The Effect of Emotional Intelligence on Turnover Intention and the Moderating Role of Perceived Organizational Support: Evidence from the Banking Industry of Vietnam

The objective of this study is to investigate the impact of emotional intelligence on turnover intention, noting the mediating roles of work-family conflict and job burnout as well as the moderating effect of perceived organizational support. Survey data collected from 722 employees at banks in Vietnam was analyzed to provide evidence. Results from the partial least squares structural equation modeling (PLS-SEM) using the SmartPLS 3.0 program indicated that there was a negative effect of emotional intelligence on employees’ turnover intention; this was mediated partially through work-family conflict and job burnout. Besides, this study indicated that perceived organizational support could decrease work-family conflict, job burnout and turnover intention of employees. It could also moderate the relationship between emotional intelligence and work-family conflict. This negative relationship was stronger for employees who work in a supportive environment. The main findings of this research provided some empirical implications for the Vietnamese banking industry. It implied that organizations in the service industry should try to improve their employees’ work-family balance, reduce job burnout and take advantage of these emotional balances and supportive environments to create beneficial outcomes.


Introduction
In recent years, the notion of turnover intention has become a central attraction of research in various fields of managerial disciplines, therefore, prompting more and more executives/managers in different business sectors to take this concept into consideration to effectively manage their employees [1]. Employee turnover can have a variety of negative effects on organizations. For example, according to Allen et al. [2], replacing an employee can cost more than a year's salary for the position being filled. Hausknecht and Trevor [3] noted that turnover results in the loss of both social and human capital. In human services organizations, those negative effects can extend to the clients and the organization it serves. In the 21st century, the concept of job turnover intention is considered as a key concept in management for a successful career of an employee and business survival at all levels, predominantly in the service industry [4,5]. The particular characteristic of a service industry is "the contact and interaction between service providers (employees) and service acceptors (customers)" that infers that employees in the service industry have become a part of service products. Being a service ambassador, employees play an important role to satisfy customers with proper services. Therefore, understanding employees' emotions and feelings toward work is another recent concern that organizations are struggling to manage. The issue, then, is how individual employees manage their emotions effectively and lessen job turnover intention. Judeh [6] stated that emotional intelligence is a significant factor in interpreting and analyzing human behavior at work. Most successful service organizations understand their people's feelings and always have special concerns about any issues that may threaten the employees. However, it is broadly termed emotional intelligence (EQ) that enables an employee to work together with others toward achieving a common goal. In addition, emotional intelligence will enhance employees' suitable emotions in fulfilling customers' expectations that help to form a positive image of the organization [4,7].
Since joining the World Trade Organization (WTO) in 2007, Vietnam has taken part in significant economic growth in the last decade due to the improvement of the business environment and the high volume of foreign investments. Along with the achievement in the economy, the Vietnamese banking industry has also seen outstanding enlargement and development in terms of scale and service quality [8]. However, the increases in competition and financial innovation have led to extremely rapid expansion which has resulted in banking system problems in the past few years. Eventually, the government-initiated project 254 called "restructuring credit institutions system 2011-2015" was implemented by the Prime Minister in 2012. In response, the State Bank of Vietnam has pushed the commercial banks to merge with the project goal and reduce half of the number of banks. In such an unfavorable condition, employees in this sector have experienced a considerable fluctuation with the highest average voluntary turnover rate compared to other service sectors. As of 2019, the turnover rate was more than 25% each year [9] and 81% for turnover intention [10]. Restructuring the banking sector caused not only workforce transition among banks, but also between banking and non-banking sectors. High turnover can lead to compromised client care and increased organizational costs [11]. Service organizations, such as the banking industry, have an added burden with turnover given that they foster individual relationships between their counselors and their clients. When an employee leaves such an organization, a personal relationship with clients is broken. Moreover, the bank incurs the expense of finding and training new employees, and clients can suffer due to discontinuity of care [12]. Additionally, turnover results in decreased efficiency in carrying out job-related responsibilities [13]. Implementing new strategies and achieving organizational goals is hampered by high employee turnover [12]. Consequently, keeping intellectual capital for organizations is a crucial strategy for maintaining sustainable development.
With an increasing interest in employee turnover intention, numerous researches have been conducted to find out the predictors of employee turnover intention and discovered several factors including emotional intelligence (e.g., [14]), work-life conflict (e.g., [15]), job burnout (e.g., [16]), and a few others (e.g., [14,16]). Research has shown that emotional intelligence helps the employees to have control of their emotions and to understand the use of social skills that are helpful to the organization [17]. Emotional intelligence in individuals who lead can help organizations to reduce turnover intentions and retain valuable employees [18]. Avey et al. [19] asserted that emotional intelligence is one of the key factors that affect the turnover intention of an employee among other factors. Besides, prior studies have also confirmed that emotional intelligence strongly affects employees who intend to leave their jobs by causing work-family conflict [20] and job burnout [16]. In general, people's emotions change quickly based on the challenges in the workplace and also in family life. The work-family balance will enhance their suitable emotions in responding to customers' expectations that help to form a positive image of their organizations [7]. Hence, understanding an employee's emotions and feelings is a crucial issue for both employees and the organization in the service industry. Moreover, employees are Electronic copy available at: https://ssrn.com/abstract=3678384 among a company's most valuable assets. Therefore, employers need to understand burnout and its causes. Lu and Gursoy [16] pointed out that burnout is one of the best predictors of turnover intention because employees in customer service-based industries have to deal with customer demands, so they are at high risk for job burnout. They also noted that burnout is costly for organizations on two fronts, not only leading to higher turnover rates but also decreasing worker productivity.
Additionally, although there are tangible factors in work environments that have been described as antecedents to work-family conflict and job burnout, theory and empirical findings indicate that the nature of the work environment as sensed by the worker, that is, perceived organizational support, may be an important determinant of work-family conflict and job burnout in employees who work in the banking industry. Perceived organizational support refers to the general belief by an employee that support will be readily available from the organization when stressful situations arise and urgent needs are addressed [21]. Theorists posit that the availability of material aid and emotional support may reduce aversive psychological and psychosomatic reactions (e.g., emotional exhaustion) when stressful situations arise [22]. In supportive work environments, employees treat each other with fairness and respect and engage in effective open communication. These characteristics contribute to win-win solutions for the employee and the organization [21]. Thus, it is plausible that employees who perceive their work environments as supportive will experience work-family conflict and job burnout, and, conversely, bankers who perceive their work environments as unsupportive will experience more job burnout and work-family conflict.
In developed countries, much research has been conducted about the perceptions of subordinates in terms of emotional intelligence, perceived organizational support, work-family conflict, job burnout, and employee turnover intention and their significant mutual relationships [20,23], however, this topic has yet to receive considerable attention from scholars in developing or less developed countries, particularly in Vietnam. An extensive literature review informed that there has been no research conducted to examine the relationship between emotional intelligence and turnover intention of employees in Vietnamese organizations. Moreover, there is no empirical evidence that confirms the relationships of emotional intelligence, perceived organizational support, work-family conflict, job burnout, and turnover intention in one model. Specific to the financial sector, a banker with high emotional intelligence would perform effectively at enhancing customer enthusiasm and reducing customer frustration [24]. So, the need for understanding employee emotional intelligence is strongly considered in this research. Therefore, these twofold gaps have led the researchers to conduct this research to examine the mutual effects of emotional intelligence and perceived organizational support on work-family conflict and job burnout and subsequently explore its impact on the turnover intention of bankers in Vietnamese commercial banks. Besides, the moderating role of perceived organizational support on these relationships will be considered as well.

