THE MULTIPLE MEDIATION ROLES OF TRUST AND SATISFACTION IN THE EFFECT OF PERCEIVED CORPORATE SOCIAL RESPONSIBILITY ON LOYALTY

. Purpose – This study aims at investigating the effect of corporate social responsibility perceptions of local people living in Diyarbakır on bank loyalty in terms of banking services and the mediator roles of trust and satisfaction in this effect. Research methodology – Cronbach’s Alpha coefficients, reliability, convergent and discriminant validity and multicollinearity were examined before testing the hypotheses of the research. The validity of the measurement model for the variables that are included in the model was examined by the confirmatory factor analysis. Partial least squares path modelling was used to analyze the research model and hypotheses, and multiple mediation analysis was used to test the mediation effects. The collected data were analyzed by the software SPSS 22.0 and SmartPLS 3.3.0. Findings – The research results exposed that the corporate social responsibility perceptions of bank customers have significant and positive effects on bank loyalty. Furthermore, satisfaction has had partial mediation effects on the relationship between CSR perception and loyalty. Research limitations – The population covers individuals with regular wage or salary who live in Diyarbakır. The survey was conducted between October 1 and December 1, 2019, with convenience sampling technique and 400 questionnaires were collected. Practical implications – When conducting CSR activities, businesses should emphasize that trust and satisfaction have an essential effect as well as loyalty. Originality/Value – This study investigated the simultaneous multiple mediator effects of trust and satisfaction in the relationship between CSR and loyalty.


Introduction
CSR is an essential issue for those seeking the recipe that today's businesses should implement and gain value in the eyes of customers. It is not easy to understand clearly whether the real intention of the companies is to take part in social projects or increase the sales volume.
Nevertheless, CSR activities should be participated and supported by stakeholders regardless of their intentions. Research on the history of CSR goes back to the 1930s when Berle and Means defended the need for more transparency and accountability for the business's stakeholders.
It can be stated that formal research on social responsibility was handled to a great extent in the 20th century; however, research in this area has increased especially in the last 50 years (Carroll, 1999).
Due to the information age we live in, the level of consciousness of consumers has significantly increased, which leads companies to take action as well. Thus, it has come to the conclusion that companies should gain a new motivation in their marketing efforts. The understanding of social marketing emerged, and the researchers emphasized that companies should not be indifferent to their social environment.
Creating plans and projects prioritizing rural development, taking steps to eliminate inequalities in education, ensuring the integration of disabled citizens with society, raising awareness of the people against infectious diseases and minimizing the social impacts of this disease to the extent possible, and many other social activities can be thought as the examples (Onaran et al., 2013).
Coronavirus , which spread via China in 2019, affected the world in a short time. As of July 2020, more than 10 million people were infected with this disease, and approximately 500000 of them died (Worldometer, 2020). The world is on alert against Coronavirus, a global epidemic. The vaccine development process is progressing rapidly. On the other side of the coin, positive outcomes of this outbreak can be seen. This virus has revealed the spirit of solidarity in people from many different nationalities. Businesses have not been silent despite this big event and have taken important steps for social responsibility. In this context, employers who took their employees on paid leave came to the agenda, and there were also businesses that used the factory band for equipment that should be used to fight disease free of charge. In addition, some sports clubs and businesses have taken steps to support the struggling healthcare workers.
In their study, Skvarciany and Jurevičienė (2017, p. 636) examined the effect of trust in internet banking and emphasized that the characteristics of the bank in Lithuania, risk perception of the customer in Latvia and characteristics of the bank, and the experience of cooperation with a bank of the respondents in Estonia are important in building trust. Besides, bank information in Lithuania, the website in Latvia, and the bank itself and trust in Estonia have been handled, and no findings have been found on the subject of satisfaction and loyalty. Nejjari and Aamoum (2020) presented evidence on which ethics, trust and university image have an important and positive effect on student loyalty. It also proved the impact of ethics and shared values on university image. However, satisfaction was not covered in the study. Tambovceva et al. (2017), Latvia business representatives primarily consider CSR as a corporate image provider. In addition, they are aware of the role of CSR in improving employee and customer satisfaction. Uhlig et al. (2020) concluded that the philanthropy dimension, one of the dimensions of CSR, has an effect on the intention to deal with socially responsible companies. Iglesias et al. (2018) showed that CSR affects customer loyalty both directly and indirectly through co-creation and customer trust, and concluded that co-creation has a direct effect on customer trust. In these studies, the issues of trust, satisfaction and loyalty are not sufficiently addressed. Therefore, in this study, it was tried closing this gap in the literature by looking at the role of trust and satisfaction in the relationship between CSR and loyalty.
This study investigated the effect of CSR perceptions of the people living in Diyarbakır regarding the banking service they received on their loyalty was examined. In addition, the effects of trust and satisfaction on this effect on loyalty have been investigated. In this context, data from 400 people were collected from people living in Diyarbakır according to the convenience sampling method.
In the study, first, a wide literature review was done on the subject. In the literature review, studies examining the relationships between CSR perception and customer loyalty were included. Then, the literature on customer trust and satisfaction was examined, and hypotheses were created accordingly. Next, the method section was started, and the model was created. Following the model, statistical analysis for the research was included, and the findings were interpreted. In the conclusion part, research-specific findings and recommendations were presented.

