The effect of religiosity and service quality on customer loyalty of Islamic banks mediated by customer trust and satisfaction Soegeng Wahyoedia*, Achmad Sudiroa, S. Sunaryoa and S. Sudjatnoa

Article history: Received: June 27, 2020 Received in revised format: August 1


Introduction
As a country with the largest Muslim population in the world, Indonesia's contribution to global Islamic banking assets is still low; merely at 1.9 percent. Meanwhile, the potential of Islamic banking market in Indonesia is immense considering that the number of Muslim populations in Indonesia has reached around 229.62 million people (Databoks, 2019). This huge potential has not been handled well as evidenced by the low number of citizens who use Islamic banks. In 2019, there were 29.21 million Islamic bank customers, or around 10.8% of Indonesia's total population of 269 million, and around 12.7% of the total Muslim population in Indonesia. The data indicate a gap between the number of Islamic bank customers and the potential of Islamic bank customers. Meanwhile, in terms of total assets, the contribution of Islamic banking assets has only reached about 4.3% (OJK, 2020) of the total national banking assets. This shows that there are still a great number of Indonesian citizens who are not banking with Islamic banks, and this indicates the problem of loyalty to Islamic banks in Indonesia. The low quantity of Islamic bank customers has an impact on the low contribution of Islamic banks in the national banking system as shown in Table 1. Loyalty is the intention or tendency of customers to conduct repurchase from the same company in the future (Amin et al., 2013). Customer loyalty refers to the commitment of a customer to remain faithful to use the services of a bank within a certain period (Amin et al., 2013: 82). A loyal customer will (a) repeat, i.e. repurchase or increase deposits in the future (share of wallet); (b) retention, i.e. not affected by goods or services offered by other banks; and (c) referral, i.e. inform positive things about the bank to other parties (Parasuraman et al., 1994;Baumann et al., 2011;Mahardika, 2018). As a bank that is run with Islamic principles, namely Islamic law, loyalty to Islamic banks should not have been debated. However, the debate still rises related to one's faith to carry out their worship in accordance with the Quran and Hadith. In this case, religiosity is needed in one's faith in carrying out their worship; therefore, the variable of religiosity became the first variable determining of loyalty. Loyalty to Islamic banks is inseparable from the role of religion, in this case the religion of Islam. In addition to religiosity, the determining factor for customer loyalty of Islamic banks is service quality. Ireland (2018) found that in the United Arab Emirates and Malaysia, many Islamic bank customers also use conventional banks. The use of dual banking systems is due to the inferiority of Islamic banks in the service quality provision compared to conventional banks. With the mastery of information technology from dominant and modern conventional banks, their service products are more attractive than those of Islamic banks. The effect of service quality on loyalty was conveyed by Awan and Bukhari (2011) in their research in Pakistan stating that service quality was more dominant in its effect on loyalty compared to religiosity. Meanwhile, Echchabi and Olaniyi (2012) conveyed their findings that Islamic bank customers in Malaysia who enjoyed good service quality became loyal to Islamic banks. Different dimensions/indicators were used to measure the service quality.
Although the use of the Servqual model has dominated research in the area of service quality, Sunaryo (2013) believes that to measure service quality in Islamic banks, the Carter model is more appropriate (compliance, assurance, reliability, tangibles, empathy, and responsiveness). Ali (2018) also used Carter's model in examining the quality of Islamic banking services in Brunei Darussalam. The use of Carter's model is inseparable from the argument that the Servqual model does not pay attention to cultural elements, especially eastern culture which is different from the western where the Servqual model was developed.
Alnasera et al. (2018) argued that service quality for countries that implement Sharia principles should be measured by using the Pakserv (Pakistan Service Quality) model. The Pakserv model places the Islamic faith as one indicator measuring the service quality. Like the Carter model, the Pakserv model also modifies the Servqual model by adding elements of Islamic creed in its measurement. These various measurements of service quality indicate that there is no standard in measuring service quality of Islamic banks. The standardized size is still very high, indicating that there is still no agreement on the right measurement tool. Considering that the country is predominantly Muslim, and most of them are in eastern jurisdiction, to add to the literature, the measurement of service quality in this study combined the Pakserv and Carter's models as a new update of this study. In addition to service quality, the next determinant of Islamic bank customer loyalty is direct and indirect customer satisfaction as stated by Sunaryo (2013) and Basalamah et al. (2018). In creating and maintaining customer loyalty, banks must provide maximum satisfaction to their customers. This satisfaction occurs when the customer's expectations of the services provided by the banking system are actually met. Before leaving or deciding not to use the bank services, most customers might have experienced various problems and submitted complaints about the problems to the bank. As a bank that is run through sharia principles, the element of trust is essential in the Islamic banking business as stated by Koufaris and Sosa (2002: 1). Simply put, trust is the belief that consumers will find what they want from the providers of goods and services. Furthermore, Wu and Liu (2007: 129) argued that trust is the belief that providers of goods and services carry out activities and act in accordance with what is desired by consumers. This trust is believed to strengthen Islamic bank customers to continue using Islamic banks (Casalao et al., 2008;Asif & Anjum, 2012;Hidayat et al., 2015;Okumus, 2005).