Emotional Intelligence
The Emotional Intelligence theory of Bar-On was first introduced in 1985 and was followed by a series of other subsequent developed versions (e.g., [25,26]). Emotional intelligence was found as having an important role in both building and maintaining successful social relationships, predicting specific aspects of situations involving social exchange, and it will either foster or hamper this exchange process [27]. Salovey and Mayer [28] stated that emotional intelligence is "the subset of social intelligence that involves the ability to monitor one's own and others' feelings and emotions, to discriminate among them and to use this information to one's thinking and actions". Besides, Ravichandran et al. [29] defined emotional intelligence as "the ability to recognize, understand, and assess one's own feelings as well as others and use this knowledge in thought and action". In line with this definition, Serrat [30] indicated that "emotional intelligence describes the ability, capacity, skill, or self-perceived ability to identify, assess, and manage the emotions of one's self, of others, Electronic copy available at: https://ssrn.com/abstract=3678384 and of groups". Possessing high emotional intelligence enables an individual employee to be more productive, optimistic, and resilient in both work and family life [31]. Mayer and Salovey [25] have established the scope of emotional intelligence into four dimensions including: (1) The self-emotions appraisal (SEA): This dimension reflects the ability of a person to understand his/her own emotions and be able to express them properly, then apply the knowledge of those emotions to create beneficial outcomes. (2) The other emotions appraisal (OEA): This component assesses the ability of an individual to observe and understand other's emotions. A person who has high capability in this dimension will be able to observe other people's emotions and predict other's emotional reactions; (3) The use of emotion (UOE): This aspect evaluates the ability of an individual to access, generate and use his/her emotions to facilitate personal performance. People who rate highly in this ability will be able to return rapidly to normal psychological states after suffering depression or feeling upset; (4) The regulation of emotion (ROE): This dimension mentions the ability of an individual to regulate his/her emotions to achieve an expected outcome and be able to remain balanced from psychological distress to solve problems.

Turnover Intention
Turnover intention is defined when an individual intends to leave their organization [32,33]. It is inevitable that turnover occurs in every organization in which some of the employees voluntarily leave the organization while others are discharged by the organization. Turnover is classified into two types, i.e., voluntary and involuntary turnover [34]. First, voluntary turnover happens when employees are not fulfilled with their current job and are ready to look for another job in another place. In other words, it is called turnover intention that refers to the desire to voluntarily leave an organization. Second, involuntary turnover happens when employees are fired by their organization. Either voluntary turnover or involuntary turnover creates serious consequences for the organization in today's business world including a number of difficulties in finding a replacement, recruitment, selection, training and development, socialization cost, and perception of service quality by customers [33,35,36]. Turnover intention involves a sequence of the process, thinking of leaving, intentions to search, and to leave [37]. Several researchers (e.g., [38,39]) also stated that turnover intention is one of the best predictors of employees' actual quitting.

Work-Family Conflict
The work-family conflict (WFC) refers to an incompatible demand that an individual employee faces difficulty in participating in both work and family roles [40]. Family and work are the inseparable two central areas of an adult's life [41,42]. Both sides assist to shape peoples' roles and define their identity. Greenhaus and Beutell [43] stated that work-family conflict arises from the "simultaneous pressures from both roles which indicate that the relationship between them is a reciprocal one". Disruptive events in either arena may have serious consequences for the individual and the conflict is bi-directional between work and family roles [44]. In other words, conflict on one side will have an impact on the other side, i.e., work stress causes family conflict (WFC) and family stress causes work conflict (FWC). In line with this notion, Boyar et al. [45] also conceptualized work-family conflict into two facets that include "work interfering with family (WIF) and family interfering with work (FIW)". Choi and Kim [46] posited that WIF appears when employees' experience at work interferes with their family lives. Conversely, FIW occurs when employees experience stress in family events that interfere with their work-life [47]. Bande, Fernández-Ferrín, Varela, and Jaramillo [23] argued that they are more common and closely associated with job satisfaction, exhaustion, and turnover intention. Moreover, Demerouti et al. [48] propositioned that work-family conflict is a crucial factor that causes job burnout. This can lead to absenteeism and increased turnover intention of an employee [49].

Job Burnout
While most people are aware of burnout, there is no specific, generally agreed definition of the term, leaving it to be defined differently in the literature over the years. Freudenberg [50], who is credited with coining the term job burnout, defined it as occurring when a person becomes psychologically worn out and exhausted due to excessive demand that could be internally or externally imposed. Pines et al. [51] viewed burnout as resulting from work tedium where the employee felt distressed and discontented with the job. In the end, the employee experienced a feeling of being emotionally and physically depleted from work. Kahn [52] viewed burnout as a syndrome of negative attitudes towards others and self, which then resulted in negative psychological and physical exhaustion for the person.
Maslach and Jackson [53] defined burnout as "a syndrome of emotional exhaustion and cynicism that occurs frequently among individuals who do people-work of some kind". Burnout is described as consisting of three aspects: emotional exhaustion, depersonalization, and reduced personal accomplishment. Firstly, emotional exhaustion is obtained when staff members feel severe emotional fatigue and feel diminished or have no ambition to work. Secondly, depersonalization is accomplished when employees get away from work and indicate an indifferent and callous attitude at firms. Lastly, reduced personal accomplishment is attained when people feel a lack of competence and successful work accomplishment. According to Maslach et al. [54], burnout has been known to lead to poor physical health. Psychological symptoms can occur, such as negative self-concept, perfectionism, irritability, feeling of being unappreciated, and depression [55].

Emotional Intelligence and Turnover Intention
Raza et al. [56] stated that employees' emotional intelligence not only decreases frustration and stress in the workplace but also helps others to have less intention to quit. Optimistic emotions are influential from an individual perception and are indicative of cooperation and fairness within the organization [57]. Pessimistic emotions have negative effects on the organization as well as individuals. Those individuals who are upset have difficulty in assessing others' emotions accurately [58]. Emotional stress results in a lack of confidence, self-esteem, or motivation to reflect these limits. Employees tend to focus more on these negative emotions than they do their work and become disconnected physically and mentally, which results in underperformance and high turnover intention [59]. Emotional intelligence may be a key component to keep employees engaged and understand the emotional reasons for leaving decisions. Employees may use their emotional intelligence to better assess and understand the situation. Thus, emotional intelligence is considered a significant factor in predicting the turnover intention of employees which leads towards actual turnover [60]. There are several proposed theories that support a positive relationship between emotional intelligence and turnover intention: human capital theory, attraction-selection-attrition theory, and job embeddedness theory [61]. As employees continue to work for the same company, they become increasingly motivated to perform well; and, as their interests increasingly align with the company's values, their chances of leaving the company decrease [61]. Prior researches demonstrated a negative association between emotional intelligence and employees' turnover intention [60,62]. They stated that employees with higher emotional intelligence have very low intentions to quit. Based on the literature it is hypothesized that: Hypothesis 1: Emotional intelligence will reduce the turnover intention of bankers.

