Customer loyalty to Islamic banks: Evidence from Indonesia

Abstract This study identifies determinants of customer loyalty to Islamic banking in Indonesia using a structural equation model (SEM) and confirmatory factor analysis (CFA). Data of 362 respondents from two cities of East Java Indonesia have been used to determine the factors of customer loyalty of state-owned and private Islamic bank. The findings suggest that the customers have different perceptions about state-owned and private Islamic banks; with customer engagement is the strongest influence on customer loyalty in state-owned and customer satisfaction is the strongest influence in private Islamic banks.


Introduction
Technological advancements of the modern era have facilitated customers to decide on retailer offerings at lower switching costs with more transparency (Grewal et al., 2017;Närvänen et al., 2020). This has created customer loyalty a tricky phenomenon (YouGov, 2018). Customer loyalty is considered as the crux of customer relationship management (Nastasoiu & Vandenbosch, 2019;Uncles et al., 2003).It is a significant factor in firm survival, given its significant relationship with profitability. The literature documents a positive relationship between customer loyalty and firm profitability (Yang & Peterson, 2004;Kaura, 2013;Bressolles et al., 2015;Kaura et al., 2015;Huang & Cheng, 2016;Makanyeza & Chikazhe, 2017).
Operations of an Islamic organization are based on customer orientation (Firdaus & Yusuf, 2014). To maintain existing customers and attract new are the primary customer relationship objectives that help sustain an organization according to Shariah objectives (Kartika et al., 2019). Seeking loyalty of customers therefore can be regarded as an ingredient of relationship management in Islam (Kartika et al., 2019). Al-Tamimi et al. (2009) opine that customer loyalty is not Bayu Arie Fianto ABOUT THE AUTHOR Bayu Arie Fianto completed his master degree in business administration from International Islamic University Malaysia and PhD in finance from Lincoln University, New Zealand. He is now an Assistant Professor at Islamic Economics Department, Faculty of Economic and Business, Universitas Airlangga, Indonesia. He has published more than thirty articles in both national and international journals. His current research interest includes Islamic microfinance, Islamic bank, and equity financing in Islamic finance.

PUBLIC INTEREST STATEMENT
Islamic bank is an institution that offers products and services which comply with Islamic law. Islamic banking customers have a special preference to use products that parallel with their religious beliefs. Islamic bank is based on Shariah compliant products such as profit and loss sharing. This concept implies that the customer and the bank share profits using pre-determined ratios. Indonesia is a country with the largest Muslim population in the world and presents a broader market for Islamic financial products and services. This study focus on the determinants of customer loyalty in private and stateowned Islamic banks in Indonesia. a matter of customer relationship management alone. It is rather having such customers who recommend the company and its products leading to an increase in sales and enhancement in company image (Dimitriades, 2006;Mosavi & Ghaedi, 2012). Customer loyalty is actually related to an image of the company and the trust of its customers (Mosavi & Ghaedi, 2012;Phan & Ghantous, 2013;Seiler et al., 2013;Stan et al., 2013;Saleh et al., 2017;Tariq & Moussaoui, 2009;Tahmasbizadeh et al., 2016;Utami, 2015;Yap et al., 2012). It intensify barriers to competition, escalate sales, expand market share, and increases firm profitability (Makanyeza & Chikazhe, 2017). Therefore, firms attempt to strengthen and maintain their relationships with customers in several ways, such as membership cards, special promotions, and targeted marketing (Kim et al., 2020).
Thomson Reuters (2018) reports that out of 69 countries in 2017 Global Islamic Banks' Performance Report 82% reported profits. Besides, Islamic banks rely heavily on customer loyalty to be profitable. Expansion has started to attract non-Muslims to Islamic banks alongside Muslims. Non-Muslim customers generally maintain the relationship with conventional banks that cross-sell Islamic banking products. However, whether the customer is Muslim or Non-Muslim, winning customer loyalty remains the top priority of banks though with different dimensions (Kartika et al., 2019).
Customer loyalty in banks has been addressed in various recent studies. Peng et al. (2019) find service quality, bank image, and trust as antecedents of customer loyalty. According to Monferrer et al. (2019) customer relations represent a factor in attaining customer loyalty through customer satisfaction because it has the strongest total effect on customers. However, studies, for instance, Kashif and Rehman (2015), Altaf (2017), Yilmaz et al. (2018), and Ali and Naeem (2019) conclude that their results cannot be generalized because cultural differences may affect customer loyalty to Islamic banks in various countries.
Indonesia has the world's largest Muslim population and presents a broader market for Islamic financial products and services (Fianto et al., 2019a(Fianto et al., , 2019b. The Islamic Financial Services Board [IFSB] (2018) reports that Islamic banking in emerging countries including Indonesia is exhibiting significant improvements in profitability. By 2019, 14 full-fledged Islamic banks operate in Indonesia; however, their assets are small relative to conventional banks (Fauzi & Suryani, 2019). Indonesian Islamic banks' total assets represent only 0.9% of the global volume (BAPPENAS, 2018). Otoritas Jasa Keuangan (2018) observes that assets of Islamic banks in Indonesia stood at IDR 444.43 trillion in June 2018 which is equivalent to 5.70% of Indonesia's total banking assets. This indicates Islamic banking less popular and less experienced having low awareness and loyalty (Souiden & Rani, 2015). For Islamic banks to develop customer loyalty is very important because it can positively impact business results given its competitive advantage (Suhartanto et al., 2019).
This study investigates the determinants of customer loyalty in private and state-owned Islamic banks in Indonesia. It contributes to the marketing of Islamic banking literature from theoretical and practical perspectives. From a theoretical perspective, it discusses the mediating relationship between customer loyalties and its various determinants viz., customer satisfaction, service quality, the value received, and brand image. From a practical perspective, it will be useful for marketers and policymakers for designing appropriate strategies to promote customer loyalty.