Research Conceptual Framework
The conceptual framework in this study was developed through 3 main concepts; i.e. experiential marketing, religiosity, and Islamic banking concepts (Rizwan et al., 2014).

Research Hypotheses
The proposed study considers the hypotheses of the survey based on the following questions.
1. Does religiosity affect customer loyalty, satisfaction, and trust? 2. Does service quality affect customer loyalty, satisfaction, and trust? 3. Does customer satisfaction affect customer loyalty? 4. Does trust affect loyalty? 5. Do customer trust and satisfaction mediate the effect of religiosity on customer loyalty? 6. Do trust and customer satisfaction mediate the effect of service quality on customer loyalty?
The study was conducted in the Special Capital Region of Jakarta. The reason of selecting the location was because the majority of Islamic bank assets, around 56 percent of the total Islamic bank assets in national scale, was in Jakarta. The analysis unit of this study was the Islamic bank customers in DKI Jakarta. The population was all Islamic bank customers in DKI Jakarta. As the number of total populations was unknown, the sampling method was carried out by non-probability sampling with a purposive method using the Hair's rules (2014). With 24 indicators, Hair (2014) suggested that the number of samples used in this study was 120.

Respondents' Descriptions
The respondents of this study were 120 people of Islamic bank customers in DKI Jakarta with the characteristics as follows. First, the gender of the respondents showed slight difference between male and female; second, the majority of respondents, or 54 customers (45.0%), aged over 40 years; third, the majority of respondents had a postgraduate education background. This highlighted that most customers had a high level of education, so that they had good knowledge and understanding of Islamic banks.

Outer Model -Validity and Reliability
Convergent validity was tested through the amount of factor loading indicator to the construct variable. A factor loading of 0.7 or greater is desirable, but a factor loading of 0.4 to 0.7 can still be counted (Hair, 2017). Convergent validity was also expressed using the average value of the extracted variant (Average Variance Extracted / AVE). The AVE value is at least 0.5 (Hair, 2017). Table 2 demonstrates that the results of data analysis using Smart PLS in this study indicated that all indicators and construct variables had met the criteria for convergent validity test.
Internal reliability consistency testing was intended to test the accuracy and precision of a measuring instrument in a measurement procedure. A measurement results can be trusted if the implementation of measurements on the same group of subjects several times obtained relatively similar results. The consistency of internal reliability was tested through Cronbach's alpha and composite reliability as shown in Table 3.  Cronbach's alpha value and composite reliability of all variables are above 0.70 as a measure of reliability eligibility (Hair, 2017). Consequently, all research variables had good reliability and had met internal consistency reliability testing.