Emotional Intelligence and Work-Family Conflict
Emotional intelligence and work-family conflict are two fundamental aspects of service organizations [63]. Both the organization and its employees are responsible for eliminating work-family conflict. Companies have their own policies and procedures to help their staff manage stress, but their employees still burn out. Certainly, individuals hold some responsibilities for regulating their own Electronic copy available at: https://ssrn.com/abstract=3678384 family balance, but they need organizational support. Organizations need to recognize and adapt employees' work and lives to win employee loyalty [64]. According to the content of resource theory, Hobfoll [65] has identified individual differences as resources causing the negative impacts of stressful events on individuals. Individuals who have more personal resources can deal with the loss of other kinds of resources, including resource loss caused by work-family conflict. "Emotional intelligence represents individual differences in the ability and capacity to monitor and recognize one's own and other's emotions and to use this information to regulate one's emotions and actions" [66]. Concerning the role of individuals, researchers have emphasized emotional intelligence is an essential factor in protecting employees from the beginning of the work-family conflict and stress to identify, acknowledge, and manage the emotions. Particularly, Suliman and Al-Shaikh [67] stated that in terms of conflict management, employees with an inflated level of emotional intelligence tended to have effective control with conflict. Indeed, people with high-ability of emotional intelligence tended to suffer less work-family conflict [63]. Accordingly, the following hypothesis is proposed: Hypothesis 2: Emotional intelligence will reduce the work-family conflict of bankers.

Work-Family Conflict and Turnover Intention
Over the past several years, many researchers have believed that when experiencing work-family conflict, employees have a tendency to quit their job to eliminate the conflict. Employees may leave an organization because of the high stress of being overworked and limited personal time available to spend away from the office [68,69]. They do not want to choose between their personal lives and work, and, if they must, they will choose their personal lives. Employees leave when organizational rules are the cause of intolerable family stress and conflicts [68,69]. Greenhaus et al. [70] proved that being disappointed with family, work, and life leads to withdrawing from work. These studies demonstrated that when job-related retention grows too strong in the organization, one solution that employees can choose is to leave their firms. Allen and Armstrong [71] suggested that work-family conflict may cause employees to leave their jobs because the demands of work lead to frustration in the workplace and the strain from work makes it difficult to fulfill family duties. Employees took the emotions felt in the workplace home, as a result, they find it hard to concentrate on tasks. Similarly, some previous researchers examined and found that WFC has a significant effect on turnover intention [33,44,45]. Based on the foregoing review and previous research, it is hypothesized that: Hypothesis 3: Work-family conflict will increase the turnover intention of bankers.

Emotional Intelligence and Job Burnout
It has been suggested that emotional intelligence enables individuals to control pressure and adapt easily with the challenge to avoid burnout [72]. In relation to job burnout, people who have elevated levels of emotional intelligence deal better with life's challenges and job stresses, which leads to good psychological and physical health [72] and makes them less likely to experience job burnout [73]. The fourth dimension of emotional intelligence enables a person to regulate their emotions quickly and work on emotive information effectively [74]. Regarding this matter, the emotionally intelligent individual helps employees choose appropriate approaches to cope with frustration and make a response more effectively relating to the emotional aspect. Employees with high levels of emotional intelligence can keep away from dysfunctional and angry emotions that would be a consequence of burnout [64]. Individuals with high-ability to manage their emotions would be less likely to experience burnout. Therefore, it could be hypothesized that: Hypothesis 4: Emotional intelligence will reduce job burnout.

Work-Family Conflict and Job Burnout
The impact of the work-family conflict was investigated and the findings showed that work-family conflict was strongly influenced by lower job and family satisfaction, greater stress, and higher turnover intention [75]. Experiencing work-family conflict, individuals are subject to burnout that leads to the consequence of a conflict between work and family domains, which seems to result in draining their emotional and physical energy [23,76]. According to Hobfoll [65] in his conservation of resource theory, energy and time are resources that people attempt to maintain, protect, and create. In addition, "during recovery periods when not confronted with an immediate need for resources, individuals attempt to recover and stockpile resources to counter future losses" [76]. Hobfoll [65] perceived that work-family conflict expends an individual's energy and other sources and keeps them from "resource stockpiling" which is likely necessary to fend off other current and future resource demands. For this reason, individuals tend to suffer higher levels of burnout. People with a high level of work-family conflict are susceptible to burnout because conflicts inhibit individuals from accomplishing their work and family obligations. Consequently, work-family conflict could cause job burnout. Thus, the following hypothesis is presented: Hypothesis 5: Work-family conflict will increase job burnout.

Job Burnout and Turnover Intention
Many researchers have found that job burnout is strongly associated with turnover intention [77,78]. Similarly, according to Layne et al. [79], the turnover intention corresponds with stress. Employees with a high level of stress are more likely to leave their organization and seek another opportunity. In addition, burnout will result in lower rates of organizational commitment and higher rates of turnover and turnover intention [80]. Hence, an employee experiencing a higher level of burnout results in a higher level of turnover intentions. Therefore, the following relationship is hypothesized: Hypothesis 6: Job burnout will increase turnover intention.

The Mediating Role of Work-Family Conflict and Job Burnout
Basically, emotional intelligence is the ability to understand and to regulate one's own emotions effectively, as well as to apply these abilities to help achieve desired outcomes [81]. Normally, people who have high emotional intelligence tend to have good control over their emotions. They won't let negative emotions affect their life, work, and relationship with others. Rather, they know how to use their knowledge about emotions to enhance their interpersonal skills and to help them make good decisions without being affected by emotions. Employees with high emotional intelligence may properly balance the relationship between work and family and reduce job burnout thus resulting in less turnover intention. Previous research revealed the direct relationships between emotional intelligence and work-family conflict, job burnout [82,83], emotional intelligence and employees' turnover intention [60,62,84], work-family conflict, job burnout and employees' turnover intention [44,60]. Linking the associations outlined above, the authors argue that work-family conflicts and job burnout are likely to mediate the relationship between emotional intelligence and turnover intention. Based on the foregoing review and previous research, the hypotheses below are proposed: Hypothesis 7: Work-family conflict mediates the relationship between emotional intelligence and turnover intention of bankers.

Hypothesis 8:
Job burnout mediates the relationship between emotional intelligence and turnover intention of bankers.

