The Impact of Human Resource Management Practices on Employee Engagement and Moderating Role of Gender and Marital Status: An Evidence from the Vietnamese Banking Industry

The aim of this research is to examine the effect of human resource management (HRM) practices on employee engagement, and the moderating roles of gender and marital status. The survey was conducted with a total of 457 frontline employees who are working in the Vietnamese banking industry. Partial least squares structural equation modeling (PLS-SEM) was estimated using the SmartPLS 3.0 program to examine the hypotheses. The results indicated that HRM practices and its four key dimensions (Training; Reward and recognition, Employee engagement, and Information sharing) positively impacted employee engagement. Moreover, this study provided an empirical evidence about the moderating roles of gender and marital status. It implied that male frontlines employees heightened the positive relationship between HRM practices and employee engagement, but this positive relationship was weakened by married employees. The finding also suggested that bank managers should develop a suitable HRM practices to reinforce employee engagement.


Introduction
In the context of the international economic tendency, WTO integration, Vietnam now has both opportunities and challenges for organizations to raise pressure on the competitiveness of the product, services, brand, and human resources. In particular, human resources is considered a key advantage for organizations to develop their businesses, it is a valuable asset to help organizations improve their competitiveness and performance (Dessler, 2019). So, human resources management is one of the prominent topics in Vietnam and gets more attention in the context that we want to keep up with the fourth industrial revolution or called new economic 4.0. Quality and productive labor is the most important to get development and attraction investment. Furthermore, human resources are part of companies' strategy to get competitive advantage and effectiveness. Its role is explained by the resource-based view theory (RBV). Based on RBV, Armstrong and Taylor (2017) stated that "competitive advantage can be achieved if a firm's resource is valuable, rare and costly to imitate". Human resource becomes significant since it can meet these criteria due to its distinctiveness (Chughtai, 2013). The organizations always make them efforts to find, develop and retain valuable workforce who are talented and skilled employees that are precious resources to afford competitive advantages. Talent staff retainment has become more important than it used to be, and it became among most concerns of organization strategies about competitive advantages and organizational effectiveness with the best resources (Juhdi, Pa'wan, & Hansaram, 2013). Currently, the reciprocity responsibility norms were popularly applied in organizations in Viet Nam. However, a majority of organizations have not sufficiently invested interest and improve the system of human resource management (HRM) in the area of relation between HRM practices and employee engagement. Due to this reason, there remain areas that both employers and employees do not have in common. As a result, the consequence of this shortcoming in human resource management, to some extent, has caused employees were not motivated to compensate beneficial treatment and the organizations did not obtain expected business outcomes. To counter this problem, employers are trying to come up organizational talent management programs and promote employees by dealing with how to engage trained and highly experienced employees and enhance job performance. The broad range of employee engagement studies (e.g., Dajani, 2015;Ismail Hussein, 2019) have shown that job performance is considered as the most popular researched phenomenon in organizational behavior and indicated the significance and complexity of the issue impacting by its all dimensions (psychological, organizational significance and economic). They indicated that there are many major variables that impact on the job performance of employees. Moreover, employee engagement is considered a significant factor in improving job performance of employees to help organizations reach their objectives. By making employee more engaged in the organization, employers can be not only continuing to get effectiveness of developing employee programs but also to get the results of retaining skilled and experienced staff. So, the concerns of organization are paid more attention to increasing engagement levels of employees.
It is to say that the HRM system is considered as playing a significant role in enhancing engagement level and improving job performance of employees. Human resource management is "the policies and procedure involving HRM aspects of management regarding, planning, recruitment, selection, training and management, performance appraisal and labor relations" (Dessler, 2019). Some researchers (e.g., Chughtai, 2013;Juhdi et al., 2013;Khoreva, 2018) suggested that HRM practices should consider as good antecedents of employee engagement and job performance. HRM practices are more important because they exist and effect on daily operations of an organization in shaping employee's attitudes, cognition and behavior. HRM practices can help organizations enhance their performance (Khoreva, 2018). Researchers (e.g., Den Hartog et al., 2012;Russo, Mascia, & Morandi, 2018) explain HRM practices role that it could impact on employee perception that the creation of a safety climate, motivation, meaningfulness, emotion... According to Robbins and Judge (2017), "HRM practices are one of the factors in the organization that could shape and determine employee responses at work". HR practices can help for explaining how these practices influence employee engagement and job performance; help managers with solutions to encourage and motivate employees to work effectively in conditions of tight labor market, where changing knowledge management practices and global convergence of technology has redefined the nature of work, and in the context of limited talent resources in manpower market. So, examining the relationship between HR practices with employee engagement and job performance is necessary. Furthermore, HR researchers have been exploring how HRM practices operate among different types of individuals and in different types of situations. Cohen et al. (2003) claimed that understanding whether a theory explains a certain phenomenon the same way in different groups of individuals is central to furthering social science research. Similarly, Frazier, Tix, and Barron (2004) stated that understanding how theories operate among different types of individuals and for whom certain predictors lead to certain outcomes is central to developing theories of human behavior (Cohen et al., 2003). Additionally, examining how theories operate under various situations is said to be "at the heart of theory in social science" and an optimal way to refine theories (Frazier et al., 2004). So far, individual characteristics have been largely neglected within organizational behavior research. Some existing researches have examined the effect of HRM practices on employee engagement, but they did not take into account the potential moderating effects of individual characteristics (e.g., gender, educational level) on this relationship. It has not also clear evidence whether frontline employees perceive HRM practices in different ways depending on their demographics. Researchers (e.g., Brinck, Otten, & Hauff, 2019) suggested that in order to understand the impact of HRM practices on employee outcomes (e.g., employee engagement) within the context of specific industry (e.g., banking), research is required to understand why some individuals have a greater tendency to engage than others. Therefore, the author will integrate individual characteristics (e.g., gender, marital status) on the relationship between HRM practices and job engagement of frontline employee in this study.
In the trend of transitioning and developing Vietnam's economy, the service sector, consisting of the banking industry, has always played an important role. It is the heart of any economy. Firstly, banks act as delegated monitors and ensure that firms use the resources allocated to them effectively. Secondly, banks also play an important role as a tool for executing macroeconomic policies. Among many sectors, the banking sector is one of the most sectors has the fastest growth in human resources in both quantity and quality. The banking industry has the highest attraction with job-seekers during the last five years. The average growth in human resources of banking sector during the period 2014-2019 was about 15% each year (Tinnhanhchungkhoan, 2019). Go along with the quick changes in business environment, the banking industry has renewed itself to catch up with the requirements of bank users. Almost banks are offering many kinds of valuable services for customers besides its traditional services in order to upgrade the satisfaction of the bank's customers. Consequently, beyond the normal expectation of providing excellent customer service, the bank's employees are also required to sell various added-value services. Frontline employees who play role as service providers will provide those services to customers (Sultana & Johari, 2017). Thus, frontline employee in banking industry becomes a part of service products and their excellent performance help to form image of organizations. Until now, there is little theory or empirical investigation that accounts for the role of HRM practices, frontline employee engagement as a means through which organizations could create a competitive advantage. Particularly, not many researches examine the role of job engagement of frontline employees as a mechanism that connects organizational factors and employee characteristics to job outcomes. This can be explained why the situation is the same in Viet Nam, not many research and specific studies on HRM practices and job engagement of frontline employees in the Vietnamese banking industry. The majority of organizations in Viet Nam has not yet invested interest and improves their system of human resource management in this specific area. Due to these discussed reasons, this study will build a human resource management practice model in the Vietnamese banking industry to improve job engagement of frontline employees. Besides, the moderating role of gender and marital status in this relationship will also be examined. For the result, the study is expected to explain how the bank community should respond to challenges of human resources so as to get satisfactory frontline employees' performance and satisfactory business outcomes through the factors of HRM practices and employee engagement.