Theoretical foundations of research
Consumer behavior is an evolving phenomenon (Shukla, 2009). Taylor et al. (2004 regard customer loyalty as an aspect of consumer behavior stemming from its roots in various theories. An important theory in this regard is the Stakeholders Theory that claims the relationships of an organization are a function of their performance. In order to understand relationships their performance should be analyzed from the perspective of its stakeholders (Parmar et al., 2010;Walsh, 2005). Another important theory is Fairness Heuristic Theory that asserts to enter into a new relationship with no prior experience masses use "fairness heuristic" (Lind, 2001;Van Den Bos, 2001). This is because "fairness rhetoric" is constructed sooner than trustworthiness and perception (Lind, 2001). Fairness impression can be achieved from respect, consistency, correctability, accuracy, and process clarifications (Van Den Bos, 2001).
Promise theory is another lens to analyze customer loyalty that regards the internal and external marketing of an organization responsible for creating customer loyalty (Bitner, 1995;Calonius, 1986;Grönroos, 2006). Customer loyalty can be explained using a combined lens of Relationship Marketing (Hallowell, 1996), Transaction Cost (Williamson, 1981), Relational Contract (Dwyer et al., 1987;Heide & John, 1992), and Relational Exchange (Macneil, 1980) theories. However, these set of theories cannot be used in the case of Islamic banks because these involve adopting opportunistic behavior on account of lack of information (Filstead, 1981;John & Weitz, 1988) that is against Islamic spirit.
Signaling theory is also used in contemporary researches to explain how the signaler (the service provider) induces the receiver (the customer) to remain loyal to them using various signals (Benlian & Hess, 2011). This is achieved with the modern-day use of information technology, security protocols, and social media (Benlian & Hess, 2011;Boateng, 2019;Mavlanova et al., 2012). Johnson et al. (2006) and Suhartanto et al. (2018) regard loyalty as attitudinal loyalty or loyalty intention that according to Yoon and Uysal (2005) extends from merely buying to spreading recommendations. In this way loyalty can be analyzed using an attitudinal approach under the Tripartite Theory of attitude that claims attitudes consisting of three elements viz., conative, effective, and cognitive (Ajzen, 2005;Johnson et al., 2006;Suhartanto et al., 2018). Analyzing customer loyalty in Islamic banking using Tripartite Theory is therefore recommended (Suhartanto et al., 2018).