Perceived Organizational Support
Perceived organizational support is theoretically defined as the employees' global beliefs about the extent to which the organization values their contributions and cares about their well-being [85]. The individual's perception about the degree to which their organization values their well-being is rooted in the nature of the give-and-take relationship between the workers and their organization [22]. Perceived organizational support theory suggests that when the emotional and social needs of the employee are fulfilled, there are positive outcomes for both the organization and the individual [21,85]. Perceived organizational support is expected to decrease negative psychological and psychosomatic reactions (e.g., burnout) to employees by providing supportive external resources.
According to the theory of perceived organizational support, the employee assigns the organization humanlike characteristics [85]. That is, employees view actions or behaviors by agents of the organizations as actions of the organization itself. Employees view favorable or unfavorable treatment directly by the organization representatives and indirectly through organizational policy as an indication of the extent to which the organization supports them. Employees who believe that their work environment is supportive are more likely to achieve personal and organizational goals, which leads to effectively balancing their work-family. Alternatively, unsupportive work environments foster a win-lose approach [21]. In unsupportive work environments, workers' individual goals are typically incompatible with other workers' goals or even organizational goals so that as one worker moves toward goal attainment, others are less likely to achieve their goals. It is plausible then that the extent to which an organization is perceived to be supportive or non-supportive can influence whether there is a positive outcome for both the employee and the organization and a reduction of work-family conflict as well as job burnout. Perceived organizational support is thought to reduce work-family conflict and job burnout by fostering the perception that material aid and emotional resources will be available when needed to deal with negative situations (e.g., work-family conflict, job burnout, and turnover intention). Besides, some scholars demonstrated that perceived organizational support could decrease work-family conflict (e.g., [86]), job burnout (e.g., [87]), and turnover intention (e.g., [88]). Therefore, the following hypotheses are suggested: Hypothesis 9: Perceived organizational support reduces the turnover intention of bankers.
Hypothesis 10: Perceived organizational support reduces the work-family conflict of bankers.
Hypothesis 11: Perceived organizational support reduces job burnout of bankers.

Moderating Role of Perceived Organizational Support
Perceived organizational support theorists indicated that perceived organizational support reduces work-family conflict and job burnout because it provides external work environment physical and emotional resources to help employees deal with burnout as well as work-family conflict [86]. Individuals who have high emotional intelligence, work in environments where cooperative (i.e., supportive) values such as fair and respectful treatment of others are supported and are likely to diminish work-family conflict and job burnout [87]. Supportive work environments are characterized by effective open communication, helpfulness, fair and respectful treatment of others, and collaboration to achieve mutual goals leading to win-win solutions [89]. Moreover, the authors state that perceived organizational support could moderate the relationship between emotional intelligence and work-family conflict and job burnout because employees with high emotional intelligence have lower work-family conflict and job burnout levels. However, if the highly emotionally intelligent employees perceived low organizational support, their work-family conflict and job burnout levels could be increased, but if the highly emotionally intelligent employees perceived high organizational support, their work-family conflict and job burnout levels could be decreased as well. Thus, it is plausible that perceived organizational support may moderate or blunt the effects of emotional intelligence leading Electronic copy available at: https://ssrn.com/abstract=3678384 to a decrease in work-family conflict and job burnout. Based on the discussion above, the following hypotheses are proposed: Hypothesis 12: Perceived organizational support moderates the effect of emotional intelligence on work-family conflict. The negative relationship between EQ and WFC will be stronger for employees who work in a supportive environment.
Hypothesis 13: Perceived organizational support moderates the effect of emotional intelligence on job burnout. The negative relationship between EQ and JB will be stronger for employees who work in a supportive environment.

Control Variables
Some researches (e.g., [90,91]) stated that employee demographic variables could be related to turnover intention. For example, Seyrek and Turan [91] argued that "both women and men are confronted with different demands from home, community and workplace in their life and these different groups compete for the person's time and energy. Especially for women, it can be difficult to cope with conflicts arising from playing different roles in workplace and family and this may lead to turnover intention". Although the impacts of individual characteristics of the employee on turnover intentions may differ in different environments, empirical studies in the banking industry context is scanty. Therefore, this study has been conducted exploring control variables (e.g., gender, marital status, age group, income, and educational level) in combination with turnover intention. The following hypotheses can be proposed (Figure 1):

Hypothesis 14:
There is a statistically significant relationship between control variables (e.g., gender, marital status, age group, income, and educational level) and turnover intention in the commercial banks of Vietnam.

Procedure and Sampling Size
Respondents were the full-time employees of banks in Vietnam. The sample was selected using nonprobability technique-convenience sampling. First of all, the researchers collected a list of commercial banks in Vietnam. Thirty-five commercial banks were chosen to conduct the survey. The researchers contacted the concerned person in charge (i.e., human resource manager) over the phone prior to conducting the survey. The topic and purposes were explained clearly. Ten out of thirty-five banks allowed the researchers to collect data.
Based on the previous studies (e.g., [74,85,[92][93][94]), a draft questionnaire was formed. The draft questionnaire was originally developed in the English language. Then the researchers translated the questions into the Vietnamese language and piloted the back-translation with the support of several English language experts. The current study consisted mainly of two stages including qualitative and quantitative research. For qualitative research, the Vietnamese version of the questionnaire was tested by an in-depth interview method in one week with ten employees from three banks to ensure if they understood the questions and revised Vietnamese terms which were unclear due to translation. Based on the comments of respondents, the survey questionnaire was modified properly.
The pilot study was sent to 50 employees working in the Vietnamese banking industry by a convenient method. The participants were asked to provide advice on elements of the survey that were confusing. They also provided recommendations on the wording and any questions they felt uncomfortable answering. Modifications were made to the instrumentation, specifically around grammatical errors and survey logic. The modified instrument was found to be reliable as the minimum Cronbach's Alpha of each factor equals to 0.743 (Table 1). The individual items were deemed to be valid for the research as for each dimension the Cronbach's alpha was above the acceptable threshold of 0.70 and the minimum value of corrected item-total correlation coefficients was higher than 0.3 [95]. Therefore, the reliability of the scales is sufficiently good to test the main survey. After finishing the main questionnaires, the authors distributed the questionnaires to 10 banks in the chosen sample. A total of 764 completed questionnaires were collected. Among these, 42 questionnaires were found invalid because the respondents answered one choice for all questions. Consequently, there were 722 valid questionnaires usable for further data analysis. Table 2 below shows the diverse information about the demographic profile of respondents.