Human resource management practices
HRM practices are not a new conception. HR practices come from Human resources management. Human resource management as conceived in the 1980s had a conceptual framework consisting of a philosophy underpinned by a number of theories (Armstrong & Taylor, 2017). Or Watson (2010) offered about HRM: "the managerial utilization of the efforts, knowledge, capabilities and committed behaviors which people contribute to coordinated human enterprises as a part of an employment exchange to carry out work task in a way which enables the enterprises to continue into the future". Armstrong and Taylor (2017) argued about HRM that can be seen in a system that starts from HR philosophies that comprise overarching values and guiding principles to adopt in practices by managing people to reality activities. The HRM system is incorporated: HRM strategies, HRM policies and HRM practices which comprise all of HRM activities involved in developing and managing people and in overseeing the employment relationship. Juhdi et al. (2013) argued that organizations could be able to achieve their objectives depending on human capital through HRM to manage and make employees act, behave and think in a way to get the goals. And HRM practices could have significant influences on what kind of employee's perception, attitude and behavior can affect organizational performance. There are many models of human resource management practices that have been developed through previous studies (e.g., Rubel et al., 2018;Ueno, 2014). Components of this model depend on many factors of national culture and the level of national development, industry, type of business, background, and perspective of the founder of the company. The dimensions of HRM practices were used in researches are different and change. So, they indicated that HRM practices are different in various organizations. However, there are also common components used in many models of HRM practices with five basic components (e.g., Chand, 2010;Othman, 1999;Rubel et al., 2018;Tsaur & Lin, 2004;Ueno, 2014;Zerbe, Dobni, & Harel, 1998), they are: Training support; Performance appraisal; Career advancement; Reward and recognition, Recruitment and selection. The other components were not used frequently in prior studies. This proved that those factors are the core components of HRM practices that any organization is very interested in, especially in the banking sector -the global business sector, fierce competition, operating in volatile business environments and integration with the global market. It requires banks must have appropriate HRM practices. Therefore, the selection of these components to the research is also in line with the realities of HRM of the majority of banks today. Moreover, based on the result of focus group with five managers in banking management field and five frontline employees of banking sector, two-thirds of the participants agreed to add two components of HRM practices such as "employee participation" and "information sharing" which were considered to have a significant influence on employee engagement.

