels of position level. The research found that maqashid sharia and market structure directly influenced the intention of BMT managers in using core banking technology. This new finding strengthens a theoretical model regarding the role of maqashid sharia in the acceptance of information technology in BMT. In addition, the perception of economies of scale has no significant effect on intention in using technology or its actual usage. The perception variable of technology procurement costs was found to have no significant effect on intention in using technology.

1. Introduction

Information technology is one of the innovations in the modern era that can be used to help various aspects of human life. Huda (2016) states that information technology has various advantages if it can be used properly, such as it can provide added value, help solve problems, and so forth. The use of information systems in a company can provide a competitive advantage and make it able to compete in the marketplace. One example of the use of information technology in financial institutions is the use of fintech. Lien et al. (2020) examines fintech in Vietnam, which is a developing country with a relatively new fintech service. The research results show that fintech services are very important for the banking sector in Vietnam. Research also confirms the factors that influence customer intention to use fintech services. In line with that, the intention to use fintech services (INT) is positively influenced by perceived of usefulness (PU), social impact (SI), customer trust (TRU), and perceived ease-of-use (PEU). Based on the results of this study, bank managers have a basis for improving the quality of fintech services.

Thus, the role and function of information systems for companies will become more extensive. Baitul Maal wat Tamwil (BMT) has started implementing information technology system to facilitate transactions with its members. Analysis of the effects of utilization, user expertise, effectiveness of use, and trust in information technology systems on individual performance in the company also needs to be conducted. Information technology has a very big role in the development of BMT in Indonesia, especially in supporting financial transactions for members and providing efficiency in the time and costs spent in BMT operational activities.

Many studies have examined the factors that influence the use of technology in the financial industry by using the Technology Acceptance Model (TAM) approach. Lee (2009) found that the factors of convenience, effectiveness, and flexibility play an important role in influencing the use of Internet banking technology. The results of research by Lee, Becker, & Potluri (2018) in Korea suggest that company needs, social expectations, ease-of-use, and usefulness should be seen as important antecedents explaining corporate behavioral intentions to use social media. This study also found that the ease-of-use and usefulness of the technology acceptance model had a positive directional mediating effect in influencing social media acceptance. This is also supported by Tahar, Riyadh, Sofyani, and Purnomo (2020), that perceived convenience and security perceptions positively affect the intention to use e-filing. Tahar’s research was conducted in the city of Semarang, Indonesia, to determine the effectiveness of the use of e-filing provided by the Directorate General of Taxes for civil servants, military and state police.

Likewise, Rahi (2017) confirms that perceived benefits, perceived ease-of-use and attitudes are the main constructs for promoting Internet banking in Pakistan. Several studies on TAM have been developed by adding new variables and integrated with the basic concept of TAM, namely, researching the intention in the use of technology (Hassan et al., 2018; Kumar & Shenbagaraman, 2017; Al Somali, Gholami, & Clegg, 2009; Rifai, Asakdiyah, & Setyawan, 2019; Saji & Paul, 2018; Bashir & Madhavaiah, 2015). The disadvantage of the limited research conducted on TAM is that there are no studies that use culture and religion variables as driving factors for intention in the use of technology. Therefore, this study adds a variable that describes a person’s perceptions based on their religious understanding, namely, perception of maqashid sharia. This is in line with the research by Phase (2005), Rahmawati (2012) and Muafi (2003), that spiritual motivation plays a major role in one’s behavior in selecting technology and performance levels.

2. Literature Review

2.1. Theoretical background

Previous research on technology adoption in the financial industry sought to understand how different factors influence individual technology adoption behavior. The TAM has gone through three phases of development: adoption, validation, and extension. TAM was developed by Davis (1986) using the basic Theory of Reasoned Action (TRA). TRA was first formulated by Ajzen (1975) who stated that intention determines whether a person does or does not carry out an activity. Sheppard et al. (1988) stated that TRA is widely used as a predictor of behavior in various ways. TRA is used by Davis as his grand theory, although not all TRA components are accommodated in his research. The component of TRA used by Davis is only the ‘attitude’ component.

2.1.1. Perceptions of Maqasid Sharia

There are three gradations in the distribution of maqashid sharia proposed by Syathibi, namely, dharuriyyat (primary), hajjiyyat (secondary), and tahsiniyyat (tertiary). Dharuriyyat includes the maintenance of basic needs in human life. These basic needs include the need for religion (al-din), soul (al-nafs), descent (an-nasl), property (al-maal), and reason (al-aql). In the concept of maqashid sharia, the use of technology in financial institutions is a facility (al-wasail) that can be used to protect assets (hidz al-maal) by customers.