Inner Model
The most commonly used measurement to evaluate structural models is the coefficient of determination (R 2 value). R 2 values range from 0 to 1, with a higher level indicating a higher level of prediction accuracy. R 2 values of 0.75, 0.50, or 0.25 for latent endogenous variables can be interpreted as substantial, moderate, or weak (Hair, 2017: 209). R 2 value of customer satisfaction of 0.704 indicates a substantial or strong prediction level which illustrates that variations in customer satisfaction can be explained by religiosity and service quality by 70.4%. Meanwhile, the R 2 value of trust of 0.659 indicates a prediction towards a substantial or strong stating that variations in trust variable can be explained by religiosity and service quality by 65.9%. Meanwhile, R 2 value of customer loyalty variable of 0.747 indicates a predictive level that is substantial or strong. This means that variations in customer loyalty variable can be explained by religiosity, service quality, customer satisfaction, and trust by 74.7%, while the remaining 25.3% is explained by other variables not examined in this study.

Hypothesis Testing
Hypothesis testing was carried out on the direct effects of each exogenous variable on endogenous variables and the indirect effects involving the mediating effect variables.

Direct Effect
The amount of direct effects of exogenous variables on endogenous variables can be seen in Table 5. The results of the hypotheses in Table 5 are illustrated in the form of a path diagram as shown in the figure.
Note: p-values are in the brackets

Indirect Effect (Mediating Effect)
Indirect testing was used to examine the effect of mediating variables on the direct effect of exogenous variables on endogenous variables. The trust and satisfaction variables acted as variables that mediate the effects of religiosity and service quality on customer loyalty. Testing was conducted by using Hair's rules (2017), and the results of the analysis of the mediating effect by using Smart PLS are presented in Table 6.

The Effects of Religiosity on Loyalty, Satisfaction, and Trust
Religiosity is an individual's appreciation and practice of the teachings of a religion or belief that is embraced, leading to a feeling of faith and resulting in an action or behavior. Religiosity does not only occur when someone does ritual worship, but also when carrying out their daily activities. Hypothesis testing in this study indicates that religiosity has a significant positive effect on customer loyalty with a coefficient of 0.487 and a p-value of 0.000. This means that the existing data support the positive effect of religiosity on customer loyalty. The religious behavior of Islamic bank customers statistically has an impact on the increasing loyalty of Islamic bank customers. The religiosity practice of Islamic bank customers will be reflected in their expectations that Islamic banks are actually carried out in accordance with sharia principles. To ensure that Islamic banks are in accordance with sharia principles, there are supervisors from the sharia supervisory board in every sharia bank tasked with providing advice to the boards and overseeing the activities of Islamic financial institutions (LKS) to comply with sharia principles. The results showed that the effect of religiosity on customer satisfaction of Islamic banks was 0.160 with a p-value of 0.041, smaller than the error tolerance limit of 0.05. This shows that there is enough evidence to state that religiosity has a significant positive effect on customer satisfaction. These findings indicate that the more religious the Islamic bank customers in the DKI Jakarta, the more satisfied they will be with the Islamic banks. Trust is a belief that customers will find what they want from goods and services providers (Raajpoot, 2004;Koufaris & Sosa, 2002;Maghfuriyah et al., 2019). In the context of Islamic banking, trust is a belief that Islamic banks are run using sharia principles that are clearly different from conventional banks. (Hoq et al. 2010). This study found enough evidence to state that religiosity has a significant positive effect on trust.
The coefficient of effect of religiosity on trust was 0.206 with a p-value of 0.006. This means that customer trust in Islamic banks is affected by their religiosity. Religiosity as a practice of the religion embraced will have an impact on trust that Islamic banks are run with sharia principles or Islamic law.