Employee engagement
Employee engagement is a concept that has gained popularity for a long period in both academic researchers and practitioners (Chughtai, 2013;Juhdi et al., 2013;Saks, 2006). Especially today, employee engagement is a top topic in human resource management, and it is considered with associating with positive and desirable outcomes in an organization. Kahn (1990) defined employee engagement as "the harnessing of organization members' selves to their work roles; in engagement, people employ and express themselves physically, cognitively, and emotionally during role performances". Rich, LePine, and Crawford (2010) stated that what engagement prefers to one-self in role that involves employees invest individually to complete and get the full self into their role performance. Employee engagement is relative to energetic, involvement and efficacy in their job (Maslach, Jackson, & Leiter, 1997). Macey and Schneider (2008) also debated that employees engaged in their tasks will have a higher level of energy, enthusiasm with their work and loyalty to their job from day-to-day. Schaufeli and Bakker (2004) also had definition of engagement. It's the psychological status that is relative to work positively and satisfactorily. It has characteristics with three levels: vigor, dedication, and absorption. They explained that vigor is a level of high energetic status, ready for efforts in their task consistently, stably to face difficulties. Dedication is higher level than vigor. It presents the closed relationship between private tasks and important emotions and enthusiasm, feeling deserved for their position and motivation. And the highest level is about absorption. It shows that the completely satisfaction on their work, present their concentration to task, spare time, cognate and enjoy their feeling with their tasks and detaching themselves from work.

HRM practices and employee engagement
HRM practices play a crucial role in improving employee engagement (Dessler, 2019). Juhdi et al. (2013) argued that "the critical role of HR practices as tools for organizations to make employees engrossed and engaged in their job performance by providing challenging job with available resources and opportunities for growth and management". The relationship of HR practices with employee engagement and suggestion of Saks (2006) using SET to explain the effects of employee engagement on antecedences, this study prefers the reciprocal exchange relations between organizations through HR practices and employee engagement. In that, employees who receive and perceive the fitness, career concern, fairness and benefits satisfaction from organization, in their turn, they will respond in the way that bring more engaged to organization. Therefore, the hypothesis is proposed.
H1: There is a positive relationship between HRM practices and employee engagement in the Vietnamese banking industry.

Recruitment and selection and employee engagement
Recruitment is those activities designed to attract and identify sufficient, qualified applicants to meet the principalship needs of the organization (Noe et al., 2010;Richardson & Denton, 2005). Besides, selection is the process of choosing best and qualified candidates from the available applicants that are most fit and likely to perform the job successfully (Byars & Rue, 2010;McCauley & Wakefield, 2006). Tinti et al. (2017) defined recruitment and selection practice "as an organization proposal to seek out employees, stimulate them to apply and select them, seeking to harmonize the person's competencies with characteristics and demands of the organization". Phillips (1998) stated that effective recruitment and selection will help improve the fit between the individual and the organization by providing applicants with suitable information or insight into the job's tasks needed for them to make informed decisions about their job. Therefore, according to the above discussions, the hypothesis is proposed.
H1a: There is a positive relationship between recruitment and selection and employee engagement (vigor, dedication, and absorption) in the Vietnamese banking industry.