Customer loyalty
The marketing goal of a firm is to achieve customer loyalty through their attitudes and purchasing behavior (Watson et al., 2015). Attitude loyalty describes customer attitudes towards a firm relative to other firms offering the same service; and behavioral loyalty (or repurchases) is represented by actual buying behavior (Salem et al., 2019). Wu and Hussein (2019) define customer loyalty as a desire to and tolerance for purchasing at a higher price. For Islamic banks, attitudinal loyalty measures the degree to which customers consider a bank as the first choice to stay in a relationship with them.
Islamic banking customers have a special reason to connect because of their religious beliefs. This is in contrast to conventional banks that are only profit-oriented. Islamic banking is based on Shariah compliant profit and loss sharing (PLS) (Hati et al., 2020). This concept implies that the customer and the bank share profits using pre-determined ratios.
PLS system has become a core value that differentiates Islamic banks from conventional (Fianto et al., 2018). Other factors include services and religiosity (Alemu, 2012). Ireland (2018)

Customer engagement
Customer engagement is a psychological state based on interactive and cooperative customer experiences in service relationships (Brodie et al., 2011). Moliner-Tena et al. (2019) define customer engagement in the banking industry as a psychological connection between banks and their customers resulting from customer satisfaction. Boateng (2019) shows that customer engagement is the intensity of customers becoming connected with the bank through an online system. This results in customers accessing websites or liking and commenting on social media platforms about banks (Shah et al., 2019). Boateng (2019) reveals that for Ghanaian retail banks, customer engagement significantly affects customer loyalty that indicates customers often connect online with banks. Bravo et al. (2019) explain that customer engagement has three determinants viz., purchase behavior, the influence of social interactions, and sharing bank knowledge. This study defines customer engagement as the frequency of customer contacts with an Islamic bank. It results in their emotional and psychological connections that affect their loyalty. Thus, the following relationship is hypothesized. H1: For Islamic banks, customer engagement and customer loyalty are correlated.

Service quality
Increased global competition has led organizations to maintain a competitive advantage of service quality (Makanyeza & Chikazhe, 2017). Feiz (2019) explains that one of the most crucial service quality factors is employee behavior. This includes helpfulness, friendliness, and maintaining an intimate and responsible image towards customers. Peng et al. (2019) reveal that Islamic bank customers have a stronger perception of quality because of Shariah compliance. Islamic banks provide Shariah compliant service quality in addition to safe banking products and services (Tabrani et al., 2018). A study by Ltifi et al. (2016) reveal that customers distinguish the service quality of a bank on the basis of physical attributes such as availability of services, the turnaround time for financing approval and the speed of financial transactions. Iqbal et al. (2018) explain that the competency of Islamic bankers to provide solutions makes customers more satisfied that in turn leads to the enhanced perceived value of Islamic principles.
Employee behavior and physical product quality significantly affect customer loyalty. Service quality is directly related to perceived value. Mahadin (2019) defines perceived value as customer perception about service performance relative to money paid. This indicates fairness of the price and customer satisfaction with banking service. Thus, a product with high price and service quality results in increased value perceived. Mahadin (2019) demonstrates that service quality has a significantly positive impact on perceived value. Hence, the following relationship is hypothesized.
H2: A positive relationship exists between Islamic banks' service quality and perceived value. Akroush and Jraisat (2016) observe that brand image improves with service quality. Quality plays an important role in forming a brand image because it is assessed using the perspective of the service quality (Feiz, 2019). Peng et al. (2019) define a brand image as a reflection of the service quality to the public. Kaabachi and Obeid (2016) report that improved image, service quality, and reputation of Islamic banks in Tunisia have attracted new customers with more loyalty. Wu and Hussein (2019) ascertain that the higher service quality of Indonesian Islamic banks has improved their image. Therefore, the following relationship is hypothesized.
H3: A significant positive relationship exists between Islamic banks' service quality and brand image. Dusuki and Abdullah (2007) demonstrate that service quality is the most significant variable that affects customer satisfaction. Yilmaz et al. (2018) observe that a bank's accessibility and staff responsiveness significantly affects customer satisfaction. Wu and Hussein (2019) state that the service quality of Islamic banks positively affects customer satisfaction. Similarly, Bazrkar et al. (2018) andDe Bruin et al. (2020) find service quality playing a vital role in increasing customers' satisfaction with Islamic banks. Therefore, the following relationship is hypothesized.
H4: A positive relationship exists between Islamic banks' service quality and customer satisfaction. Athanasopoulou and Tsoukatos (2015) and Huang and Cheng (2016) find a positive relationship between service quality and customer loyalty. Several studies such as Jamal and Naser (2002), Ali and Naeem (2019), and Boonlertvanich (2019) find service quality as the strongest influence on customer loyalty in the banking sector. Thus, better service quality is likely to keep customers loyal. Haron et al. (2020) and Garepasha et al. (2020) use Malaysian and Iranian Islamic banking data to ascertain that service quality positively affects customer loyalty. They recommend that Islamic banking staff should be polite, professional, and pay more attention to delivering appropriate financial advice for easy access to account information. Service quality is found to an essential role in shaping customer loyalty to Indonesian Islamic banks (Fusva et al., 2020). Therefore, this study asserts that service quality has a direct effect on customer loyalty. This leads to the hypothesis: H5: A positive relationship exists between Islamic banks' service quality and customer loyalty. Zietsman et al. (2019) and Abubakar (2019) define perceived value as a customer perspective of what they receive and pay relative to service quality they receive. Kaura et al. (2015) explain that perception of fairness is important in bank selection because banking price structures are relatively more complex. Mahadin (2019) reveals that Islamic bank customers are very concerned with perceived value. Perceived value is therefore a response of the customers to final value relative to prices that affect customer satisfaction. Therefore, the following relationship is hypothesized.