Literature review and research hypotheses
It is stated that the sustainability issue has three dimensions in general. These are environmental sustainability, economic sustainability and social sustainability. Although social sustainability is mentioned less than other dimensions, it actually has a very important place. Likewise, it covers vital issues such as humanitarian standards, equality, justice, community development, humanitarian support and assistance. Social sustainability overlaps with the goals of corporate social responsibility. It also addresses the welfare of humanity and society through the management of social resources (Govindan et al., 2020). Increased awareness of sustainability and CSR has increased the interest of human resources in efforts for a cleaner business (Amrutha & Geetha, 2020). CSR has very long and different historical backgrounds. Indeed, CSR history can go back to the history of concern of society from business communities. In addition, official writings on social responsibility are largely a product of the 20th century, especially the last 50 years. It is possible to see the footsteps of CSR thought worldwide, mostly in developing countries. However, most of the official sources state that CSR practices are evident in the USA, where the literature is accumulated (Carroll, 1999). As a social structure, CSR is voluntary participation to fulfil its responsibilities toward the institutions' elements and solve some social problems (de Quevedo-Puente et al., 2007;Brammer & Pavelin, 2006;Cingöz & Akdoğan, 2012).
All these crises motivate companies to be more social and create social value. CSR activities enable many multinational companies to lead corporate strategies. For example, Unilever's Paul Polman claims that businesses can be a positive force for good in the world, and this perspective is in the interest of all company stakeholders (van Doorn et al., 2017). Therefore, creating something useful for humanity does not only bring material gain to businesses but is also important for motivating the shareholders.
Having an important place in the development of the concept of CSR, Carroll examined social responsibility in four dimensions and defined the social responsibility of the companies as "the sum of the economic, legal, ethical and voluntary responsibility activities that the society expects from companies at a certain time" Carroll (1979). These responsibility activities are as follows: Economic responsibilities: In the classical approach, economic responsibilities are accepted as the most important and fundamental responsibility of the business, both in the modern and postmodern periods. Economic responsibilities can be regarded as the most important existence causes of businesses, as they act with the motive of making a profit. It is not possible for a company that does not have any financial gain to survive (Torlak, 2007).
Legal Responsibilities: The legal responsibilities of companies cover issues such as compliance with the law, compliance with central and local regulations, being a law-abiding citizen and fulfilling the legal obligations of companies (Carrol, 1991).
Ethical Responsibilities: The enterprise should take precautions to minimize the potential harm to society and environment while performing its own goals and activities. In addition, to gain the trust of its stakeholders, it should determine the ethical performance conditions in all its activities and manage the process with this reference (Yakar & Özhan, 2017).
Discretionary Responsibilities: These responsibilities include acting in line with the charitable and generous expectations of society, supporting the arts, participating in volitional and charitable activities in their own regions, providing assistance to private and public educational institutions, and voluntarily supporting projects that contribute to the quality of life of the society (Carroll, 1991).