Rizky (2017) explains that maqashid sharia is the main support in every operational and product development in Islamic banks. Therefore, understanding the importance of maqashid sharia will affect the perceived ease of operation of Islamic financial institutions. Likewise, Firdaus (2018) shows that the perception of maqashid sharia affects patient preferences and perceptions of convenience towards a hospital service. According to Jamal (2016), the understanding of maqashid sharia allows ijtihad to be developed, especially in dealing with various new problems that are not mentioned by the texts. Islamic law will remain dynamic in responding to various social phenomena that are constantly changing and developing. Therefore, this study will develop the perception of BMT employees about maqashid sharia that will affect their intention in the use of technology. This is in line with Nurfalah and Rusdiana (2019) that the ease of technology that was developed allows people to easily access Islamic financial services and products wherever and whenever they need it, according to the concept of maqashid sharia.

2.1.2. Perceptions of Economies of Scale

Based on some previous studies, it was found that technology is an important factor for increasing company productivity. Several studies in various business sectors show that technology supports the strengthening of a company’s economies of scale. Baumers, Dickens, Tuck, and Hague, (2016) found that the use of technology has an impact on increasing manufacturing production. It was found that, overall, the estimated technology parameters were in line with expectations. Meanwhile, Aivazian (2013) explains that economies of scale are influenced by technological changes as the main determinant, especially in driving productivity growth. It found that technological change explained productivity growth with certainty and even more so in economies of scale. The growth of the economies of scale in the company, which is supported by the use of technology, is also found by Mitchell (2002). He found that the role of technology in increasing the rate of plant growth processes greatly, so that companies become more efficient. Bilgihan (2011) found that developing and implementing IT projects can lead to competitive advantages in hotel companies. He showed that IT capabilities and IT competencies can help hotel companies achieve a competitive advantage. Likewise, digitization or utilization of Information Technology in BMT operations is becoming an important tool to achieve performance efficiency.

2.1.3. Perceptions of Market Structure

The implementation of the use of technology can be in the form of adding BMT service product facilities, where it is in line with a market structure characterized by monopolism with the same product but has different services offered. In its operations, BMT requires innovation and creativity in the service products offered. The use of technology will encourage the emergence of new services based on accuracy and speed. Thus, technology will make the BMT market even wider in reach. Several studies have confirmed that technology can increase market coverage in the industry.

Hannan and Mcdowell (2014) show that banks have successfully used ATM (Automatic Teller Machine) to attract customers from competitors. Consequently, the impact of technology adoption on market structure depends on the size of the company adopting new technology. Karaca Mandic et al. (2017) concluded that medical technology diffuses faster in more competitive areas. Furthermore, they found that a more competitive physician market structure could promote faster diffusion by increasing patient access to the latest medical technology. Sutton (2001) found that technological change determines market structure. The findings by Weeds (2012) also support other research that technological changes have changed the creative media industry. The results of his research found that supply-side factors can produce advantages and long-term sustainability, where the use of technology can reduce fixed costs for basic products or decrease market size. Likewise, Elberfeld and Gotz (2002) found that technology is an endogenous variable in the relationship between market size and market structure. Their research yields a simple and fairly intuitive explanation for the phenomenon of industrial shocks. The emerging market is bigger due to the large-scale adoption of technology. They found that technology switching always leads to a reduction in the number of active companies.

2.1.4. Perception of Technology Procurement Costs

Flor and Hansen (2013) found that technological advances have an impact on future income, the results of the study show an important long-term effect, that is, the investment threshold increases while the value of options decreases. The application of investing in technology applies to long-term profitability calculations. This is consistent with the results of research by Bessen (2002), that in the manufacturing industry in the United States, technology investment causes productivity to grow around 0.4% per year faster. Technology procurement costs are included in the long-term investment category. Bessen (2002) also found evidence linking increased productivity with greater investment in IT. Such dramatic changes suggest that companies experience large adoption costs when switching to new technologies so that it will have long-term implications for higher productivity.