Perceived value
H6: A positive relationship exists between Islamic banks' perceived value and customer satisfaction.
An existing study in the banking context shows that perceived value is significantly related to customer loyalty (Lewis & Soureli, 2006). Moreover, Jiang et al. (2015) reveal that for e-banking services perceived value significantly influences customer loyalty. Boonlertvanich (2011) and Rahi (2016) observe that perceived value is an important factor that affects customer loyalty in Thailand and Pakistan retail-banking. Thus, a positive experience with perceived value encourages maintaining contact with the bank and strengthening loyalty. Therefore, this study formulates the following hypothesis.
H7: A positive relationship exists between Islamic banks' perceived value and customer loyalty. Moliner et al. (2018) explain that customer engagement results from customer satisfaction with bank service quality. The higher service quality relative to price affects perceived value. Thus, an increase in the perceived value of Islamic bank service quality leads to stronger customer engagement. Therefore, the following hypothesis is formulated.
H8: A positive relationship exists between Islamic banks' perceived value and customer engagement. Sayani (2015) and Kashif and Rehman (2015) explain that satisfaction is a happy feeling when a product meets expectations. Satisfaction as a result of dis-confirmation of desires is evaluative satisfaction, whereas satisfaction as a result of non-rational forms is emotion-laden (Suhartanto et al., 2019). Islamic banks must be careful when displaying their image in products and services (Dusuki & Abdullah, 2007). Competition between Islamic banks, especially in developing countries guarantees that these banks distinguish their appearance and offer significantly different services (Iqbal et al., 2018). Al-Tamimi et al. (2009) observe that Islamic banks portray different images from a customer perspective. Thus, customer satisfaction for Islamic banks is derived not only from returns but is also related to their ability to produce Shariah compliant services. The ability of Islamic banks to improve customer satisfaction creates a good brand image. Therefore, the following hypothesis is formulated.

Customer satisfaction
H9: A significant relationship exists between Islamic banks' customer satisfaction and brand image.
The theory of engagement explains that satisfaction and emotional connection are primarily why customers engage with a brand or product (Pansari & Kumar, 2017). Moliner-Tena et al. (2019) observe that customer satisfaction plays an important role as an antecedent of customer engagement. They explain customer engagement as an accumulation of customer satisfaction and emotions during a service. Therefore, increasing customer satisfaction increases the degree of engagement. The following relationship is hypothesized.
H10: A significant relationship exists between Islamic banks' customer satisfaction and customer engagement. Fornell (1992) and Kaura et al. (2015) show that higher customer satisfaction improves customer loyalty. Lewis and Soureli (2006) and Yilmaz et al. (2018). Fauzi and Suryani (2019) find customer satisfaction with Islamic banks in Indonesia directly related to customer loyalty. Recent studies by Mulia et al. (2020) and Haron et al. (2020) explain that satisfied customers tend to reuse and spread positive stories about Islamic mobile banking to their families and friends. Therefore, the following relationship is hypothesized. H11: A significant relationship exists between Islamic banks' customer satisfaction and customer loyalty.

Brand image
A company's image is closely related to its reputation. Altaf (2017) defines brand image as the sum of customers' impressions that form trust in Islamic banking. Ali and Naeem (2019) find that Shariah compliance is one of the most important factors that attracts customer attention creating a positive image. A positive brand image enhances the value of the brand and protects it from threats. Sumarwan (2014) and Simanjuntak et al. (2020) observe positive brand image enhancing commitment, and emotional and psychological attachment. Moliner et al. (2018) find a strong emotional connection between customers and the brand image of the banks. Such a connection encourages customers to defend the brand, spread positive comments, and perform actions that reflect customer engagement. Lewis and Soureli (2006), Amin et al. (2013), Makanyeza and Chikazhe (2017), and Simanjuntak et al. (2020) demonstrate that brand image positively affects customer loyalty. In summary, improving brand image increases customer loyalty and engagement. Therefore, the following relationships are hypothesized.
H12: A significant relationship exists between Islamic banks' brand image and customer loyalty.
H13: A significant relationship exists between Islamic banks' brand image and customer engagement.