The link between corporate social responsibility and customer loyalty
The banking sector has a critical place in meeting the demand for funds to implement the increasing needs of individuals today. In this regard, the number of bank branches is increasing rapidly. In Turkey, the banking system, as of June 19, 2020, has 48 banks in operation; the number of domestic branches of these banks is 10079, and the number of branches abroad is 71 (The Banks Association of Turkey, 2020). When one looks at the activities of banks, it is seen that the environment they affect does not cover only bank owners, employees and customers but also they interact with many different institutions and organizations and can directly or indirectly affect the commercial activities of these institutions. Furthermore, in the postmodern period, all people on a global scale need to make bank transactions. It is not possible for such a huge sector to attract customers only with its financial instruments. Especially in parallel with the development of consumer awareness, the sensitivity of banks toward social responsibility has become more critical.
In the research carried out by Özmen (2009) on bank customers, the effect of various CSR activities on the overall brand image was examined. Accordingly, it was concluded that there is a relationship between the social responsibility activities related to the social environment, consumption and consumer, the environment, health, economic development, and sports and cultural activities and the brand image and these activities have a positive effect on the brand image.
In their study, Doğan and Varinli (2010) examined the relationships between social responsibility understanding and corporate image of bank customers. According to the findings obtained from the study conducted with 537 participants, it was concluded that there is a relationship between the corporate image perceived by the customer, social responsibility activities and the brand image, and these activities affect the brand image positively.
In the study by Özgüven (2013), the impact of the concept of CSR on the banking sector was examined. According to this research, it was concluded that social responsibility activities positively affect the customers' trust in the bank and customer satisfaction. Accordingly, it also positively affected customer loyalty.
Respect for employees, philanthropic activities and respect for consumers, which constitute the perception of CSR, have had a significant impact on brand awareness, quality and brand loyalty, which are components of consumer-based brand equity (Koçoğlu, 2016). Chen, Chang, and Lin (2012) have examined whether CSR impacts customer loyalty in their study on Taiwan airlines. Hence, it has been observed that CSR has an important and positive effect on the formation of behavioural and attitudinal commitment.
In a study on mobile telecommunication service, it was determined that CSR has a direct relationship with satisfaction and customer loyalty and satisfaction have a mediating effect (He & Li, 2011). Köse (2015) concluded that the economic, ethical, legal and charitable sense of responsibility, which is one of the dimensions of CSR, is effective in creating customer loyalty and customer satisfaction. In addition, Chung, Yu, Choi, and Shin (2015) concluded that CSR positively influences customer loyalty. Similarly, Fatma and Rahman (2016) concluded that CSR is effective on the intention to repurchase in their study in the banking sector in India. Skvarciany and Jurevičienė (2017) conducted research in Baltic countries to evaluate the determinants of trust in traditional banking. Hence, it has been emphasized that the characteristics of the bank in Lithuania, the risk perception of the customer in Latvia and the characteristic of the bank, and the cooperation experience of respondents in Estonia with a bank are essential in building trust. In another study, Skvarciany and Jurevičienė (2018) investigated the effect of trust in internet banking. Based on the views of 1013 participants and 21 experts, country-based evaluations were made as factors necessary for trust-building. Thus, it has been observed that bank information in Lithuania, website in Latvia and bank factors in Estonia are effective in building trust.
The study of Nejjari and Aamoum (2020) shows how the image of the university, ethical rules and common values of graduates positively affect trust affecting loyalty. According to the results of the research, ethics, trust and university image have an important and positive effect on student loyalty. It also proves the impact of ethics and shared values on university image. Tambovceva et al. (2017) conducted a survey of Latvian business representatives. Latvian business representatives see CSR primarily as a corporate image provider. In addition, they are aware of the role of CSR in improving employee and customer satisfaction. Uhlig et al. (2020) aimed to learn the relationship intention of Brazilian consumers. Accordingly, it was concluded that the philanthropy dimension, one of the dimensions of CSR, has an effect on the intention of dealing with socially responsible companies. Iglesias et al. (2018) examined the impact of CSR on customer loyalty, taking into account the roles of co-creation and customer trust. At the same time, the effect of co-creation on customer confidence has been explored. In this study conducted in Spain, it was determined that CSR affects customer loyalty both directly and indirectly through co-creation and customer trust. It was also concluded that co-creation has a direct impact on customer confidence. In line with the information obtained, as seen in Figure 1, hypotheses from H 1 to H 3 were formed as follows: H 1 : Corporate social responsibility positively affects customer loyalty.