Measurement
All variables in the conceptual model were measured with multiple items, which were developed by prior scholars [74,85,[92][93][94]. Particularly, an emotional intelligence scale was measured by 16 items of Wong and Law [74]. Sample items included (Self-emotion appraisal: e.g., "Self-emotion appraisal I have a good sense of why I have certain feelings most of the time"); (Others' emotion appraisal: e.g., "I always know my friends' emotions from their behavior"); (Use of emotion: e.g., "I always set goals for myself and then try my best to achieve them"); (Regulation of emotion: e.g., "I am able to control my temper so that I can handle difficulties rationally"). Perceived organizational support was measured by 7 items of Eisenberger, Huntington, Hutchinson, and Sowa [85]. Sample items for this construct were "The organization really cares about my well-being" and "Help is available from the organization when I have a problem". The work-family conflict scale was measured by 5 items about work interference with family from Netemeyer, Boles, and McMurrian [92]. A sample item for this construct was "The amount of time my job takes up makes it difficult to fulfill family responsibilities". The job burnout scale was measured by 4 items from Lee [93]. Sample items for this construct were "I worry that this job is hardening me emotionally" and "I leave work feeling tired and rundown". Turnover intention scale was measured by 4 items of Vigoda [94]. A sample item for this construct was "Next year I will probably look for a new job outside this organization". All items of these constructs were measured using a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree).

Partial Least Squares Regression
The statistical technique chosen to test the stated hypotheses was partial least squares (PLS) path analysis, otherwise known as partial least squares structural equation modeling (PLS-SEM). PLS-SEM is a statistical analysis technique for data exploration within the quantitative research discipline used to measure the observed variables collected from instruments to determine their influence on latent or unobserved variables [96]. Researchers have suggested the use of PLS-SEM due to its effective use as an analysis tool to support prediction models from empirical data [97,98]. Moreover, PLS-SEM is appropriate for analysis when different measurement scales are used in the research model because it allows optimal empirical assessment of a structural model. The use of PLS-SEM data analysis utilizes multi-item observations to measure the area under review as this has shown to be more reliable than using a single-item of observation, as this is shown to provide more accurate reflections on the latent variable. Moreover, PLS-SEM has the capability to calculate p-values through a bootstrapping technique if samples are independent and if the data is not required to be normally distributed [39,98]. Therefore, data analysis for this study was performed using PLS-SEM with the SmartPLS 3.0.

