Determinants of customer satisfaction in commercial banks: A case study in quang nam province, Vietnam

AGU International Journal of Sciences – 2019, Vol. 7 (2), 42 – 55 42 DETERMINANTS OF CUSTOMER SATISFACTION IN COMMERCIAL BANKS: A CASE STUDY IN QUANG NAM PROVINCE, VIETNAM Huynh Tan Nguyen1, Nguyen Hoang Hung1 1Dong Nai Technology University Information: Received: 18/09/2018 Accepted: 29/03/2019 Published: 11/2019 Keywords: Customer Satisfaction, Structural Equation Modelling (SEM), Commercial Banks. ABSTRACT Currently, customer satisfaction is often considered as the sustainable and long-term strategy to expand business operations for any banking institutions. Hence, the purpose of this article is to explore and measure factors affecting customer satisfaction in commercial banks via both qualitative and quantitative approaches. To collect data, the questionnaire was designed to officially survey 500 customers at nine commercial banks in Quang Nam province, Vietnam. Next, the study used Structural Equation Modelling (SEM) to test the proposed research hypotheses. It is found that there are four factors influencing customer satisfaction, including Corporate Image, Security, Service Price, Customer Expectations, Empathy. Based on the findings, the paper proposed some managerial implications to enhance customer satisfaction in the next period. Yet, the study had some limitations, so more studies with alternative methods should be conducted to confirm the results of this study for better policies and practices. 1. INTRODUCTION Doing business in the context of such competitive environment nowadays, customer satisfaction is considered as a main factor affecting business operations of any commercial bank (Colm & Parasuraman, 2017), in which the bank retains customer interest and loyalty. Doing so will provide opportunities to widen market share, increase profit and develop business operation sustainably (Hennig & Klee, 1997). Thus, the commercial banks are shifting business strategies towards improvement of customer satisfaction (Yang & Peterson, 2004). Many empirical studies have illustrated that profits and business efficiency only come from satisfying customer expectations and demands (Le Thị Thanh Giang, 2016; Lone, Aldawood, & Bhat, 2017; Wolter et al., 2017). However, today's competitive market has forced banks to only focus on reducing operating cost for gaining short-term high profit instead of improving service quality and enhancing customer satisfaction, which is regarded as a long-term business strategy. That’s why exploring determinants of customer satisfaction is presently a popular research topic in research as regards human behavior. The empirical research of Sirdeshmukh, Deepak, Singh, and Sabol (2002) demonstrates that customer satisfaction is significantly different between the state-owned commercial banks and the private commercial banks with higher satisfaction in the private banks. The reasons for this phenomenon may stem from deliverability, accessibility as well as transaction convenience, AGU International Journal of Sciences – 2019, Vol. 7 (2), 42 – 55 43 which lead to the advancement of customer service capacity. This finding is consistent with the study of Janahi and Mubarak (2017) and Sirdeshmukh et al., (2002). There are presently numerous ongoing empirical researches in the context of the Vietnamese banking industry, yet, descriptive statistics of customer satisfaction instead of testing the relationship among research constructs is one of the main limitations of these studies. Furthermore, the other limitation is that the studies are only conducted at one commercial bank, so the results are generally difficult to apply for overall system of commercial banking in Vietnam. Hence, this research aims (1) to explore and measure the factors influencing customer satisfaction in commercial banks in Vietnam; (2) from that, to propose effective recommendations to improve service quality as well as boost customer satisfaction. 2. LITERATURE REVIEW AND RESEARCH MODEL 2.1 Customer Satisfaction Oliver (2014) states that satisfaction is the understanding of the consumer comparison between the expectation and performance. Kotler et al. (2009) postulate that satisfaction is “a person's feelings of pleasure or disappointment that results from comparing a product's perceived performance or outcome with his/her expectations”. According to Fečiková (2004), satisfaction is the feeling of customers originated from the procedure of assessing what they obtain and what they expect. Abdullah and Rozario (2009) emphasize that customer satisfaction may be affected by several various internal and external elements. Hence, it's difficult to determine client satisfaction. In the opinion of Veloutsou et al. (2005), overall customer satisfaction is a moving process and will change the clients’ experience with the service and products that they purchase. 