The mediating role of customer trust in the corporate social responsibility-customer loyalty link
In the study by Özgüven (2013), the impact of the concept of CSR on the banking sector was examined. According to this research, social responsibility activities positively affected the customers' trust toward the bank and customer satisfaction. In another study, it was determined that there was a significant and positive relationship between the perceived corporate social responsibilities of the customers and the corporate image, customer loyalty and customer trust (Saraçoğlu, 2013). On the other hand, Kim et al. (2008) examined brand equivalence within the scope of customer relations management through an application they performed on hospitals and put forward that 5 factors affect brand equivalence through successful customer relations. Thus: trust, customer satisfaction, relational loyalty, brand loyalty and brand awareness are the factors that help hospitals create brand equity and can be managed through customer relations. The findings also indicate that hospitals would be successful in obtaining an image and creating positive brand equity if they manage customer relations well (Kim et al., 2008). Uslu et al. (2020) found that the brand equity dimensions of hotel businesses have a mediator effect of trust on customer satisfaction. On the other hand, Diallo and Lambey-Checchin (2016) and Afridi et al. (2018) found the mediator effect of customer trust between CSR and customer loyalty in their study. Hence, we formed the following hypothesis: H 2 : Customer trust has a positive mediate effect on the relationship between corporate social responsibility and customer loyalty.

The mediating role of customer satisfaction on the corporate social responsibilitycustomer loyalty link
In the research conducted in the durable consumer goods sector by Eren and Eker (2012), they found that the corporate social responsibility activities that they have implemented in the companies significantly affect the brand image and perceived value. In addition, in this research, it was determined that the brand image has an impact on the perceived value, customer satisfaction and brand loyalty, and the perceived value has an effect on customer satisfaction. In addition, Hoeffler and Keller (2002) concluded that CSR phenomenon creates brand awareness, brand image, trust in the brand and loyalty to the brand.
There are studies in the literature that determine the mediating effect of customer satisfaction between CSR and customer loyalty, though in a limited number (Isa, 2012;Rivera et al., 2019;Latif et al., 2020). Thus, we constructed the hypotheses: H 3 : Customer satisfaction positively mediate the relationship between corporate social responsibility and customer loyalty 2. Method

Research model
It has been determined that customer satisfaction has a mediating effect between CSR and brand loyalty (Rivera et al., 2019). Afridi et al. (2018) found the mediator effect of customer trust between CSR and customer loyalty in their study. Are there multiple mediation effects of customer satisfaction and customer trust in the relationship between the Perception of CSR and Customer loyalty? Based on the research question, as can be seen in Figure 1, the research model was created, and the analysis was carried out. In the context of the hypotheses established in line with the literature in the research, the research model was created as in Figure 1 below to test the causal structures.

Purpose of the study and variables
The purpose of this study is to investigate the effects of corporate social responsibility (CSR) perceptions of people with a certain wage or salary who are living in Diyarbakır on the banks they receive service on trust, satisfaction and loyalty. Furthermore, we aimed at determining the mediation role of relationship quality (trust and satisfaction) in the relationship between CSR and loyalty. The fact that the mediation role of both variables was examined by multiple mediation analysis is considered as a unique approach. In this framework, the study of Swaen and Chumpitaz (2008), which includes 18 expressions from the scales used in previous studies, was taken as a basis to measure CSR dimensions. Five expressions to measure the variable of trust and 3 expressions to measure customer satisfaction, which are the sub-dimensions of relationship quality, were developed using the Aydın and Özer (2005) study.

Population and sampling
The population of this study is the people having a certain wage or salary who are living in Diyarbakır. According to the 2019 census of the population by Turkey Statistical Institute (TSI), the province of Diyarbakir consists of 1,756,353 people. Sample size calculated for this population calls for at least 384 individuals (Cohen et al., 2017). In this respect, we assumed that the 400 questionnaires collected for our study have a sample size that can represent the population.