Research Results
Before performing PLS-SEM estimation for hypotheses testing, the validity and reliability of the multi-item measures should be assessed [99]. Convergent validity is the amount of variance between two or more items that agree when measuring similar constructs and is calculated using factor loadings. Hair, Hult, Ringle, and Sarstedt [98] stated that convergent validity will be suitable when the factor loadings are above 0.5. In Figure 2, the statistical results showed that all factor loadings were more than the threshold of 0.5. For example, the minimum factor loading for Use of emotion = 0.791, Regulation of emotion = 0.825, Self-emotion appraisal = 0.680, Others' emotion appraisal = 0.834, Emotional intelligence = 0.609, Perceived organizational support = 0.632, Work interference with family = 0.833, Job burnout = 0.818, and Turnover intention = 0.604. Besides, convergent validity will be confirmed when the average variance extracted (AVE) for each of the latent variables is greater than 0.5 [96]. As shown in Table 3 below, AVE values were reported for each of the variables ranging from 0.611 to 0.759 (perceived organizational support and emotional intelligence, respectively). Therefore, all constructs showed good convergent validity. Additionally, Hair, Hult, Ringle, and Sarstedt [98] recommended that discriminant validity can be shown through the correlation matrix. The square root of a construct's AVE value should be greater than the squared correlation with any other construct "since a construct shares more variance with its associated indicators than it does with any other construct" [96]. Table 2 indicated the correlation matrix of the constructs with the diagonal values. Square roots of AVE of latent constructs which were shown in the parentheses were higher than the correlation coefficient with any variable. For example, the AVE value of Others' emotion appraisal was 0.734 and the square root of its AVE was 0.857. This value was both higher than the correlation values in its column (0.069, 0.819, 0.050, −0.014, 0.055, −0.490, and 0.106) and its row (0.362, 0.481, 0.652, −0.383, 0.505, and −0.267). Therefore, discriminant validity for the constructs was established.
Next, the Cronbach's alpha and composite reliability values were used to assess construct reliability. According to Hair, Hult, Ringle, and Sarstedt [98], Cronbach's alpha and composite reliability values should be 0.7 or greater to be considered reliable in a model. As shown in Table 3 below, Cronbach's alpha and composite reliability values were more than 0.7. The minimum composite reliability and Cronbach's alpha values were 0.875 and 0.807, respectively. Furthermore, Hair, Hult, Ringle, and Sarstedt [98] also suggested the rho_A coefficient to measure the reliability for the partial least squares and this value should be greater than 0.7. According to Table 3, the rho_A values varied from 0.819 to 0.922. Consequently, the authors can verify that all constructs achieved good reliability. Additionally, Hair, Hult, Ringle, and Sarstedt [98] recommended that discriminant validity can be shown through the correlation matrix. The square root of a construct's AVE value should be greater than the squared correlation with any other construct "since a construct shares more variance with its associated indicators than it does with any other construct" [96]. Table 2 indicated the correlation matrix of the constructs with the diagonal values. Square roots of AVE of latent constructs which were shown in the parentheses were higher than the correlation coefficient with any variable. For example, the AVE value of Others' emotion appraisal was 0.734 and the square root of its AVE was 0.857. This value was both higher than the correlation values in its column (0.069, 0.819, 0.050, −0.014, 0.055, −0.490, and 0.106) and its row (0.362, 0.481, 0.652, −0.383, 0.505, and −0.267). Therefore, discriminant validity for the constructs was established.
Next, the Cronbach's alpha and composite reliability values were used to assess construct reliability. According to Hair, Hult, Ringle, and Sarstedt [98], Cronbach's alpha and composite reliability values should be 0.7 or greater to be considered reliable in a model. As shown in Table 3 below, Cronbach's alpha and composite reliability values were more than 0.7. The minimum composite reliability and Cronbach's alpha values were 0.875 and 0.807, respectively. Furthermore, Hair, Hult, Ringle, and Sarstedt [98] also suggested the rho_A coefficient to measure the reliability for the partial least squares and this value should be greater than 0.7. According to Table 3, the rho_A values varied from 0.819 to 0.922. Consequently, the authors can verify that all constructs achieved good reliability.
Electronic copy available at: https://ssrn.com/abstract=3678384  Electronic copy available at: https://ssrn.com/abstract=3678384 Finally, multicollinearity was assessed for all of the constructs. The variance inflation factor (VIF) indicator was suggested to measure multicollinearity issues. The VIF value should be less than a 5.00 tolerance level [95,98]. As shown in Table 4, the maximum inner VIF value of constructs was 2.190. Therefore, the collinearity of the latent variables was not a concern. Results from the PLS-SEM analysis are shown in Figure 3. Standardized path coefficients and p-values are reported. The findings are presented as follows: Finally, multicollinearity was assessed for all of the constructs. The variance inflation factor (VIF) indicator was suggested to measure multicollinearity issues. The VIF value should be less than a 5.00 tolerance level [95,98]. As shown in Table 4, the maximum inner VIF value of constructs was 2.190. Therefore, the collinearity of the latent variables was not a concern. Results from the PLS-SEM analysis are shown in Figure 3. Standardized path coefficients and pvalues are reported. The findings are presented as follows:  Table 5). The results indicated that the more emotional intelligence, the greater the possibility that bankers will have low levels of turnover intention, which was consistent with the findings of the previous study of Akhtar, Shabir, Safdar and Akhtar [62] and Riaz, Naeem, Khanzada and Butt [60]. Emotional intelligence was related to bankers who had the propensity to leave his or her job based on a prolonged period of time being emotionally low regulated by employees' work. Thus, Hypothesis 1 was supported.  Table 5). The results indicated that the more emotional intelligence, the greater the possibility that bankers will have low levels of turnover intention, which was consistent with the findings of the previous study of Akhtar, Shabir, Safdar and Akhtar [62] and Riaz, Naeem, Khanzada and Butt [60]. Emotional intelligence was related to bankers who had the propensity to leave his or her job based on a prolonged period of time being emotionally low regulated by employees' work. Thus, Hypothesis 1 was supported.
Electronic copy available at: https://ssrn.com/abstract=3678384 Hypothesis 2: the result found that emotional intelligence had a negative impact on work-family conflict with a standardized coefficient of −0.140 and p-value = 0.000, which corresponds with many studies (e.g., [63]) which suggested that emotional intelligence is one of the preliminary factors preventing employees from establishing work-family conflict. Subordinates with a good understanding of emotional intelligence would gain feelings of balancing her/his work and family. He or she not only senses that the work being performed is meaningful, but also exhibits the capability to take advantage of work-life balance to have mutual results: completing the job successfully and having a happy family life. Moreover, emotional intelligence helps employees figure out work-family conflicts and helps them to manage their emotions. In terms of conflict management, employees with better emotional intelligence tend to have better effective control with work-family conflict than others. Thus, emotional intelligence (self-emotion appraisal, others' emotions appraisal, use of emotion, regulation of emotion) are important for individual difference effects in regulating emotion in work-family life. Therefore, Hypothesis 2 was supported.
Hypothesis 3: This analysis found that work-family conflict had a positive and significant relationship with the turnover intention of bankers (p-value = 0.000 and beta coefficient = 0.169). It was connected to the studies of Khan, Nazir, Kazmi, Khalid, Kiyani, and Shahzad. [33] and Wang, Lee, and Wu [44]. Vietnamese bankers with higher work-family conflicts were found to have greater job turnover intentions. With a regression weight of 0.169, the results showed that bringing conflict from the workplace to employees' homes would likely force them to think of leaving their job. Moreover, in the banking sector, the majority of employees are women, and they may choose to leave an organization voluntarily because of family responsibilities, such as childbearing or child-rearing. Women are generally regarded as the caretakers for an elderly parent as well. As the secondary income earner in the household, a woman's income is characterized traditionally as non-crucial income, which is another common reason identified for turnover. Similarly, from a conventional perspective, women require the flexibility or support from an organization to fulfill family responsibilities. If these requirements cannot be fulfilled by a firm, the employee is more likely to leave voluntarily. Therefore, Hypothesis 3 and physical health and seems to lessen employees' suffering from job burnout. Thus, Hypothesis 4 was supported.
Hypothesis 5: The results showed that work-family conflict had a positive impact on job burnout with the standardized coefficient of 0.164 and p-value =0.003, which is consistent with the findings of the previous studies of Bande, Fernández-Ferrín, Varela, and Jaramillo [23] and Golden [76]. Employees who experience stress from work-family conflict over a prolonged period of time get drained of energy, which eventually results in job burnout and vice versa. Indeed, the findings also showed that work-family conflict is associated with negative consequences (i.e., burnout) that affect both the work and family. Additionally, high values of work-family conflict may lead to contemporary feelings of emotional exhaustion, reduced personal accomplishment, and depersonalization. Therefore, Hypothesis 5 was supported.
Hypothesis 6: The results showed that job burnout had a positive impact on the turnover intention with the standardized coefficient of 0.345 and p-value = 0.000, which is consistent with the findings of the previous study of Gharakhani and Zaferanchi [78] and Scanlan and Still [77]. Job burnout is related to bankers in the commercial banks propensity to leave his or her job based on a prolonged period of time being emotionally overextended and exhausted by the employees' work. Thus, Hypothesis 6 was supported.
Hypothesis 7: According to Table 6, work-family conflict mediated the relationship between emotional intelligence and turnover intention due to the following reasons: first, the results in Table 5 revealed that the p-value for the direct path EI→TI was 0.015; EI→WFC was 0.000; WFC→TI was 0.000, which were statistically significant (p < 0.05). Second, the p-value of the indirect effect (EI→WFC→TI) was 0.000 (Table 6) which was statistically significant as well. Hence, the mediating role of work-family conflict exists [95]. Therefore, Hypothesis 7 was supported and this mediation was partial. Hypothesis 8: According to Table 6, job burnout mediated the relationship between emotional intelligence and turnover intention due to the following reasons: first, the results in Table 5 revealed that the p-value for the direct path EI→TI was 0.015; EI→WFC was 0.000; WFC→TI was 0.000, which were statistically significant (p < 0.05). Second, the p-value of the indirect effect (EI→JB→TI) was 0.000 (Table 6) which was statistically significant as well. Hence, the mediating role of job burnout exists [95]. Therefore, Hypothesis 8 was supported and this mediation was partial.
Hypothesis 9: The results showed that perceived organizational support had a negative and significant relationship with the turnover intention of bankers, (p-value = 0.000 and beta coefficient = −0.261) ( Table 5). The result indicated that the more perceived organizational support, the greater the possibility that bankers will have low levels of turnover intention, which was consistent with the findings of Marchand and Vandenberghe [88]. Thus, Hypothesis 9 was supported.
Hypothesis 10: The results showed that perceived organizational support had a negative and significant relationship with work-family conflict, (p-value = 0.000 and beta coefficient = −0.600) ( Table 5). The result indicated that the more perceived organizational support, the greater the possibility that bankers will have low levels of work-family conflict, which was consistent with the findings of Gurbuz, Turunc, and Celik [86]. Thus, Hypothesis 10 was supported.
Hypothesis 11: The results showed that perceived organizational support had a negative and significant relationship with the job burnout of bankers, (p-value = 0.000 and beta coefficient = −0.175) ( Table 5). The results indicated that the more perceived organizational support, the greater the Electronic copy available at: https://ssrn.com/abstract=3678384 possibility that bankers will have low levels of job burnout, which was consistent with the findings of Caesens, Stinglhamber, Demoulin and De Wilde [87]. Thus, Hypothesis 11 was supported.
On the other hand, the findings showed that the component structure of emotional intelligence was a second-order structure. This result was consistent with emotional intelligence concepts [74], which stated that emotional intelligence was composed of four sub-components: self-emotion appraisal, others' emotions appraisal, use of emotion, and regulation of emotion. Regarding the dependency level of each variable to their subscales via arrows, the subscales of use of emotion had the smallest share and others' emotion appraisal had the largest share in stating emotional intelligence.