2.2 Development of hypothesis 2.2.1 Corporate image Corporate image may be regarded as “the perceptions of an enterprise” that exists lastingly in consumers' memory, and also helps customers to recognize the difference between the product and service of the organization from their competitors as well (Elgin & Nedunchezhian, 2012). In the opinion of Connor and Davidson (1997), an enterprise with a positive corporate image is likely to lead to attraction of new customers and retain existing ones. Because of the current competitive nature of the banking sector nowadays, corporate image plays a crucial role in seeking and attracting new customers, owing to the fact that the more reputable banks will be easy to appeal to new customers. Various empirical evidence have suggested that corporate image is a vital factor of customer satisfaction (Johnson et al., 2001). Thus, it is hypothesized as follow: H1: Corporate image has a positive influence on customer satisfaction. 2.2.2 Security Cyberattacks in the financial industry, especially the banking sector, are recently increasing more and more seriously when so many recent attacks have become more complex than before (Amin, 2016). The reason may be explained as the quite rapid increase of the quantity of financial transactions and the continued advancement of new technologies in the banking sector (Fabrigar & Wegener, 2011). Therefore, security is an extremely important factor for financial operations. Although there is much technological progress towards security namely cryptography, electronic signatures, as well as multi-factor authentication, consumers are still at risk in monetary transactions, particularly online transactions (Hazen et al., 2017). Consequently, if the security is assured, customer satisfaction may increase. Several previous studies in retail banks also demonstrate that security is a significant determinant of customer satisfaction. Therefore it is hypothesized that: AGU International Journal of Sciences – 2019, Vol. 7 (2), 42 – 55 44 H2: Security has a positive influence on customer satisfaction. 2.2.3 Service price For a specific bank, the service price is one of the factors of the marketing strategy. The marketing perspectives evaluate the service price in the banking sector, especially, from the customer’s point of view. The service price is regarded as what customers actually pay for their transactions including the loan interest, the money transfer cost, the other service fees. Many recent researches have shown that service price is one of the significant factors influencing service quality and customer satisfaction in the banking industry (Hazen, Boone, Wang, & Khor, 2017). Hence, with this factor, the hypothesis is posited as:
 H3: Service price has a positive effect on customer satisfaction. 2.2.4 Customer expectations Customer expectations are defined as the perceived value or benefits that the customer perceives before consuming goods or services from providers. They may be derived from the “experiencing” process and can also be shaped immediately with even first impressions during purchasing (Cronin Jr et al., 2000; Hult et al., 2017). Once formed, these expectations are hardly to change and are likely to considerably affect clients’ decision-making processes. Furthermore, customer expectations might be established from customers' consumption experience in the past and largely depend on the firm's product or service quality as well as promotion strategies and word-of-mouth information (Johnson et al., 2001). Therefore, customers will evaluate the performance efficiency of a bank via their expectations. Thus, it is worth mentioning that the commercial banks ought to identify customer expectations and to seek to meet their expectations with the aim of maximizing their satisfaction (Jamal & Naser, 2002). Accordingly, the following hypothesis is proposed as: H4: Customer expectations have a positive impact on the customer satisfaction. 2.2.5 Empathy Empathy refers to the possibility to understand another person’s feelings and experiences within their frame of reference (Kotler, Kartajaya, & Setiawan, 2018). The different types of empathy rely mainly on the range of emotional states, which result in three types of empathy including emotional empathy, somatic empathy as well as cognitive empathy (Levesque & McDougall, 1996). Some researchers have demonstrated role of empathy as an essential part of service quality in any enterprise, and certainly affecting customer satisfaction. Many prior studies have postulated that empathy has a direct and considerable effect on re-purchasing behavior of the clients (Tadic et al., 2018; Yang, 2017). Empathy in the banking sector is also reflected through free service capacity, which relates to caring, listening and careful compensation in case of any mistake (Ekiz et al., 2006; Farooq, 2016). Hence, it is hypothesized that: H5: Empathy has a positive impact on customer satisfaction. 