Training and employee engagement
Training is generally thought of as a set of activities that are designed to improve the knowledge, skills, and abilities of employees in order to provide a better match between employee and job characteristics (Milkovich & Boudreau, 2008;Patterson et al., 2005). Liao and Chuang (2004) stated that service training programs should be set "clear behavioral standards across all aspects of a service encounter". It should also be designed in line with the specific needs of the service offered. Organizations should ensure that their employees "know and understand what they do and why, to be able to provide a high level of service to the customers" (González & Garazo, 2006). The association between the training practice and employee engagement is described by the norms of social exchange theory (SET) which proposes that frontline employees feel obligated if they realize that their bank has concerns for their progress. (Karatepe, 2013) stated that as a company invests resources through suitable training programs, employees may reciprocate by engaging themselves more to their duties. Based on the aforementioned discussions, the hypothesis is proposed: H1b: Training has a positive impact on employee engagement (vigor, dedication, and absorption) in the Vietnamese banking industry. Coens and Jenkins (2002) defined performance appraisal as a process by which managers evaluate, judge and describe the job performance of subordinates over a specified period of time, and the results of the evaluation are kept by the organizations for future reference. Apak et al. (2016) proposed that organizations should give feedback the results of performance appraisal to employees about their contributions in order to increase employees' performance because it enables employees to identify their anxiety and strengths that help to overcome their weaknesses, develop skills, and absorbed them towards their task. Therefore, the hypothesis is proposed as follows:

Performance appraisal and employee engagement
H1c: Performance appraisal has a positive impact on employee engagement (vigor, dedication, and absorption) in the Vietnamese banking industry.

Reward and recognition and employee engagement
Armstrong and Taylor (2017) defined a rewards system as the processes, policies, practices, and strategies for rewarding its employees in line with their abilities, contribution, and artifice by proving many suitable types, levels of pays, benefits, and other forms of reward. The rewards system exits in order to motivate employees to work towards achieving strategic goals. While most employees appreciate monetary rewards for job well done, many people simply need to be recognized as an individual or a member of a group when they did a good job. Harrison (2004) defined recognition as "the timely, informal or formal acknowledgment of a person's or team's behavior, effort or business result that supports the organization's goals and values, and which has clearly been beyond normal expectations". Truss et al. (2013) advocated that rewards and recognition can reinforce desired behaviors and simultaneously control behavior externally as well as prompt employees to modify their behaviors leading to higher performance outcomes. Rewards and recognition are also considered as critical practices that successful organizations use to enhance employee engagement (Crawford et al., 2013;Vance, 2006). With these arguments, the hypothesis is proposed as follow: H1d: Reward and recognition have a positive impact on employee engagement in the Vietnamese banking industry.
Electronic copy available at: https://ssrn.com/abstract=3655357 2.3.5 Career advancement and employee engagement Herzberg (1966) defined that career advancement is the degree to which a firm provides prospect for the progression of employees' career as well as employees have a clear career path ladder within their firm. Brown (2002) stated that career advancement mentions a series of developmental and progressive changes (e.g., higher salary, greater freedom to pursue interests) that occur in a person's career. The concept of career advancement can be considered as a platform that encourages employees look beyond their present works and prepare for brighter future positions within their organization (Byars & Rue, 2010). If organizations provide career advancement opportunities to their employees, this will trigger employees' positive attitudes and psychological safety and make them more motivated, dedicated and enthusiastic towards their job (Aktar & Pangil, 2017;Kahn, 1990). Therefore, the hypothesis is proposed as follows: H1e: Career advancement has a positive impact on employee engagement (vigor, dedication, and absorption) in the Vietnamese banking industry.

Employee participation and employee engagement
Employee participation refers to the part of the decision-making process in which employees are involved (Heery & Noon, 2008). The definition applies whether they are consulted individually or as a group by their leader who then makes the final decision, or they share the final decision making with the leader, or they are delegated responsibility for making the decision by the leader (Wilkinson, Gollan, & Marchington, 2012). According to Cantor, Morrow, and Montabon (2012), employee participation led to increased individual integration into the organization and employees became more involved in the work. Employee participation could make employees understand the basic need of creativity and offer viable and creative solutions for their job through transformed relationships with employees that enhance superior levels of their behavior outcomes (e.g., employee engagement) (Marchington & Kynighou, 2012). Therefore, the hypothesis is proposed as follows: H1f: Employee participation has a positive impact on employee engagement (vigor, dedication, and absorption) in the Vietnamese banking industry. Wood and Wall (2007) defined that information sharing as members within an organization obtain the correct information regarding business results, customer feedback, and quality. Chiang et al. (2008) argued that when employees experience higher information sharing, they are more motivated in their work to contribute towards the achievement of the organization's goals. Frontline employees represent the relationship between the organization and its customers. They usually contact customers and receive customers' responses to organization's products or services. In the case of banks, when customers intend to seek more information (e.g., credit card, balance checking, loans), they normally approach employees at the frontlines (e.g., teller, personal banker, customer service officer). Therefore, greater information sharing (e.g., imparting knowledge or offering advice to customers) may be useful for frontline employees in responding to customer's inquiry (Suan & Nasurdin, 2014) and help them build rapport with customers, which in turn may enhance the customer service experience (Gremler & Gwinner, 2008). Therefore, the hypothesis is proposed as follows:

Information sharing and employee engagement
H1g: Information sharing has a positive impact on employee engagement (vigor, dedication, and absorption) in the Vietnamese banking industry.