The Moderating Role of Perceived Organizational Support
Hypothesis 12 predicted that perceived organizational support would moderate the relationship between emotional intelligence and work-family conflict. The study showed that the moderating effect 1 of the interaction between emotional intelligence and perceived organizational support with work-family conflict was negative and statistically significant (p-value = −0.072 and beta coefficient = 0.006) (Figure 4). This finding proposed that perceived organizational support negatively moderated the relationship between emotional intelligence and work-family conflict. In other words, the negative relationship between EQ and WFC was stronger for employees who work in a supportive environment ( Figure 5). Therefore, Hypothesis 12 is supported. possibility that bankers will have low levels of job burnout, which was consistent with the findings of Caesens, Stinglhamber, Demoulin and De Wilde [87]. Thus, Hypothesis 11 was supported. On the other hand, the findings showed that the component structure of emotional intelligence was a second-order structure. This result was consistent with emotional intelligence concepts [74], which stated that emotional intelligence was composed of four sub-components: self-emotion appraisal, others' emotions appraisal, use of emotion, and regulation of emotion. Regarding the dependency level of each variable to their subscales via arrows, the subscales of use of emotion had the smallest share and others' emotion appraisal had the largest share in stating emotional intelligence.

The Moderating Role of Perceived Organizational Support
Hypothesis 12 predicted that perceived organizational support would moderate the relationship between emotional intelligence and work-family conflict. The study showed that the moderating effect 1 of the interaction between emotional intelligence and perceived organizational support with work-family conflict was negative and statistically significant (p-value = −0.072 and beta coefficient = 0.006) (Figure 4). This finding proposed that perceived organizational support negatively moderated the relationship between emotional intelligence and work-family conflict. In other words, the negative relationship between EQ and WFC was stronger for employees who work in a supportive environment ( Figure 5). Therefore, Hypothesis 12 is supported.  Hypothesis 13 predicted that perceived organizational support would moderate the relationship between emotional intelligence and job burnout. The study showed that the moderating effect 1 of the interaction between emotional intelligence and perceived organizational support with job burnout was not statistically significant (p-value = 0.701) (Figure 4). Therefore, Hypothesis 13 was not supported.
Hypothesis 14: The results indicated income negatively related to turnover intention (p-value = 0.065) at the significance level of 10%, which means that bankers who had a high income tended to Electronic copy available at: https://ssrn.com/abstract=3678384 Hypothesis 13 predicted that perceived organizational support would moderate the relationship between emotional intelligence and job burnout. The study showed that the moderating effect 1 of the interaction between emotional intelligence and perceived organizational support with job burnout was not statistically significant (p-value = 0.701) (Figure 4). Therefore, Hypothesis 13 was not supported.
Hypothesis 14: The results indicated income negatively related to turnover intention (p-value = 0.065) at the significance level of 10%, which means that bankers who had a high income tended to reduce turnover intention more than other bankers (beta coefficient = −0.019). This finding was different from some prior results of Seyrek and Turan [91] and Liu, Liu and Hu [90]. However, this study did not find the relationship between turnover intention and gender (p = 0.559; β = −0.015), age (p = 0.630; β = −0.013), educational level (p = 0.432; β = −0.049), and marital status (p = 0.496; β = −0.017), which was consistent with the findings of Seyrek and Turan [91] and Liu, Liu and Hu [90].

Model Fit
The coefficient of determination (R 2 ) is a measure of the model's predictive power. R 2 is the amount of variance in the endogenous (dependent) latent variables in the structural model explained by the exogenous (independent) constructs connected to it. R 2 values range from 0 to 1. The higher the R 2 coefficient, the better the construct is explained by the latent constructs in the structural model. The high R 2 coefficient also reveals that the values of the variables can be well predicted by the PLS path model [95,98,100]. The R 2 value for the turnover intention was 0.515 which indicated that 51.5% of the total variation of the endogenous construct turnover intention may be explained by the exogenous construct such as emotional intelligence, perceived organization support, work-family conflict, and job burnout ( Figure 2). Moreover, R 2 values and the effect for endogenous latent variables in behavioral sciences can be assessed as 0.26 (large effect), 0.13 (moderate effect), and 0.02 (weak effect). Due to the fact that R 2 values for the turnover intention were greater than 0.26, the model of this study proved the model-data fit.

Discussion
Turnover is a costly consequence as recruiting, selecting, and training new members requires additional resources that might negatively affect the performance of a service organization. This study contributes to the existing body of literature supporting the role of emotional intelligence and perceived organizational support in turnover intention. It also focusses on the additional variables of work-family conflict and job burnout having a significant relationship in conjunction with emotional intelligence and perceived organizational support in influencing the turnover intention. Emotional intelligence and perceived organizational support are accounted for 47.2% and 37.5% of the variance in work-family conflict and job burnout, respectively. When work-family conflict and job burnout are added as the additional independent variables with emotional intelligence and perceived organizational support, the combination is accounted for 51.5% of the variance in turnover intention. These numbers may appear to be large; they are very important to banks looking to improve their retention. Being a banker can be an extremely stressful profession, emotionally intelligent bankers are better equipped to work in teams, deal with the job burnout, and interact with their customers. A better understanding of the relationship between emotional intelligence, perceived organizational support, work-family conflict, and turnover intention has the potential to help enhance the workplace as a whole.
An emotionally intelligent banker is more likely to be able to manage her/his tasks properly. Possessing the ability to communicate efficiently, appropriately, and quickly with customers, bankers will have more time for other tasks. It also can increase the accuracy of the banker's work if he or she is better able to understand the needs and wants of the customers in addition to a better understanding of the needs and wants of his/her family. A non-emotionally intelligent banker might struggle to communicate with the customer or misinterpret the customer's needs which can result in the delivery of inappropriate service. A less emotionally intelligent banker may also struggle to interact appropriately with the family members. This could leave the family with the impression that the banker is not caring Electronic copy available at: https://ssrn.com/abstract=3678384 for their loved one which could also leave the family dissatisfied with him/her, leading to an increase in work-family conflict.
Another way emotional intelligence can impact turnover intention is its effect on work-family conflict and job burnout. Employees with higher levels of emotional intelligence are more balanced in their job with family which in turn results in a lower level of job burnout. Emotionally intelligent bankers who can deal with the stressors of the workplace feel more confident and successful, so they have the potential to enjoy their jobs. This would decrease their likelihood of leaving the workplace. They would provide a greater contribution to the overall success of their unit and the bank. Bankers not possessing emotional intelligence would struggle to deal with the stress and demands of the workplace, which may make them more vulnerable to continue their jobs.
Additionally, the recent study of perceived organizational support included plentiful direct causal relations with turnover intention, and it illustrated that elevated employee perceptions of organizational support have been supportive of the reduced intention to quit. Besides, this research found that POS also reduced work-family conflict and job burnout. The negative relationship among POS, work-family conflict and turnover intention was consistent in relation to the previous researches of Marchand and Vandenberghe [88], Gurbuz, Turunc and Celik [86] and Caesens, Stinglhamber, Demoulin and De Wilde [87]. POS were key mechanisms in connecting voluntary turnover among employees. This research found that POS contributed to proactive employees' intent to retain their position in the organization. A positive perception of human resource management practices could lead employees to be less likely to quit as trust could form under a social exchange theory. Perceived organizational support is a belief emanating from the social exchange and the norm of reciprocity. According to organizational support theory, employees exchange valued organizational outcomes for the belief that the organization values their contribution and cares about their well-being as evidenced by tangible support as well as the fulfillment of intangible socio-emotional needs. In other words, employees who feel greater inducements might sense the obligation to repay the organization. Moroever, individuals who have stronger POS might be less likely to seek other jobs or consider turnover.