2.3 Proposed research model Based on the literature review from the previous studies, the proposed theoretical model includes six constructs. These are depicted in Fig. 2.1 below. AGU International Journal of Sciences – 2019, Vol. 7 (2), 42 – 55 45 Figure 2.1. Proposed research model 3. METHODOLOGY 3.1 Measures of the constructs The items used to establish constructs are mainly adapted from prior studies to assure content validity. Measures for the independent constructs were taken from Lone et al. (2017), Shahin et al. (2017), Verhoeven and Sha (2017), Fabrigar and Wegener (2011), Hazen, Boone, Wang, and Khor (2017), Hult et al. (2017). Measures of customer satisfaction construct was taken from Fabrigar and Wegener (2011), Lone et al. (2017), Verhoeven and Sha (2017), Hazen, Boone, Wang, and Khor (2017). This research used five-point likert scales to collect customer responses for each construct, in which 1 represented as “strongly disagree” while 5 reflected as “strongly agree”. 3.2 Sampling Hair, Black, Babin, Anderson, and Tatham (2006) suggest that minimum sample size in EFA is 50 observations, and preferably sample size should be 100 or larger and observations/measured variables is 5:1, more acceptable sample size would have a 10:1 ratio. From this general rule, this study surveyed 500 customers at 09 commercial banks in Quang Nam province, Vietnam. Then, cleaning inappropriate responses, official data was used to analyze as follow: Table 1. The result of survey responses at each bank Frequency Percent Valid Vietcombank 39 8.46 BIDV 34 7.38 MHB 44 9.54 Dong A Bank 37 8.03 (+) (+) (+) (+) (+) Corporate image (F1) Security (F2) Service price (F3) Customer expectations (F4) Empathy (F5) Customer satisfaction (Y) AGU International Journal of Sciences – 2019, Vol. 7 (2), 42 – 55 46 Frequency Percent Techcombank 53 11.5 Maritimebank 61 13.23 ACB 73 15.84 VPbank 33 7.16 LienViet Post Bank 87 18.87 Total 461 100 3.3 Data analysis methods After finishing the phase of qualitative research, the official questionnaires are delivered to clients transacting at the commercial banks to collect necessary information for the study. Then all data is cleaned missing information, coded, and analyzed through SPSS 22 and AMOS 22. Some techniques are utilized for quantitative research, consisting of: Firstly, the study carries out descriptive statistic technique to examine mean, standard deviation, and minimum value and maximum value of observed variables. Furthermore, this technique is also exploited to test the difference between controlled variables in proposed research model at a particular significant level. Secondly, the reliability of overall measurement scale is tested by Cronbach's alpha, which is a statistical test used to estimate the reliability or internal consistency of a composite scale, originates from the paradigm of (Cronbach, 1951). According to Nunnally and Bernstein (1994) and Hair Jr et al. (2016), a scale is considered reliable if Cronbach's Alpha coefficient greater than 0.7 and observed variables have corrected item-total correlation greater than 0.3. Thirdly, exploratory factor analysis (EFA) is carried out to identify trends in the sample data and reduce a set of observed variables by grouping them together (Ho, 2006, Hair et al., 2015). In addition, the aim of applying factor analysis is to explore the most appropriate items from a set of items for a specific construct, in which their internal consistency has been analyzed (Everitt and Dunn, 2001). Everitt and Dunn (2001) and Esbensen et al. (2002) state that the standard required in EFA comprising of: (1) Eigenvalue ≥ 1, (2) total variance explained ≥ 50%, (3) Kaiser-Meyer-Olkin (KMO) ≥ 0.5 and sig. coefficient of the KMO test ≤ 5%, (4) factor loadings of observed items ≥ 0.5, (5) difference between the loading of two factors greater than 0.3. Fourthly, the study utilizes Confirmatory Factor Analysis (CFA) to further confirm the unidirectionality, scale reliability, convergence value and distinctive value; next the Structural Modelling Equation (SEM) is use to test research hypothesis. According to Hair Jr et al. (2016) and Everitt and Dunn (2001), a model is considered consistent with market data if satisfying some criterion, including: (1) significance value of Chi- square test is less than 5%, (2) CMIN/df ≤ 3 (if CMIN/df ≤ 2 is demonstrated best), (3) GFI, CFI, TLI ≥ 0.9. Furthermore, some scholars illustrate that RMSEA ≤ 0.08, GFI should be greater than 0.8, composite reliability (CR) ≥ 0.6 and the average variance extracted (AVE) ≥ 0.5 (Esbensen et al., 2002, Ho, 2006). Additionally, to further confirm the exactness of estimated results, the study utilizes the boostrap technique, which is suggested by Yung and Bentler (1996). AGU International Journal of Sciences – 2019, Vol. 7 (2), 42 – 55 47 4. RESULTS 4.1 Scale reliability tests Table 4.1 briefly summarizes the results of scale reliability tests through Cronbach’s Alpha coefficients. The overall Cronbach’s Alpha coefficients of all constructs are greater than 0.7 and the corrected item-total correlations are greater than 0.3. Hence, six constructs and relevant items (exempt from "BM_3", which corrected item-total correlations are smaller than 0.3) are accepted and included in the next analysis. Table 4.1. Scale reliability tests Construct Cronbach's Alpha Corporate image 0.815 Security 0.880 Service price 0.810 Customer expectations 0.786 Empathy 0.871 Customer satisfaction 0.862 4.2 Exploratory factor analysis (EFA) 4.2.1 EFA for independent constructs Table 4.2 shows that the items of independent constructs are converged on original five constructs with the total variance explained of 61.05%; KMO = 0.946 with the significance value of 0.000; meaning that, EFA is used appropriately in this study. Additionally, the factor loadings are greater than 0.5 and the weight differences among the loadings are greater than 0.3, demonstrating that these observed variables are acceptable for further analysis (exempt from “VC_7”). Table 4.2. Exploratory factor analysis Observed variables Independent Constructs Corporate image Security Service price Customer expectations Empathy DA_2 0.758 DA_1 0.749 DA_4 0.694 DA_3 0.693 DA_5 0.616 DA_6 0.575 VC_1 0.701 VC_2 0.636 VC_4 0.611 VC_3 0.608 AGU International Journal of Sciences – 2019, Vol. 7 (2), 42 – 55 48 Observed variables Independent Constructs Corporate image Security Service price Customer expectations Empathy VC_5 0.575 VC_6 0.519 VC_7 TL_1 0.694 TL_2 0.685 TL_3 0.679 TL_4 0.631 TL_6 0.592 TL_5 0.578 GDV_1 0.799 GDV_2 0.698 GDV_3 0.670 GDV_4 0.534 BM_4 0.737 BM_1 0.654 BM_2 0.552 BM_5 0.535 Eigenvalues 11.156 1.7071 1.488 1.1142 1.0194 % of variance 41.317 6.3227 5.5114 4.1267 3.7754 Kaiser-Meyer-Olkin Measure of Sampling Adequacy. .946 Bartlett’s Test of Sphericity Approx. Chi-Square 6434.638 df 351 Sig. .000 4.2.2 EFA for “Customer Satisfaction” As shown in Table 4.3, the results of EFA for the scale “Customer Satisfaction” with 5 items illustrates that using EFA in the scale is also appropriate and the items can be used for the next analysis. AGU International Journal of Sciences – 2019, Vol. 7 (2), 42 – 55 49 Table 4.3. EFA for construct “Customer Satisfaction” Customer Satisfaction SHL_3 0.847 SHL_4 0.809 SHL_2 0.798 SHL_5 0.786 SHL_1 0.77 Eigenvalue 3.221 Total variance 64.442 KMOs 0.859 Approx. Chi-Square 982.507 df 10 Sig. 0.000 4.3 Confirmatory factor analysis (CFA) The results from CFA model is assessed via three criteria including unidirectionality, convergence values, composite reliability (CR) and average variance extracted (AVE). Unidirectionality: Figure 4.1 illustrates that CFA model is suitable with the market data. Specifically, Chi-square = 927.132 with 332 degree of freedom; p = 0,000; CMIN/df = 2,793; GFI = 0,875, TLI = 0,9; CFI = 0,912; RMSEA = 0,062. Convergence values: Figure 4.1 also shows that the weights of CFA meet required standards that is greater than 0.5. Table 4.4. Confirmatory factor analysis Scale Notation Numbers of items CR AVE Conclusion Corporate image F1 6 0.865543 0.518827 Accepted Security F2 5 0.849550 0.531231 Accepted Service price F3 6 0.816107 0.525269 Accepted Customer expectations F4 4 0.797496 0.496602 Accepted Empathy F5 4 0.785127 0.528049 Accepted Customer satisfaction Y 3 0.817978 0.500277 Accepted AGU International Journal of Sciences – 2019, Vol. 7 (2), 42 – 55 50 Composite Reliability and Average Variance Extracted: The calculation of CR and AVE in the Table 4.4 indicates that the scales meet criteria in terms of Composite Reliability and Variance Extracted. Figure 4. 1. The CFA results of the saturated model 4.4 Analysis of Structural Equation Modelling The results of SEM in Figure 4.2 show that the proposed research model is consistent with the market data. Particularly, Chi-square = 845.819; df = 306; p = 0,000; CMIN/df = 2.764; GFI = 0.881; TLI = 0.904; CFI = 0.916; RMSEA = 0.062. AGU International Journal of Sciences – 2019, Vol. 7 (2), 42 – 55 51 Figure 4. 2. Standardized SEM Model The results of hypothesis testing shown in Table 4.5 clearly illustrate that 4 out of research hy