Data collection and analysis
In order to collect the research data, a questionnaire consisting of pre-prepared close-ended questions were used. The survey comprises three parts. In the first part, gender, age, marital status and monthly income questions were included in determining the demographic characteristics of the local people. In the second part, there are 3 questions asking for information about the bank from which the service is received. In the third part, there are 30 questions that measure the variables of the research (CSR, trust, satisfaction and loyalty). Respondents were asked to answer the questionnaire using the 7-point Likert scale (1 -Strongly Disagree, 2 -Disagree, 3 -Partially Disagree, 4 -Neither agree nor disagree, 5 -Partially Agree, 6 -Agree, 7 -Strongly Agree).
During the survey application period, a total of 500 questionnaires were distributed in October-December 2019 to be made to those who live in Diyarbakır and have a regular salary or wage regularly received from the bank. The surveys were applied to people visiting historical and touristic places such as Ulu Camii (Great Mosque), Diyarbakır Surları (city walls), On gözlü (ten eyed) bridge located in the city centre of Diyarbakır. The surveys were conducted by one of the authors and 3 interviewers (students) who knew the purpose of the research in this region, which is the city centre and one of the most crowded places in the city. The survey was carried out on the basis of voluntary participation according to the convenience sampling method. As a result, 420 were returned from 500 respondents, which means 84% of the distributed questionnaires. Twenty unusable questionnaires that were unusable and left blank were excluded from the study, and the remaining 400 questionnaires were evaluated.
In this study, it is aimed to investigate the trust, satisfaction and loyalty relationships of corporate social responsibility (CSR) perceptions of banks, where the participants receive a salary or other banking services (credit, credit card, etc.). PLS-SEM (partial least squares structural equation model) was used to determine the role of relationship quality (trust and satisfaction) in the relationship between CSR and loyalty. Validity, reliability and confirmatory factor analysis (CFA) were performed using the partial least squares measurement model. Then, in the structural model, all hypotheses of the research were tested by conducting path analysis and mediation test. SPSS 22.0 and SmartPLS 3.3.2 statistical packages were used during the analysis of the data.

The measurement model
Prior to the estimation of the model, the scales' psychometric properties were tested for reliability, discriminant and convergent validity and also multicollinearity.

Reliability, discriminant and convergent validity
Before the research model was analyzed, the validity and reliability of the research variables were performed with the measurement model with the purpose of testing the validity and reliability of the measurement model. In this context, internal consistency reliability, convergent validity and discriminant validity were checked out. For internal consistency reliability, Cronbach's Alpha and Composite Reliability values were examined. In determining the convergent validity, factor loads, and average variance extracted values were used. Factor loads, CR values and Cronbach's Alpha (α) coefficients are required to be over 0.70 while the average variance extracted value is expected to be over 0.50 (Fornell & Larcker, 1981;Bagozzi & Yi, 1988;Hair et al., 2017). Table 1 shows the results regarding internal consistency, reliability, convergent validity, and Cronbach's Alpha (α) coefficients.
We have stated that Cronbach's Alpha (α) coefficients and CR values should be greater than 0.70 to ensure reliability and internal consistency (Nunnally, 1978;Gefen et al., 2000). Cronbach's Alpha (α) coefficients of the variables used in the research were between 0.953 and 0.912, and CR values are between 0.962 and 0.943. It can then be stated that internal consistency reliability is provided. The rho_A coefficients that show whether the expressions in the factors are reliable were found above 0.70 for all factors. The rho_A coefficients have recently been reported in the literature to support the Cronbach's Alpha coefficient and are more reliable than the Cronbach's Alpha value (Henseler et al., 2016). It can be stated that due to factor loadings of all variables between 0.938 and 0.646 and AVE values between 0.865 and 0.558, convergent validity is provided.
Fornell and Larcker Criteria and HTMT (Heterotrait-Monotrait Ratios) criteria were used to determine the discriminant validity (Fornell & Larcker, 1981;Chin, 2010;Hair et al., 2014;Henseler et al., 2015). According to the Fornell and Larcker (1981) criteria presented in Table 2, the square root (bold values) of AVE values should be higher than the correlation values between the other structures in the study (Fornell & Larcker, 1981;Chin, 2010). The results presented in Table 2, according to Fornell and Larcker criteria show that the discriminant validity is provided. Another way to check discriminant validity is HTMT rates. If these rates are less than 0.90, it is considered acceptable (Gold et al., 2001;Henseler et al., 2015). As shown in Table 2, it is understood that the HTMT ratios obtained are below the threshold value of 0.90 and that the model provides the discriminant validity.
The goodness of fit values is not required when using PLS-SEM (Henseler, 2018). However, Henseler et al. (2014) suggested that the value of SRMR (Standardized Root Mean Square Residual) can be used to determine the model-data fit in PLS-SEM. The fact that the SRMR value is at a sufficient level such as 0.062 and that the SRMR <0.08 (Hu & Bentler, 1999) condition is met in the research shows that the model-data fit is sufficient. As a result, The bank from which I receive services develops projects for poor countries. The bank that I receive service supports philanthropy.