Conclusions
Keeping positive and motivated staff is a vital factor for commercial banks and has become a pressing matter for the board of directors. When a banker leaves his/her organization, the organization must absorb the cost of losing an employee such as the cost of recruiting a new employee, which consists of advertising, interviewing, choosing and hiring, and the cost of management activities, which consists of orientation and training. Thus, researchers have examined and proved the effect of emotional intelligence in jobs that required high interaction with others, such as bankers. Particularly, the study examined the effect of the emotional intelligence of commercial bank employees in Vietnam on the turnover intention of employees through the mediation of work-family conflict and job burnout. The findings strongly supported most of the proposed hypotheses. The results of testing have found correlations between a system of variables. The findings showed that emotional intelligence had a negative relationship with work-family conflict, job burnout, and turnover intention. The respondents in this study who had a high level of work-family conflict and job burnout will experience a high level of turnover intention. On the other hand, the high ability of emotional intelligence could lower the level of work-family conflict and job burnout which later could reduce employees' turnover intention.
In terms of emotional intelligence, this research reviewed the definition and dimensions of the emotional intelligence construct and argued its important role in reducing work-family conflict and job burnout issues which also decreased the turnover intention. Due to the importance of this issue, it is strongly recommended to promote the development of emotional intelligence and perceived organizational support to reduce work-family conflict, job burnout, and turnover intention.
Besides, the moderating role of perceived organizational support has not been found in any previous papers about emotional intelligence and turnover intention. Therefore, this study demonstrated the existence of the moderating role of perceived organizational support on these Electronic copy available at: https://ssrn.com/abstract=3678384 relationships. POS could decrease work-family conflict, job burnout, and turnover intention of employees. It could also moderate the relationship between emotional intelligence and work-family conflict. This negative relationship was stronger for employees who work in a supportive environment.

Managerial Implications
The results of this research study lead to many practical recommendations for both banking practitioners and managers of organizations desirous of improving work-related outcomes.
First of all, hiring employees with high levels of emotional intelligence could have a considerable impact on reducing organizational work-family conflict, job burnout, and turnover intention. Effective retention strategies often begin during the employee recruitment process. When choosing the right employee, the manager should strongly consider the personality, temperament, and capacity of candidates. People who are passionate and active will have strong emotion regulation ability. Thus, during the recruitment process, psychological tests and interviews are suggested methods to determine emotional intelligence. The psychological assessment usually includes a clinical interview, assessment of intellectual functioning (IQ), personality assessment, and behavioral assessment. A personality test is important because it allows the interviewer to thoroughly understand a person and their behavior which includes temperament, personality, interests, attitudes, values, motivation, and other non-cognitive factor characteristics. When interviewing candidates, the interviewer can be flexible and extend the interview to various aspects such as education, knowledge, experience, and achievement. By these approaches, the employer can properly judge the interviewee and select the qualified staff for the organization. In these ways, the organization can hire an employee with high levels of emotional intelligence which can help the organization keep its critical workforce.
Secondly, enhancing training for existing staff is considered. Since practical training can improve individual emotional intelligence levels, organizations may consider implementing Employee Assistance Programs (EAPs) to foster bankers' emotional competence. EAPs have been formed in American companies. EAPs are employee benefit programs that provide professional advice and guidance to help staff with personal problems or work-related conflicts that have a positive impact on reducing work-family conflict and turnover throughout the enterprise. Managers should develop and utilize EAPs in training employees.
Third, perceived organizational support could reduce work-family conflict, job burnout, and turnover intention. Thus, managers of organizations have fresh insight regarding how the delivery of supportive practices to their employees may result in an accretive improvement in organizational outcomes. Specifically, engaging in practices that instill within employees the belief that the organization values their contributions and cares about their well-being may achieve better, stronger outcomes when complemented by related but different organizational practices that increase job satisfaction and decrease work-family conflict, job burnout, and turnover intention to the organization. An employee's belief that the organization is supportive may better explain to management the variance in the levels of worker commitment, which in turn may add insight into how the effect of organizational support is transmitted to outcomes meaningful for organizational strategy execution.
Fourth, emotional intelligence interacts with perceived organizational support upon the intention to quit in a manner such that employees with higher levels of emotional intelligence will experience less work-family conflict and job burnout which will reduce their quitting intention. Management may, therefore, wish to screen potential hires for emotional intelligence, as well as implement training programs that will help employees perceive and better manage their emotions. Of equal importance is the finding that emotional intelligence appears to have an incremental effect upon outcomes as opposed to a basic foundational relation.
Finally, one of the positive ways to reduce the negative effects associated with work-family conflict, job burnout, and turnover intention is to provide a fair work environment. Fairness is a basic element that human beings need and is also a key factor in influencing people's emotions. This will create a pleasant atmosphere, and inspire all people in the organization to build up enthusiasm and be more Electronic copy available at: https://ssrn.com/abstract=3678384 positive. Besides, bankers must know clearly what to do for their customers and establish realistic expectations regarding performance. Managers should always be concerned about their employees and recognize and assist them to deal with problems effectively. A fair atmosphere is beneficial for the employee and the employer; if employees are treated well, then they are more likely to treat customers well. Furthermore, employees will retain positive emotions in the workplace, and the level of work-family conflict, job burnout, and turnover intention will reduce.

Limitations and Recommendations for Future Research
In this study, there were some limitations. First, this was a cross-sectional study. Future research should test the relations proposed with this study in longitudinal studies using a probability sampling technique, simple random sampling, to increase the generalization of the study. Secondly, multi-culture was another limitation of this project. Although emotional intelligence may be a universal construct, it is believed that the characteristics of the emotional intelligence of individuals may vary across cultures. However, emotional intelligence across cultural boundaries was not mentioned in the literature review. Finally, the survey method was the only method used in this study. Experimental study and field study should be considered in any future research.
Despite these limitations listed above, the present research has practical implications for managers in commercial banks. It is expected to expand the literature on turnover intention in Vietnamese commercial banks and will motivate further research from this scope. Obtaining actual turnover data is much better than turnover intention since there is still a gap between people who just have the intention to leave and those who ultimately leave their organization. This paper mainly studied the relationship between emotional intelligence, perceived organizational support, work-family conflict, and turnover intention. There are also other antecedents that predict turnover such as job satisfaction and employee engagement. These factors should also be involved in investigating the relationship between emotional intelligence and turnover intention.