The moderating role of gender
A large body of research exists relating HRM practices to various work outcomes, including employee engagement. Researchers began to look beyond biological gender to the psychological characteristics and behaviors that distinguish men from women in a given society (Deaux & Major, 1987). The acquisition of gender roles occurs during the socialization process, whereby men and women develop attitudes, behaviors, emotional reactions, interests and motives that are culturally defined as appropriate for members of their gender (Bem, 1974). Gender roles are thought to be quite stable throughout the life span but may not manifest themselves in all situations and may be more or less salient under certain circumstances (Eagly, 1987). An assumption developed early in the history of gender role studies was that masculine and feminine were polar opposites on a continuum. The masculine role is associated with instrumental traits, such as self-assertive and goal-oriented. Femininity is conventionally associated with expressive traits, such as emotional and relationship orientation. Masculinity has been associated with work-related behaviors and attitudes, such as stronger feelings of personal accomplishment, higher occupational status, greater problem-solving skills and lower levels of stress, less perceived isolation, and more adaptive coping in workrelated environments (Hofstede, 2011). The attributes associated with femininity, such as the tendency to seek approval from others and to have avoidance and support-seeking coping styles (Gianakos, 2000;Mayor, 2015), are not as valued in many work environments and may hinder feminine individuals from achieving HRM practices. It seems plausible that the instrumental traits of the masculine gender role will enhance perception of HRM practices, consequently, job engagement. In other words, gender is likely to moderate the relationship between HRM practices and employee engagement. Therefore, the following hypotheses can be detailed: H2: Gender moderates the effect of HRM practices on employee engagement. The positive relationship between HRM practices and employee engagement will be stronger for males than females.

The moderating role of marital status
In this study, partnership status as a moderating characteristic with two states -married versus single. Specifically, in taskoriented settings, marriage is expected to operate as a negatively evaluated state for mothers. According to Correll et al. (2007), married mothers will face penalties while married fathers will reap rewards. However, in a subpopulation of single parents, I argue, the roles of caregiving and breadwinning will be dissociated from the gender of the parent, resulting in a significantly reduced difference in workplace outcomes for mothers and fathers. The scope of our definition and predictions are limited to heterosexual employees. Below the authors elaborate why and how partnership status is expected to further modify the interaction between HRM practices and employee engagement. Status connotations of marriage can be most directly linked to the notion of domesticity -the gender system which separates family work and market work (Williams & Dempsey, 2014). Domesticity as a gender system emerged around 1780 when men started working in factories and offices while women specialized in child-rearing and homemaking. In its original context, domesticity justified and reproduced the breadwinner (husband) and housewife (wife) roles. Despite the large share of women in the labor market, domesticity is still entrenched in Vietnamese society and operates as both a norm and a practice (Williams & Dempsey, 2014). It primarily manifests itself in two ways: first, it organizes market work around the idea of a worker who is primarily focused on work and takes no time for childbearing. Second, it marginalizes caregivers, impeding them from positions of authority (Williams & Dempsey, 2014). Even though the role of marriage in domesticity is more implied than overtly discussed, I argue that marriage is both an inseparable element of the domesticity system and a set of complementary roles that triggers expectations for behavior dictated by gendered norms. Marriage is the necessary condition that makes the separation between the two spheres-home and works-theoretically possible. Since married women and mothers, in particular, have been traditionally associated with homemaking and child-rearing while their partners earned money to support the family, they are expected to prioritize family work over market work. Due to the link outlined above, marital status is expected to produce negatively biased performance expectations for married mothers, which will result in diminished workplace outcomes (e.g., job engagement) for mothers. In service industry (e.g., banking), most of frontline employee were female, if they get married, the authors predict that married employees will significantly weaken the interaction between HRM practices and employee engagement will be significantly weakened. Therefore, the following hypotheses can be proposed: H3: Marital status moderates the effect of HRM practices on employee engagement. The positive relationship between HRM practices and employee engagement will be stronger for single employees than married employees.

Fig. 2. A conceptual framework
Control variables

H3
Electronic copy available at: https://ssrn.com/abstract=3655357 This research has been conducted exploring demographic variables (e.g., age, income, experience, and education level) in conjunction with employee engagement. the following hypotheses can be proposed: H4: There are statistically significant in the relationship between control variables (e.g., age, income, experience, and education level) and employee engagement in the Vietnamese banking industry.