36.511
The bank from which I receive service reduces the consumption of natural resources. The bank from which I receive service carries out environmentally friendly production practices.

38.111
The bank from which I receive service maintains an ecological balance as much as possible.

37.159
The bank from which I receive service protects consumer rights. (postloan services etc.

Multicollinearity
An indicator's information can become redundant due to high levels of multicollinearity. The variance inflation factor (VIF) is identified in the literature as a good indicator to determine whether there is multicollinearity (Kock, 2015). When the variance inflation factor (VIF) values are less than 5, the multicollinearity problem is not expected . The variance inflation factor (VIF) values in Table 2 that it is much lower than 5 points out that there is no multicollinearity problem for the research.

Findings
This part of the study includes the demographic characteristics of the local population, descriptive statistics, correlation coefficients and testing the hypotheses formed by the structural equation model.

Demographic characteristics of the local population
It is determined that 65.3% (261 people) of the respondents were male and 34.8% (139 people) were female; 57.5% (230 people) of the respondents married and 42.5% (170 people) single. It is seen that more than half of the respondents (51.7% = 207 people) are in the age range of 25-34, and 27.0% (108 people) in the age range of 35-44. Monthly income levels range from 3501-4500 TL for 42.5% (170 people), and 4501-5500 TL for 38.5% (154 people). The local people responding to the research are mostly male, married, middle-aged young people with a monthly income between 3500 TL and 5500 TL (Table 3).

Descriptives and correlation coefficients
When Table 4 is examined, it is seen that the variable with the highest arithmetic average x = 4.35 according to the perceptions of local people about banking services is the variable of trust. Loyalty was found to have an arithmetic mean of x = 4.28. It is seen that the lowest arithmetic average is the perceived corporate social responsibility dimension with x = 3.96. It can be stated that the respondents generally agreed with low arithmetic mean for the scale questions. The correlation coefficients were found to be between ±0.3 and ±0.9 values (Raykov & Marcoulides, 2006). This shows that there are significant and positive relationships between variables.

Testing the research model and results
Partial least-squares path analysis (PLS-SEM) was used to analyze the structural equation model. For the research model, the PLS algorithm was used in calculating linearity, path coefficients and R 2 . In the calculation of t-values used to evaluate the significance of the PLS path coefficients, 5000 subsamples (resamples) were taken from the sample and bootstrapping analysis was performed. The results of the structural equation model formed to test the research hypotheses are shown in Figure 2 and Table 5. This model shows whether established relationships are supported by the model. Table 5 also shows the explained variance (explained variance) (R 2 ) values, path coefficients and t-values obtained by bootstrapping method for endogenous variables.   Table 5 includes the direct paths for both models A and B. Model A solely comprises the corporate social responsibility-customer loyalty direct link. When Figure 2 and Table 5 are examined, it was determined that perceived corporate social responsibility according to model A affects customer loyalty (β = 0.666; t-value = 23.681; p < 0.001). However, according to Model B, it was determined that customers' perceived corporate social responsibility of banks do not affect customer loyalty (β = 0.027; t-value = 0.560; p > 0.575) along with the mediation effects of customer trust and satisfaction. H 1 hypothesis for this purpose is supported. H 1 hypothesis was supported because the p-value of the standardized (β) coefficient between perceived corporate social responsibility and customer loyalty, according to Model A is less than 0.05. On the other hand, considering the R 2 value of Model A, it was found that the perceived corporate social responsibility explained customer loyalty at a moderate level of 44% (Henseler et al., 2014). When the R 2 values are examined according to Model B, it is determined that perceived corporate social responsibility explains customer loyalty by 72%, customer satisfaction by 54% and customer trust by 60% (Henseler et al., 2014).