The Effects of Service Quality on Loyalty, Satisfaction, and Trust
Butt and Aftab (2013) formulate that good service quality that meets the expectations of customers will be able to build customer loyalty because good service quality will make the customers maintain their commitment to buy goods and services in the future. Thus, the good service quality provided by Islamic banks will be able to make customers continue to use Islamic bank services in the long run, and they will be willing to provide recommendations to others to participate in using Islamic banks. In this study, the coefficient of effect of service quality on loyalty was 0.237 with a p-value of 0.006, smaller than the significance level of 0.05. Hence, there is enough evidence to state that service quality has a significant positive effect on customer loyalty of Islamic banks. Good service quality provided by Islamic banks will continuously bring the impact of confirming the customers' expectations so that they are satisfied with the services. This study found a significant positive effect between service quality and customer satisfaction of Islamic banks with a coefficient value of 0.727 and a p-value of 0,000, smaller than the significance level of 0.05. This highlights that the quality of Islamic banking services will have an impact on increasing satisfaction of Islamic bank customers. Service quality that does not meet the customers' expectations causes the reduce of customer trust (distrust) on the providers of goods and services. Trust is formed by past experiences and previous interactions. The services provided by Islamic banking with better quality and carried out continuously will foster trust from Islamic bank customers that Islamic banks are indeed implemented with sharia principles. Based on the empirical finding, the effect of service quality on trust was 0.663 with a p-value of 0,000, smaller than the significance level of 0.05. Therefore, there is enough evidence to state that service quality has a significant positive effect on trust. The better the quality of services provided by Islamic banks, the more growing sense of trust from customers towards Islamic banks in DKI Jakarta.

The Effect of Trust on Loyalty
The existence of Islamic banks, which are younger than conventional banks, causes customers to question the compatibility of Islamic banks with sharia principles. This doubt has an impact on the lack of trust in Islamic banks. Trust that has not yet been built will bring a reluctance of customers to continue using Islamic banks or recommend them to others resulting in the decreased customer loyalty to Islamic banks. This phenomenon is confirmed in this study where there is not enough evidence to state that trust affects the loyalty of Islamic bank customers. This study empirically found that the effect of trust on the loyalty of Islamic bank customers was -0.031 with a p-value of 0.747, greater than the significance level of 0.005. Hence, trust does not affect the loyalty of Islamic bank customers.

Customer trust and satisfaction mediating the effect of religiosity on customer loyalty
In the context of Islamic banking, Muslim et al. (2013) define trust as the moral obligation of each individual in carrying out their duties in society. Sauer (2002) states that in the banking system, honesty, trust, moral values, and equality among customers are symbols of trust to develop relationships between banks as financial service providers for their customers. This study found the direct effects of religiosity on loyalty and trust. However, the indirect effects of religiosity on loyalty mediated by trust produces a coefficient value of -0.021 with a p-value of 0.752, indicating an insignificant mediation effect. By including trust variable as a mediation, it was revealed that trust was not able to mediate the effect of religiosity on loyalty. The choice of customers to use Islamic banks is because of their religiosity to run sharia-based economic activities. By using Islamic banks, they hope that the performance of Islamic banks can meet their expectations in carrying out Islamic banking activities. Fulfillment of expectations from customers will cause customers to be satisfied, and if this cycle repeats in the long run, it will have an impact on growing loyalty of Islamic bank customers. Subsequently, the effect of religiosity on loyalty can occur directly and indirectly through the mediation of satisfaction. Religiosity perspective as an important factor that identifies customer satisfaction is explained by Metawa and Almossawi (1988) stating that people consider two main factors when choosing a bank; i.e. 1) banks are run based on Islamic law 2) rates of return and profitability. This study found that religiosity has a significant positive effect on satisfaction, and satisfaction has a significant effect on loyalty, and religiosity has a significant positive effect on loyalty. The indirect effect of religiosity on customer loyalty through satisfaction as the mediation was 0.043 with a p-value of 0.127, greater than 0.05, meaning that there is not enough evidence to state that satisfaction mediates the effect of religiosity on customer loyalty of Islamic banks in DKI Jakarta.