Indirect effects among the investigated variables
Perceived value mediates the relationship between service quality and customer loyalty (Hapsari et al., 2017). Raza et al. (2020) find that online service quality increases customer loyalty to online banking services in Pakistan. Jiang et al. (2015) observe that customer-perceived value partially mediates the relationship between e-service quality and customer loyalty in e-banking. Therefore, the following relationship is hypothesized.
H14a: The perceived value received mediates the effect of service quality on Islamic banks' customer loyalty.
Feiz (2019) explains that brand image represents the customer perspective of Islamic bank service quality. Alnaser et al. (2018) report that service quality increases the perceived value, satisfaction, and loyalty of the customers. Besides, the process of service quality is also observed by the customers that contributes to creating loyalty (Othman & Harun, 2020). Quality is the most important factor in building brand image because the brand appraisal is based on the perception of service quality. Feiz (2019) also reveals that brand image effectively extends customer loyalty when Iranian banks provide better physical service quality. Thus, the following relationship is hypothesized.
H14b: Brand image mediates the effect of service quality on Islamic banks' customer loyalty. Bloemer et al. (1998) observe that customer loyalty is both directly and indirectly influenced by service quality through customer satisfaction. Janahi and Al Mubarak (2017) reveal service quality positively influencing satisfaction leading to increased customer loyalty. Kaura et al. (2015) find that customer satisfaction acts as a mediating variable between service quality and customer loyalty in Indian retail banking. Asnawi and Fanani (2019) report customer satisfaction partially mediating the relationship between service quality and customer loyalty. These findings indicate that customer satisfaction is a mediator between service quality and customer loyalty. Therefore, the following relationship is hypothesized.
H14c: Customer satisfaction mediates the effect of Islamic banks' service quality on customer loyalty. Lewis and Soureli (2006) explain that in banking, perceived value positively impacts customer loyalty and customer satisfaction. Mahadin (2019) states that customer satisfaction in Islamic banks is directly proportional to service quality and related cost. Furthermore, Shariah compliant service quality leads to increased customer satisfaction (Mahadin, 2019). Thus, the following relationship is hypothesized.
H15a: Customer satisfaction mediates the effect of Islamic banks' perceived value on customer loyalty. Sondhi et al. (2017) observe that customer engagement is the most influential factor on the length of customer association with banks; and perceived value is determined by customer loyalty. Mahadin (2019) reveals that customers buy a product that meets their expectations depending on what they sacrifice and what they receive. Thus, higher customer loyalty might be caused by the perceived value through customer engagement. Thus, the following relationship is hypothesized.
H15b: Customer engagement mediates the effect of Islamic banks' perceived value on customer loyalty. Hapsari et al. (2017) conclude that an indirect relationship exists between brand image and customer loyalty mediated by customer involvement. Bravo et al. (2019) observe that customer engagement is determined by brand trust and commitment. Brand image results in a higher degree of customer engagement that in turn leads to stronger customer loyalty. Specifically, a Shariah compliant brand image indicates that a bank complies with Islamic principles. This affects their customer engagement and influences customer loyalty. Thus, the following relationship is hypothesized.
H16: Customer engagement mediates the relationship between Islamic banks' brand image and customer loyalty.

Customer loyalty and the relationships between variables
The banking industry literature indicates that direct and indirect relationships exist between customer satisfaction, customer engagement, and customer loyalty. Amin et al. (2011) explain that satisfied customers tend to engage with their banks in long-term relationships. The engaged customers have a higher intensity and stronger emotional bond that is created from satisfaction with the bank. Thus, the following relationship is hypothesized.
H17: Customer engagement mediates the relationship between Islamic banks' customer satisfaction and customer loyalty.

Data
The research sample has been drawn from customers of full-fledged state-owned and private Islamic banks in the cities of Surabaya and Sidoarjo located in the East Java province of Indonesia. A face-to-face survey has been conducted using convenience sampling. This is because of the practical difficulties faced in obtaining a comprehensive list and detailed information on the target population of Islamic bank customers needed for a probabilistic sample. Thus, the results of the survey cannot be interpreted beyond the sample.