Multiple mediation analysis
In the model built with PLS, besides the PLS-SEM method, the mediator effect between variables can also be tested. Recently, the bootstrapping method has been used in PLS, which suggests that it provides more reliable results than the classical mediation method of Baron and Kenny (1986) in mediation variable analysis (Hayes, 2018). Therefore, in this study, as shown in Table 6, bootstrapping method (Williams & MacKinnon, 2008) and VAF values (Nitzl et al., 2016) were used to test the mediation hypotheses (H 2 and H 3 ).
Chin (2010) suggests a two-step procedure for assessing indirect effects on PLS. The first step deals with using the specific model in question, including both direct and indirect paths, performing N-bootstrap resampling and finally multiplying the direct paths that make up the indirect path under evaluation. The second step is the estimation of significance and the size of the indirect effects in relation to the total effect, through the assessment of the variance accounted for (VAF). In this case, it is possible to determine to what extent the variance of the dependent variable can be indirectly explained through mediator variables. As seen in Figure 2, (a) is the path value between the independent variable and the mediator variable, while (b) is the path value between the mediator variable and the dependent variable. In the calculation of VAF, the formulae of VAF = a × b / (a × b + c) is used (Nitzl et al., 2016). On the other hand, c value is the value of the path coefficient between perceived corporate social responsibility and customer loyalty. In single-step multiple mediator models, c value can be calculated using the formulae of c= c' +a 1 b 1 + a 2 b 2 (VanderWeele & Vansteelandt, 2014). If VAF values are below 20%, there is a zero mediator effect (there is no mediation), whereas 20% to 80% of the VAF value means partial, and more than 80% means the full mediator effect (Hair et al., 2017).
When the VAF values in Table 6 are examined, it is determined that the calculated VAF value is 19% and there is no mediating effect between perceived corporate social responsibility and customer loyalty to determine the mediator effect of trust. Therefore, the H 2 hypothesis was not supported. On the other hand, when the VAF value calculated to determine the mediating effect of customer satisfaction between perceived corporate social responsibility and customer loyalty was determined, it was determined that it was calculated as 42% and partial mediation effect was supported that means the H 3 hypothesis was also supported. Considering VAF is over 60% measuring the total effect of both customer satisfaction and trust together points out the joint indirect effect of SAT and TRU partial mediate the influence of CSR on LOY link.