Customer trust and satisfaction mediating the effect of service quality on customer loyalty
The mediating role of trust in the effect of service quality on customer loyalty of Islamic banks in this study is shown by a coefficient of -0.021 with a p-value of 0.751, greater than the significance level of 0.050. Thus, there is not enough evidence to state that trust variable mediates the effect of service quality on customer loyalty. The findings of this study differ from Rinaldy's study (2017) examining BNK bank customers in Aceh and revealing that customer trust mediated the effect of service quality on customer loyalty in full. Engaging with customers in banking context is a core business process, so customer satisfaction is the key to success. One main element that determines customer satisfaction is the customer perception on service quality. Customer satisfaction is described as a result of comparison between customer expectations and the perceived service quality performance (Oliver & Swan 1989;Yi 1991). According to this conceptualization, the perception of service quality is one antecedent of overall customer satisfaction. Furthermore, customer satisfaction can lead to customer loyalty. As people tend to be rational and avoid risk, they will have a tendency to reduce risk and stay with service providers that they had good experience with. When a customer of an Islamic bank has felt good service quality, he/she will have a higher level of satisfaction increasing his/her intention to be loyal to the Islamic bank. The mediating effect of satisfaction on the effect of service quality on customer loyalty in this study is shown by the direct and indirect effects of service quality variables on customer loyalty by mediating customer satisfaction. The direct effect of service quality on satisfaction and loyalty both directly and indirectly is statistically significant because it has a p-value of 0.00, smaller than the significance level of 0.05. Hence, customer satisfaction mediates the effect of service quality on customer loyalty of Islamic banks by mediating partial mediation.

Conclusion and suggestion
The conclusions of this study are as follows.
1. Religiosity has a significant positive effect on customer loyalty, satisfaction, and trust.
2. Service quality has a significant positive effect on customer loyalty, satisfaction, and trust. 3. Customer satisfaction has a significant positive effect on customer loyalty. 4. Trust does not significantly affect loyalty. 5. Customer trust and satisfaction does not mediate the effect of religiosity on customer loyalty of Islamic Banks. 6. Trust does not significantly mediate the effect of service quality on customer loyalty. 7. Customer satisfaction significantly mediates the effect of service quality on loyalty.
This study on religiosity, service quality and its effect on customer loyalty of Islamic banks indicate that religiosity variable significantly affects customer trust, satisfaction, and loyalty. This shows that Islamic banks must always raise people's awareness of the need to use Islamic banks as an embodiment of religiosity in Islam, namely the implementation of faith, sharia, and morals in running their economic lives. The empirical findings of this study show that trust variable does not significantly affect loyalty. Hence, this study recommends that the community needs to be educated continuously regarding Islamic bank products and services that they are in accordance with Islamic law and have been guaranteed by the Islamic supervisory boards in every Islamic bank. One renewal of this study is the use of service quality indicators that integrate the Servqual, Carter, and Pakserv models. These indicators were well received by Islamic bank customers. Therefore, further research is advised to use the service quality indicators compiled in this study in measuring service quality of Islamic banks. The indicators of quality in this study involve elements of compliance, personality, sincerity, and formality which are considered appropriate for Muslim community in Indonesia, in addition to Servqual indicators. Meanwhile, this study found that the coefficient of determination (R square) of the endogenous variable of loyalty was 0.747, meaning that variations of Islamic bank customer loyalty can be explained by religiosity, service quality, trust, and satisfaction by 74.7%. The remaining 25.3% was explained by other variables not examined in this study. Subsequently, further research is expected to elaborate on these other variables, such as the brand image of Islamic banks. The findings in this study can only be outlined for the population of Islamic bank customers in DKI Jakarta. For this reason, further research is expected to examine same variables as this study, but in different units of analysis. Further research is also expected to conduct research in different areas of the study, especially in areas with majority Muslim population, so that findings can be compared.