Technology procurement is a long-term management decision. Thatcher and Oliver (2001) examined the impact of technology investment on various performance measures. They found that the impact of technology investment increases efficiency. The results show that investment in technology encourages certain efficiency improvements that can maximize profit. More specifically, research shows that investing in technology also reduces a company’s fixed overhead costs by not affecting product quality and pricing decisions, even increasing profits and increasing productivity. In addition, research also shows that investment in technology can reduce costs for designing, developing, and manufacturing products which encourages company to improve product quality and charge higher prices.

An efficient company always tries to suppress various investments that absorb large funds. The financial industry will avoid procurement for additional unproductive assets. Every money owned by the company must be able to become productive assets in the form of financing. Therefore, the procurement of fixed assets, both tangible and intangible, will always be considered carefully.

Likewise, the procurement of investment for technology is currently still in the expensive category, so BMT will be careful in carrying out the technology implementation process. Akbar and Noviani (2019) found that the procurement of information technology requires a sufficiently large budget or funds.

2.1.5. Behavioral Intention

Behavioral intention has been observed as an indicator of system success (Venkatesh, Morris, Davis, & Davis, 2003). Intention is defined as the desire to perform behavior. Intention can change over time. Behavioral intention according to Fishbein and Ajzen and many other researchers is a strong predictor of how someone will behave in certain situations. Intention is defined as a condition where someone tends to choose to do or not to do a job. One of the motivating factors in individuals that influence behavior is Intention. Thus, Ajzen (1991) concluded that intention can be reflected in how much someone wants to try and the level of motivation to realize certain behaviors.

The results of previous studies indicate that behavioral intention is the best predictor of technology use by system users. Behavioral intention to use is a behavioral tendency to keep using a technology. The level of use of a computer technology in a person can be predicted from the attitude of his attention to the technology, such as the desire to add supporting peripherals, motivation to continue using it, and the desire to motivate other users. The results of the research prove that usefulness and attitude toward using affect a person’s tendency to behave.

2.1.6. Actual System Usage

The actual system usage is the real situation of a person using a technological system in his life (Davis, 1989). Actual system usage is the same as behavior in TRA. However, it is used in a technological context. This construct is a variable that is directly influenced by intention and perceived ease of use. The actual system usage construct can be measured by 3 (three) indicators, namely real use, actual frequency and user satisfaction (Wibowo, 2006). When technology users have the perception that technology systems are easy to use and will increase their work productivity, they will feel satisfied and will prove the use of technology in real conditions (Tangke, 2004).

Figure 1: Theoretical Framework

2.2. Hypotheses

This study develops the TAM variable and builds the following hypotheses:

H1: Perception of Maqashid sharia has a positive effect on intention in using technology.

H2: Perception of market structure has a positive effect on intention in using technology.

H3: Perception of economies of scale has a positive effect on intention in using technology.

H4: Perception of technology procurement costs has a negative effect on intention in using technology.

H5: Intention in using technology has a positive and significant effect on actual system usage.

3. Methods

3.1. Survey Design and Sampling

This research used a quantitative method to determine the effect and relationship between the independent variable on the dependent variable. To prove the correlation and causal relationship, researchers used quantitative research method of causal research. According to Istijanto (2009), in causal research, researchers will try and aim to determine the variables that influence or cause changes in other variables. Meanwhile, according to Rangkuti (2005), causal research aims to determine the relationship or involvement between these variables. Thus, a quantitative approach is used to verify the effect of the technology acceptance model on the intention to adopt core banking. The questionnaire was developed to measure respondents’ observations. The survey was distributed to BMT employees in the Central Java, Indonesia. The sample consisted of 35 BMT and the respondents consisted of six levels of positions at BMT: (a) Manager/Director, (b) Branch Head, (c) Head of Division, (d) Human Resource & Development, (e) Marketing/ Funding /AO, and (f) Teller/CS (Customer Service).

Data were collected from May to June 2020 from 22 districts in Central Java, Indonesia with cluster random sampling. The cluster random sampling method divides the population into several separate groups known as clusters and a number of randomly selected samples are taken. Cluster random sampling is a sampling technique that is applied when a population is found in groups that appear uniform but internally remain different. Cluster random sampling is a sampling technique that is widely used in geographic statistical groups. This study used a sample size of 300.