Procedure and sampling size
The survey was conducted in Vietnamese banks. The sample was selected using a nonprobability sampling with techniqueconvenience. The target population of this study was frontline employees who work in Vietnamese banks, especially frontline employees who have over one-year experiences. The current study consisted mainly of two stages including qualitative and quantitative research. For qualitative research, the questionnaire was originally developed in English language and then translated into Vietnamese language by the authors with the assistance of English language experts. In the qualitative research, the Vietnamese version of the questionnaire was checked by in-depth interview method in one week with ten people who work in the Vietnamese banks to ensure if they understood the questions and revised Vietnamese terms which were unclear during due to translation. Based on the comments of respondents, the survey questionnaire was modified properly. For quantitative research, after the modifications for the questionnaire, the survey was issued to all respondents who work in the banks in Vietnam. Finally, 457 questionnaires were used as valid data for this study. The individual characteristics were presented in Table 1.

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 (Fornell & Larcker, 1981). Researchers have suggested the use of PLS-SEM is suggested due to its effective use as an analysis tool used to support prediction models from empirical data (Goodhue, Lewis, & Thompson, 2012;Hair et al., 2014). Moreover, PLS-SEM is appropriate for analysis when different measurement scales are used in the research model and it allows optimal empirical assessment of a structural model (Keil et al., 2000). 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 (Diamantopoulos et al., 2012). Therefore, data analysis for this study was performed using PLS-SEM with the SmartPLS 3.0 (Ringle, Wende, & Becker, 2015).

Results
Partial Least Squares-Sequential Equation Modeling was utilized to analyze the survey response for both the measurement and structural model. The measurement model identifies the variables with their associated latent variables. Before performing PLS-SEM estimation, the validity and reliability tests for the multi-item measures were assessed. First, convergent validity is the amount of variance when two or more items agree when measuring similar constructs and is calculated using factor loadings. Convergent validity is said to be reliable when the factor loadings are above 0.50 (Fornell & Larcker, 1981;Hair et al., 2014). The results indicated that there were 3 items out of 33 items for human resource management practices and 2 items out of 17 items for employee with lower factor loadings than the threshold of 0.50 (for example, factor loadings for IS5 = 0.499, RR5 = 0.408, RS4 = 0.498, ABS6 = 0.484, DED = 0.456). Thus, these items were removed from the analysis. Besides, 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" Fornell and Larcker (1981). Tables 2 was the correlation matrix of the constructs with the diagonal values. The square roots of the AVE values of each construct were indeed greater than the correlation coefficient with any other construct. Therefore, discriminant validity has been established for the constructs. Finally, the authors checked reliability of each construct. According to Bagozzi and Yi (1988), who indicated that the Cronbach's Alpha and composite reliability values should be 0.7 or greater to be considered reliable in a model, each variable was evaluated to verify reliability. From Tables 2 presented, it is clearly stated that all the constructs used in this study were reliable since it obtained the Cronbach's Alpha and composite reliability values more than 0.7 (For example, the minimum Cronbach's Alpha and composite reliability values were 0,784 and 0,861, respectively). Therefore, all values fall within the acceptable range to conclude good reliability. Moreover, multicollinearity was assessed for all of the constructs. Hair et al. (2014) suggested that indicators that show the presence of multicollinearity is a problem, as the indicator has the possibility of inflating bootstrap standard errors, thus increasing the probability of type II errors, or failing to detect that an effect is present in the research. The authors used a 5.00 tolerance level for Variance Inflation Factor (VIF) Hair et al. (2014), no items were outside the threshold (maximum outer VIF of constructs = 3.437). Additionally, to ensure the predictors are not collinear, the block Variance Inflation Factor is calculated for all predictor variables. All inner VIF were below 5.0 (maximum inner VIF of constructs = 2.421), and as such collinearity of the variable is not a concern. The hypotheses were tested for significance, strength, and polarity for the inner model reporting. This was accomplished in SmartPLS 3.0 using bootstrapping for 500 samples. The results are displayed in Fig. 3 and Table 3, showing hypothesis, path coefficients, P-values and whether the hypothesis is accepted or rejected based on the model analysis. The figures were calculated and assessed according to the above-named criteria of p-values being significant if they were 0.05 or smaller. Table 2 Correlations of constructs, internal consistency, convergent and discriminant validity Hypothesis 1: the result showed that HRM practices had a positive and significant relationship with employee engagement in the Vietnamese banking industry, (p-value = 0.000 and beta coefficient = 0.593). The result indicated that the more effective HRM practices, the greater is the possibility that frontline employees will engage in their tasks. Thus, hypothesis 1 was supported. Educational level -0.006 No Note: ***=p < 0.001; **= p < 0.01; *= p < 0.05 Gender*HMRP = Moderating Effect 1; Martial*HMRP = Moderating Effect 2 Gender (male was the reference), marital status (married was the reference) ABS = absorption, CA = career advancement, EP = Employee participation, HMRP = Human resource management practices, IS = information sharing, PA = performance appraisal, RR = reward and recognition, RS = recruitment and selection, TR = training, VIG = vigor Hypothesis 1a: the result showed that recruitment and selection had not a significant relationship with employee engagement in the Vietnamese banking industry, (p-value = 0.842 and beta coefficient = -0.012). Besides, recruitment and selection showed a positive relationship with absorption and dedication, and a negative relationship with a vigor which means that frontline employees who had good perception of recruitment and selection practices tended to show a higher level of absorption and dedication, and a lower level of vigor. However, these relationships were not statistically significant (p-value for absorption, dedication, and vigor were 0.478; 0.475; 0.183 with beta coefficient of them were 0.044; 0.045; -0.076, respectively), which mean that there is a high potential that these relationships may occur purely by chance. Therefore, hypothesis 1a was rejected.