Figure 1. Conceptual model of the study relationships.
Four hundred questionnaires have been distributed and 362 responded making a response rate of 90.5% (38 responses have been eliminated because of inadequate information). The responses have been divided into customers of private banks (69 respondents) and state-owned banks (293 respondents).

Method
A pilot study has been carried out on 30 customers that has enhanced the reliability by ensuring content validity. All questionnaire items use a standard 7-point Likert-type scale ranging from Very Strongly Disagree (1) to Very Strongly Agree (7). A performance-based measure has been used to assess the interrelationships between constructs (Tabrani et al., 2018). Levene's test for equality of variances and t-test for equality of means has been conducted to ensure that the data set is free of early-late response bias (Armstrong & Overton, 1977). Missing values account for less than 5% and appear randomly. Mean substitution method has been used to substitute all missing values.
The first stage of the analysis applies confirmatory factor analysis (CFA) to test whether the model specifications fit and match the actual conditions of the sample (Harrington, 2008). Structural equation modeling (SEM) has been used to test hypotheses H1 to H17. Table 1 lists demographic characteristics. The results suggest that there exists a significant gender difference (χ 2 = 7.396, significant at the 1% level) between the customers of private and stateowned Islamic banks with females dominant in the former and males in the later. The dominant respondent age group is 17-25 years with the Muslim faith. A majority of private Islamic bank respondents hold a bachelor's degree; whereas in state-owned, a majority own senior high school certificate and bachelor's degree. This information confirms that private Islamic bank customers have better education, difference significant (χ 2 = 18.742, significant at the 1% level).

Results
The most common occupation for private Islamic bank customers is "other" (40.6%) while for state-owned banks "student" is dominant (48.5%). Respondents have been asked to specify the other occupation, and most answered "housewife." The difference in occupation is significant (χ 2 = 19.860, significant at the 1% level). A majority of customers from private and state-owned Islamic banks fall in the 50-100 million IDR income category. Also a majority of respondents from both types of banks have not reported any family member working as a government officer.
The CFA of the original model indicates unfit indices (χ 2 /df = 2.234, PGFI = 0.630, GFI = 0.721, RMSEA = 0.059, RMR = 0.077, NFI = 0.822, CFI = 0.893). This result requires model modifications to improve the goodness-of-fit indices because the model estimate has fit indices that do not fall within acceptable cut-off limits (Boateng, 2019). An improved model fit has been achieved by deleting 20% or less of the total items, measuring each construct, and then re-specifying the model (Hair et al., 2009;Hapsari et al., 2017). The model modification involves removing five customer engagement items, two service quality items, two customer loyalty items, and one item each from brand image, customer satisfaction, and perceived value. These modifications resulted in a close, marginal, and good model fit (χ 2 /df = 1.400; PGFI = 0.734; GFI = 0.868; RMSEA = 0.040; SRMR = 0.038; NFI = 0.893; CFI = 0.967).
Discriminant validity has been assessed using the average variance extracted (AVE), construct reliability (CR), and Cronbach's alpha. The recommended factor loadings and AVE values supporting discriminant validity must be higher than 0.5 (Hair et al., 2009;Iqbal et al., 2018), and an acceptable Cronbach's alpha value must be higher than 0.7 (Hair et al., 2009;Wu & Hussein, 2019). Table 2 shows that the factor loading, AVE, CR, and Cronbach's alpha for all discriminants exceeds the threshold values of 0.5 and 0.7, respectively. H2, H3, H4, H6, H8, H9, H10, H11, H12, and H13 are supported, and H5 and H7 are not supported. Table 3 also shows that customer satisfaction is the most important factor affecting customer loyalty to private Islamic banks. This finding is confirmed by the highest coefficient for customer loyalty at 1.267 significant at 10%. For state-owned Islamic banks, service quality exhibits the highest impact on customer loyalty (coefficient of 1.067, significant at 1%). Table 4 presents the results of seven hypotheses; service quality (SQ) through customer satisfaction (CS), perceived value (PV) through customer satisfaction, customer satisfaction through customer engagement (CE), and brand image (BI) through customer engagement are supported for both type of banks (see Table 4):