3.2. Instrument Development

The survey instrument consists of two parts. The first part asks about demographic characteristics such as region, age, gender, qualifications, while the second part contains six latent constructs, namely, perceptions of maqasid sharia, perceptions of economies of scale, perceptions of market structures, perceptions of technology procurement costs, intention in using technology, and actual system usage. All construction items were adopted from previous research work. The items of maqasid sharia was adapted from Noordin (2019), economies of scale adapted from Priyonggo (2019), market structure adapted from Kidwell and Peterson (1981), technology procurement costs adapted from Akbar and Noviani (2019), intention and actual usage adapted from Davis (1989). All items were given a choice of a 7-point Likert scale (1 for strongly disagrees and 7 for strongly agree).

3.3. Respondent’s Profile

Demographics of 300 respondents are tabulated in Table 1. There were more males (61.3%) than females (38.7%). The 25.7% of respondents were less than 30 years old, 43% were counted between the ages of 30 to 40 years, 27.7% for 40 to 50 years and around 3.7% of respondents were over 50 years old. The data were further analyzed by demographic characteristics such as level of education. Most respondents’ level of education was undergraduate degree (60.3%), which means that the respondents were educated, for graduate (4.7%) and Senior High School (20.3%). The composition of respondents based on position was Teller/CS (42 people), Marketing (41 people), HR&D (34 people), Head of Division (37 people), Branch Head (111 people), and Manager (35 people). The majority of respondents in this study came at the level of the branch head.

Table 1: Demographic profile of the respondents

4. Data Analysis Technique

The model was analyzed by using the Partial Least Square (PLS) analysis technique with SmartPLS 3.2.8 (Starstedt, 2017). The analytical procedure was carried out in two stages, the researcher tested the measurement model (validity and reliability) and the structural model (Hypothesis testing) recommended by Hair Jr. et al. (2014).

4.1. Measurement Model

Research must begin by evaluating the latent construct measurement model for its validity and reliability by going through a process of confirmatory factor analysis (CFA). Because this study is quantitative, it usually used Cronbach (α) to ensure reliability. However, Composite Reliability (CR) is recommended more (Henseler et al., 2009). Two types of validity used are convergent and discriminative validity.

4.2. Convergent Validity

The convergent validity of the measurement model is usually ascertained by examining the loading factor, extracted mean variance, and reliability (Hair et al., 2010). Figure 2 illustrates the result where the loading factor value is supported by Chin (1998), the recommended threshold level is 0.6. All values above 0.6 indicate convergent validity. Likewise, all composite reliability values were above 0.7 (Sarstedt, 2019).

Figure 2: Measurement Model

Convergent validity was confirmed by the Average Variance Extracted (AVE) estimation with the value recommended by Fornell and Larcker (1981), which is greater than 0.5. Average Variance Extracted which reflects the total variance in the indicator is accounted for the latent construct.

Furthermore, this study conducted measurements by assessing composite reliable. Table 2 illustrates the degree of composite reliability (CR) where the value must exceed 0.7 recommended by Hair et al. (2010). The result of calculating all CR values above 0.7 indicates that all constructs are reliable.

Table 2: Results of Measurement Model

4.3. Discriminant Validity

Discriminant validity is to compare each square root of AVE to the correlation value between constructs. The calculation is carried out by taking into account the value in the test of Fornell-Larcker Criterion (1981) and the crossloading value. The variable is said to be fit if the square root value of AVE is higher than the correlation between constructs. Table 3 shows that the square root of the AVE as shown in the bolded value is greater than the corresponding row and column value, it means the measure is discriminant.

Table 3: Discriminant validity of Measurement Model

Note: Bolded values indicate the square root of AVE of each construct.

The next step is to examine the cross-loading value, which is to evaluate the correlation value of variables with indicators that are greater than the correlation among other constructs (Hair Jr. et al., 2016). Table 4 illustrates that the correlation value of each variable on the indicator is above 0.7 and all of them are greater than the correlation among other constructs. Based on the cross-loading test above, it can be concluded that the constructs built in this research model are valid.

Table 4: Loading and Cross Loadings

4.3.1. Heterotrait-Monotrait Ratio (HTMT)

Another approach to assessing discriminant validity suggested by Henseler et al. (2015) is by using multi-trait and multi-method matrices, namely, Heterotrait-Monotrait Ratio (HTMT). The recommended HTMT value of smartPLS should be less than 0.9 (Gold & Malhotta, 2001). From Table 5, it can be seen that all HTMT values are less than 0.9, it means, the model is valid.