Fig. 3. Structural Model
Hypothesis 1b: the result showed that training had a positive and significant relationship with employee engagement in the Vietnamese banking industry, (p-value = 0.000 and beta coefficient = 0.200). The result indicated that the more effective training practice, the greater is the possibility that frontline employees will engage in their tasks. Besides, they showed a higher level of absorption, dedication, and vigor (p-value for absorption, dedication, and vigor were 0.037; 0.005; 0.000 with a beta coefficient of them were 0.125; 0.160; 0.227, respectively). Moreover, the result revealed that training practice had a stronger effect on vigor than absorption and dedication. Thus, hypothesis 1b was supported.
Hypothesis 1c: the result showed that performance appraisal had not a significant relationship with employee engagement in the Vietnamese banking industry, (p-value = 0.703 and beta coefficient = 0.022). Besides, performance appraisal showed a positive relationship with dedication and vigor, and a negative relationship with absorption which means that frontline employees who had good perception of performance appraisal practice tended to show a higher level of dedication and vigor and a lower level of absorption. However, these relationships were not statistically significant (p-value for absorption, dedication, and vigor were 0.602; 0.279; 0.606 with beta coefficient of them were -0.032; 0.067; 0.032, respectively), which mean that there is a high potential that these relationships may occur purely by chance. Therefore, hypothesis 1c was rejected.
Hypothesis 1d: the result showed that reward and recognition had a positive and significant relationship with employee engagement in the Vietnamese banking industry, (p-value = 0.000 and beta coefficient = 0.219). The result indicated that the more effective reward and recognition practices, the greater is the possibility that frontline employees will engage in their duties. Besides, frontline employees showed a higher level of absorption, dedication, and vigor (p-value for absorption, dedication, and vigor were 0.000; 0.002; 0.000 with a beta coefficient of them were 0.295; 0.150; 0.160, respectively). Moreover, the result revealed that reward and recognition practices had a stronger effect on absorption than vigor and dedication. Thus, hypothesis 1b was supported.
Hypothesis 1e: the result showed that career advancement had not a significant relationship with employee engagement in the Vietnamese banking industry, (p-value = 0.770 and beta coefficient = -0.019). Besides, career advancement showed a positive relationship with absorption and vigor, and a negative relationship with a dedication which means that frontline employees who had good perception of career advancement practice tended to show a higher level of absorption and vigor and a lower level of dedication. However, these relationships were not statistically significant (p-value for absorption, dedication, and vigor were 0.981; 0.350; 0.905 with beta coefficient of them were 0.002; -0.069; 0.007, respectively), which mean that there is a high potential that these relationships may occur purely by chance. Therefore, hypothesis 1e was rejected.
Hypothesis 1f: the result showed that employee participation had a positive and significant relationship with employee engagement in the Vietnamese banking industry, (p-value = 0.000 and beta coefficient = 0.234). The result suggested that the more effective employee participation practice, the greater is the possibility that frontline employees will engage in their tasks. Besides, they showed a higher level of dedication and vigor (p-value for dedication and vigor were 0.000; 0.000, with a beta coefficient of them were 0.309; 0.263, respectively). Moreover, the result revealed that employee participation had a stronger effect on dedication than vigor. Thus, hypothesis 1f was supported in general. However, employee participation didn't impact on absorption (p-value = 0.115 and beta coefficient = 0.098).
Hypothesis 1g: the result showed that information sharing had a positive and significant relationship with employee engagement in the Vietnamese banking industry, (p-value = 0.000 and beta coefficient = 0.232). The result indicated that the more effective information sharing practice, the greater is the possibility that frontline employees will engage in their jobs. Besides, they showed a higher level of absorption, dedication, and vigor (p-value for absorption, dedication, and vigor were 0.000; 0.015; 0.000 with a beta coefficient of them were 0.242; 0.162; 0.221, respectively). Moreover, the result revealed that information sharing practice had a stronger effect on absorption and vigor than dedication. Thus, hypothesis 1g was supported.