Discussion and conclusions
CSR finds a significant place on the agenda of all companies that aim to prioritize community interests. The trio of price-quality-presentation that determines the customer value in the banking sector, which is mentioned more frequently with its increasing business volumes and the harsh competition conditions among themselves, is homogenized similar to other sectors. Therefore, the factor that brings a bank to the forefront in the eyes of the customer is its capacity to show itself in CSR activities.
Banks can ensure customer loyalty with CSR, but if it is determined which other variables are effective in the formation of this loyalty, customer loyalty will be achieved more easily. As a matter of fact, it is assumed that the variables of customer trust and satisfaction, which are considered effective in the formation of this loyalty, will have mediating effects. Thus, the purpose of this research has been to investigate the relationship between customer loyalty and customer loyalty of banks benefiting from banking services, and the mediating role of customer trust and satisfaction in this relationship. It is expected that this study makes an important contribution to the literature in terms of the city where the research was conducted and the variables it deals with. As a matter of fact, the meaningful results obtained in the study provide important clues for professionals in the banking sector.
The results of this research revealed that perceived corporate social responsibility has effects on customer loyalty. Studies by Chen et al. (2012), Özgüven (2013), Köse (2015), Chung et al. (2015), Fatma and Rahman (2016), Gültekin (2018) have similar results. In addition, this research determined that customer satisfaction has mediating role between perceived CSR and customer loyalty. It is seen that this result shows parallel results with the study conducted by Isa (2012), Rivera et al. (2019), Latif et al. (2020).
Consequently, the better the CSR perceived by customers regarding banking services, the better the satisfaction towards the bank is. Perceived CSR is associated with customer loyalty. With this study, it is revealed that the bank loyalty of customers is due to perceived CSR and if the customers are satisfied with banks, the loyalty of the bank increases. According to another important result obtained in the study, it was determined that customer trust does not play a mediating role between customers' CSR perceptions and customer loyalty. This result does not match the results obtained in the studies of Diallo and Lambey-Checchin (2016) and Afridi et al. (2018).
It was assumed that customer trust and customer satisfaction would have a mediator effect on customer loyalty. Likewise, the effect of trust and satisfaction on loyalty is known. As a matter of fact, Nejjari and Aamoum (2020) study show that ethics, trust and university image have a significant and positive effect on student loyalty. Also, Iglesias et al. (2018) show that CSR affects customer loyalty both directly and indirectly through co-creation and customer trust in a study examining the effect of CSR on customer loyalty by considering the roles of co-creation and customer trust. In these studies, strong links between trust and loyalty were examined, and determinations were made based on clear findings. However, when looking at the literature, no study has been found that the relationship between CSR and customer loyalty would affect customer trust and customer satisfaction together. In this respect, the study has made an important contribution to the literature.
Banks must provide enlightening information in determining customer-oriented policies. Similarly, banks should explain themselves well to customers regarding CSR. Indeed, it is very well known that CSR practices are essential for the client both in this research and in the literature. In this way, it is possible to establish trust and loyalty-based relationship.
In order for loyalty to occur, a sense of satisfaction must develop. This will enable potential or real customers to become loyal customers. In addition, trust and satisfaction together create a mediating effect in the building of loyalty and its relationship with CSR, revealing that being a reliable business and a satisfactory role is a determinant. Likewise, in today's world, where trust in CSR activities has been shaken to a great extent, it will be important to build trust by focusing on policies that prioritize CSR's philanthropy dimension.
Nowadays, while purchasing a product of a company, consumers with a sense of responsibility make their choices by considering criteria such as the quality and price of the product, as well as how effective the company is in social issues. This is one of the most important factors that push businesses to corporate social responsibility efforts. Businesses that can effectively implement corporate social responsibility practices are generally perceived more positively by their consumers. It is necessary to motivate the customers to ensure that their loyalty to the bank is based on CSR practices. In fact, sharing the gains with society prevent economic crises globally. In this way, the trust of the customers toward the bank is ensured, and even conflicts between classes are prevented. Customers would respect the bank, and the bank would see customers as a "value" rather than a "customer" aspect.

Theoretical and managerial ımplications and limitations
The theoretical contribution requires specific kinds of research findings that can present novel insights into a phenomenon that is deemed important for improving CSR on banking industries. Our study offers an original insight based on the empirical data on the perception of CSR, customer satisfaction, customer trust and customer loyalty, as well as various contributions to professionals and researchers. It thereby contributes by determining between the perception of CSR and customer loyalty with multiple mediating roles of customer satisfaction and customer trust. Hence, this study expanded research on customer relationship by examining how CSR, customer satisfaction and customer trust determine the customer loyalty of the banking sector. Moreover, this study identifies how large banking organizations control their perception of CSR, customer satisfaction and customer trust in achieving customer loyalty.
In addition to its theoretical contributions, this research has important implications for professionals. First, if professionals want to increase customer loyalty derived from their CSR activities, they need to link with customers. The results of our study build on this literature by suggesting that managers should embrace CSR to increase customer trust and customer satisfaction into enhanced customer loyalty.
The research has some constraints. The main constraint of the research is that the data collected from the local people is of only the city of Diyarbakır, and the results of the research cannot be generalized. In future studies, it is evaluated that people from different cities should also be involved so that comparisons can be made, and the research model can be enriched with variables such as perceived quality and price. On the other hand, it can be considered that this study, which is generally carried out for the banking sector, should be done separately for public, private and participation banks customers. At the same time, more holistic results can be obtained by conducting in-depth research with qualitative research methods.