Table 5: Heterotrait-Monotrait Ratio (HTMT)

Note: Heterotrait-Monotrait Ratio (HTMT) discriminant at (HTMT <0.9).

4.4. Structural Model Evaluation

Structural equation modeling (SEM) was conducted to assess the strength of the proposed model for this study. To assess the lateral collinearity test of the structural model (VIF), R2 value, and t-value, it was adopted from Hair Jr. et al. (2016). The proposed hypothesis was tested by running a bootstrap procedure with resample 5000, as suggested by Hair Jr. et al. (2014).

4.5. Collinearity Statistics

In the first step of structural equation modeling, lateral collinearity is assessed by the VIF collinearity statistic. Collinearity statistics is known by looking at the value of the Variance Inflation Factor (VIF) and free from multi-collinearity phenomena, the VIF value must be below 5. This collinearity occurs when the two variables which are hypothesized to have a causal relationship measure the same construct. This type of collinearity is assessed by the VIF value, If the VIF value is 3.3 or higher, it indicates the potential for collinearity (Diamantopoulos & Siguaw, 2006). Table 6 shows the Inner VIF value is below 5, which means the model is fit (Diamantopoulos & Siguaw, 2006).

As described in Table 6, the VIF values are less than 5 and 3.3, indicating that there is no lateral multi-collinearity in this study based on the statement of Hair Jr. et al. (2014).

Table 6: Results of Collinearity Statistics

4.6. Hypotheses Testing

The hypothesis development in this study was tested by running a bootstrap procedure with resample 5000, as suggested by Hair Jr. et al. (2014). Table 7 describes the path coefficient of each construct with its level of significance.

The results showed that the three hypotheses had a significant relationship with each endogenous variable. H2 is not significant in the relationship between economies of scale and technology procurement cost. Table 7 illustrates that the relationship between maqashid sharia and intention in using technology as stated by H1: (t-statistic = 5.346>1.96 and p <0.05). Furthermore, the relationship between perceived market structure and intention in using technology is supported as stated by H3: (t-statistic = 3.122> 1.96 and p <0.05). H5 shows that intention in using technology is positively related to actual system usage (t-statistic =10.656>  1.96 and p <0.05.

Table 7: Results of Structural Model Analysis (Hypothesis Testing)

Note: Significance level where, *p < 0.05.

4.7. Evaluating Effect Size

R^₂  values for interest in using technology and actual system usage are 0.483 and 0.293, both are accepted based on the suggestion of Cohen (1988). We also assessed the effect size (f ²). As stated by Cohen (1988), the P value can indicate an influence but does not reveal the magnitude of the influence. In Table 8, the effect size (f ²) can be seen where H5 describes a large effect size, while H1 and H3 have a medium effect size. H2 and H4 present small effect sizes as suggested by Cohen (1988). Furthermore, the relevance of the model predictions was assessed by using a blindfolding procedure. The blindfolding procedure should only be applied to endogenous constructs that have reflective measurements (Hair Jr et al., 2016). If the Q² value is greater than 0, it indicates that the model has predictive relevance for certain endogenous constructs (Cohen, 1988; Hair Jr. et al., 2016). Table 8 shows that the value of Q² is greater than 0, meaning that the proposed model has significant predictive relevance.

Table 8: Evaluating Effect Size

Note: f² 0.02, small; 0.15, medium; 0.35, large.

5. Discussion and Conclusion

This study investigates the relationship between the proposed variables and examines their effect on the intention of BMT employees to use core banking. The character of BMT as a microfinance institution has similarities with financial institutions in general. Furst et al. (2002) stated that the profitability of financial institutions that use Internet banking is higher than non-Internet banking users. Furthermore, Internet banking creates new marketing opportunities and increases customer loyalty (Rahi, 2016). Thus, the use of technology on BMT will have the same impact as the impact obtained by other financial institutions. BMTs in Indonesia need to know the factors that influence the adoption of core banking. The model built on the research guides finding the core factors of technology acceptance from BMT employees. Intention in core banking is measured by the perception of maqasid sharia, perceptions of economies of scale, perceptions of market structures and perceptions of costs of technology procurement. This concluded that maqashid sharia and market structure have a positive and significant relationship with the behavior intention in using core banking technology in BMTs. Meanwhile, economies of scale and technology procurement costs do not have a significant influence on the intention of BMT managers to use core banking technology.