Discussion and implications
Results confirm that most of the hypotheses are significant and explain relationships between the variables used in this study.
Between CE and customer loyalty (H1) only state-owned Islamic banks exhibit significantly positive relationships. This conforms to the findings of Boateng (2019) who report that CE with Islamic banks plays important role in strengthening customer loyalty. This result might be caused by customers' knowledge and understanding of private and state-owned Islamic banks. Abubakar (2019) explains that different understandings and perceptions create varying degrees of CE. This study reveals that CE is positively influenced by a strong commitment and a supportive government. In Indonesia, a widely held stereotype belief in society is that a state-owned institution is better and cheaper than a private institution. Thus, customer loyalty to private Islamic banks might be influenced by lower CE.
SQ and PV (H2) have significantly positive relationship for private and state-owned Islamic banks. This result conforms to Mahadin (2019) who observe that SQ significantly positively affects development of customers' PV in Islamic banking. Wu and Hussein (2019) state that Islamic banks' SQ is not only based on profit orientation but also on Shariah compliance. Customers who believe that Islamic banks run their business in compliance with Shariah have higher PV; even if they spend more time and pay a higher price.
SQ and BI (H3) relationship is significant only for state-owned Islamic banks. This is because they gradually improve in SQ by ensuring Shariah compliance, thus creating a significant good BI for customers. Feiz (2019) observe that, generally, Islamic banks in Indonesia do not execute their businesses in true compliance with Shariah, thus influencing the image of Islamic brands not different from conventional. Iqbal et al. (2018) explain that Islamic bank employees must improve their understanding of bank regulations and customer needs to serve effectively that can lead to higher SQ and BI.
The relationship between SQ and CS is significantly positive for both private and state-owned Islamic banks (H4). This conforms to Amin et al. (2013) Bazrkar et al. (2018), Dusuki and Abdullah (2007) Kant et al. (2019), Wu andHussein (2019)De Bruin et al. (2020), and Abror et al. (2019) explain that when customers perceive good quality service their satisfaction improves. This indicates that Islamic banks have improved their SQ based on Shariah and profit. Asnawi and Fanani (2019) explain that as Islamic banks in Indonesia improve their SQ customers become more satisfied. In the Iranian Islamic banking context Feiz (2019) shows that customers are not satisfied with the SQ. The contradictory results of his study might due to different cultures and locations. Aaker (1997) and Aaker et al. (2001) state that several countries have different personality dimensions that reflect their cultures. In the context of SQ, the findings of Aaker (1997) and Aaker et al. (2001) might be different for every country. Therefore, this study suggests that Islamic banks should consider their customer needs, present greater care, and maintain consistency in services while earning profits.
The PV and CS of Indonesian Islamic banks have a significantly positive relationship (H6). This finding supports Abubakar (2019), Hamouda (2019), Kaura et al. (2015), and Mahadin (2019), who explain that PV represents customer perception of the balance between the benefits they receive and the money they pay. For Islamic banks in Indonesia, Rahayu et al. (2020) reveal that Muslims perceive value in the form of price, emotional value, and social value that directly affect CS. Customers assess bank performance against their expectations that determines CS. In summary, customers believe that the services of Islamic banks are affordable and meet their expectations.
Private and state-owned Indonesian Islamic banks' PV and CE have a significantly negative relationship (H8) that indicates a change in PV significantly affects CE. This supports Carlos Fandos Roig et al. (2006) who report that increasing PV can be encouraged by providing higher quality services. Moreover, Abubakar (2019) explains that Islamic bank customers purchase products and services that are equivalent to their PV. This is important in enabling customers to become more engaged with banks.
CS and BI (H9) have significantly positive relationships for both types of banks. This supports Amin et al. (2013) and Wu and Hussein (2019) who observe that BI of Islamic banks is influenced by Shariah compliance. When Islamic bank customers strongly believe that their banks are Shariah compliant they are more satisfied. These results reflect that private and state-owned Islamic banks produce good BI.
CS and CE indicate a significantly positive relationship for private and state-owned Islamic banks (H10). This supports Marino and Lo Presti (2018) and Abror et al. (2019) who observe the significant effect of customer emotional engagement on CS. This study suggests that Islamic banks should improve their services to improve CS and CE.
Hypothesis (H11) infers that CS has a significantly positive influence on customer loyalty in private and state-owned Islamic banks. This conforms to Lewis and Soureli (2006) Amin et al. (2013) show a significant relationship between CS and loyalty in Muslim and non-Muslim Islamic banking customers in Malaysia. Asnawi and Fanani (2019) demonstrate that in Indonesia a more satisfied customer is more loyal to Islamic banks.
BI has a significantly positive impact on customer loyalty in state-owned Islamic banks only (H12). This extends Amin et al. (2013)  BI has a positive and significant relationship with CE for both private and state-owned Islamic banks (H13). Moliner et al. (2018) and Yasin et al. (2020) recommend that to integrate its brand with customer self-concept an Islamic bank should manage its brand with great care and focus on every aspect of its services.
Regarding indirect relationships, brand image mediates the relationship between SQ and CL in the case of state-owned banks only (H14b). This extends the findings of Feiz (2019), Alnaser et al. (2018), and Othman and Harun (2020) who do not make any distinction between private and state-owned banks in their works.
H14c infers that CS in both private and state-owned Islamic banks mediates between SQ and customer loyalty. Several studies such as Butt and Aftab (2012), Kaura et al. (2015), Janahi and Al Mubarak (2017), and Asnawi and Fanani (2019) observe CS as a mediator between SQ and customer loyalty. These results signify that when Indonesian Islamic bank customers receive better quality service their satisfaction is enhanced and loyalty to the bank is strengthened. This study suggests that Indonesian Islamic banks should endeavor to improve CS and customer loyalty.
CS mediates between PV and customer loyalty in private and state-owned Islamic banks (H15a). This study conforms to Rahayu et al. (2020) who report mediation of CS between customer loyalty and PV. These results indicate that neither private nor state-owned Islamic banks operate in line with customer expectations and doing so could improve CS and strengthen customer loyalty. CE mediates the relationship between BI and customer loyalty (H16) in state-owned Islamic banks only. This result confirms that CS for Indonesian-state-owned Islamic banks is affected by banks' BI consequently influencing customer loyalty. This extends the research of Greve (2014) who observes that BI has a significantly positive impact on engagement activities; a stronger BI drives higher CE; and BI positively affects brand loyalty irrespective of the type of banks.
These results conclude that BI plays an important role in building CE that in turn leads to customer loyalty in state-owned Islamic banks. This also extends the scope of Rahmawati et al. (2019) who reveal that Indonesian Islamic banks use Instagram to enhance their promotional activities and provide better information about their BI that drives CE and improves customer loyalty irrespective of the type of bank. CE mediates the relationship between CS and customer loyalty (H17) in state-owned Islamic banks. This extends to Thakur (2019) who observes that CE affects the relationships between CS and loyalty to mobile shopping, and CS and travel planning apps. Moreover, Monferrer et al. (2019) also observe that satisfaction is the most important variable for improving CE and loyalty to Islamic retail banking. Thus, the results suggest that Islamic banks should improve CS to increase customer loyalty through stronger CE.
Finally, H1, H3, and H12 confirm that customers of state-owned Islamic banks are more loyal. This is because Indonesian customers have a better opinion about government institutions. Mahadin (2019) and Abubakar (2019) state that government support also influences CE with Islamic banks. Specifically, Sayani (2015) states that a key success factor of Islamic banks in Malaysia is the strong government commitment to actively promoting Islamic banks. Thus, this study suggests the government should emphasize and promote customer loyalty to both private and state-owned Islamic banks to reduce different perceptions of private and state-owned Islamic banks.

Study limitations and directions for future research
The results of this study are based on a convenience sample of Islamic bank customers in the East Java province of Indonesia. The loyalty measurement is based on behavioral intentions that are also used as a predictor of actual behavior. Loyalty behavioral intention might not translate into actual behavior if customers face obstacles to loyalty such as an unavailable product, switching incentives, and customer idiosyncrasies. An exploratory study can provide a deeper understanding of the CE. Some scholars (e.g., Moliner-Tena et al., 2019;Van Doorn et al., 2010) consider CE to be a single construct; however, other scholars suggest that CE is multi-dimensional. Future research can replicate the conceptual model used in this study and apply it to predict loyalty to Islamic banks in other regions with different cultures.
Finally, this research identifies that customer loyalty has different dimensions in private and state-owned Islamic banks. This is specifically with respect to brand image and customer engagement that mediate various relationships in state-owned Islamic banks. Exploring these issues in detailed private and state-owned Islamic bank comparative studies can enhance understanding of the factors affecting customer loyalty in customer relationship management. Such studies are important because these can help streamline different patterns of growth in Islamic banks. The relationships identified in this study can be expanded to include Islamic bank customers in all Islamic and non-Islamic countries for additions to the current body of knowledge.