The moderating role of gender and marital status
Hypothesis 2 predicted that gender would moderate the relationship between HRM practices and employee engagement. The study showed that the moderating effect 1 between gender and HRM practices with employee engagement is positive and statistically significant (p-value = 0.014 and beta coefficient = 0.112). This discovery proposed that frontline employees with male gender positively moderated the relationship between HRM practices and employee engagement. In other words, Frontline employees with male gender heighten the positive effect of HRM practices on employee engagement is stronger with male employees (Fig. 4). Therefore, Hypothesis 2 is supported. Hypothesis 3 expected that marital status would moderate the relationship between HRM practices and employee engagement. The study showed that the moderating effect 2 between marital status and HRM practices with employee engagement is negative and statistically significant (p-value = 0.009 and beta coefficient = -0.100). This discovery proposed that married frontline employees negatively moderated the relationship between HRM practices and employee engagement. In other words, married frontline employees weaken the positive effect of HRM practices on employee engagement (Fig. 5). Therefore, Hypothesis 3 is supported. For the effect of the control variables, the result indicated that age positively related to employee engagement (p-value = 0.027 and beta coefficient = 0.078), which means that elder employees tended to engage in their duties than young employees. However, this study didn't find that the relationship between employee engagement and income (p = 0.327; β = -0.034), experience (p = 0.609; β = 0.018), and educational level (p = 0.913; β = -0.006).

Model fit
The coefficient of determination (R 2 ) is a measure of the model's predictive power. R 2 the amount of variance in the endogenous (dependent) latent variables in the structural model explained by the exogenous (independent) constructs connected to it (Hair et al., 2014). 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 whose arrows point to it (Hair et al., 2014). The high R 2 coefficient also reveals that the values of the variables can be well predicted by the PLS path model (Hair et al., 2014). The R 2 value for employee engagement (0.498) indicates that 49.8% of the total variation of the endogenous construct employee engagement may be explained by the exogenous construct, HRM practices (Fig. 6). Moreover, Cohen (1988) suggested that R 2 values and the effect for endogenous latent variables in behavioral sciences be assessed as 0.26 (large effect), 0.13 (moderate effect), and 0.02 (weak effect). Because R 2 values for employee engagement was greater than 0.26. Therefore, the model of this study proved the model-data fit.

Conclusion
The purposes of this study are testing the relationship of HRM practices on employee engagement and examining gender and marital status as moderating variables in the context research such as frontline employees who are working in the Vietnamese banking industry. This study is around two concept's constructs: HRM practices, and employee engagement. In there, an initial model of HRM practices consists of 7 components: Recruitment and selection, Training; Performance appraisal; Career advancement; Reward and recognition, Employee engagement, and Information sharing. In this study, measurement scales of all concepts have been adjusted and developed to match the research context of frontline employees in the Vietnamese banking industry. By testing with data collected from 457 questionnaires, the set of scales consists of 7 components of HRM practices, and three dimensions of employee engagement with total 50 observed variables to better suit for research about frontline employees. The set of scales after adjustment can continue to be used for further studies. Additionally, the PLS-SEM results of the research model showed that this model is suitable for market data and most of hypotheses in the research model are accepted. So, the purposes of study are achieved. The study results provided strong support to the notion that HRM practices significantly influence employee engagement, as suggested in previous studies of Saks (2006), Crawford et al. (2013), (Suan & Nasurdin, 2014). Besides, there was 4 out of 7 dimensions of HRM practices had significant impact on employee engagement. So, this indicated that HRM practices with 4 components (Training; Reward and recognition, Employee engagement, and Information sharing) are crucial and significant to employee in creating and boosting the sense of engagement. Regard to the path coefficient among variables, employee participation (β = 0.234) has been defined as the strongest predictor of employee engagement in banking sector in Vietnam, followed by information sharing (β = 0.232), reward and recognition (β = 0.219), and training (β = 0.200) respectively. When organizations through HRM practices can create perception in employees about care, fairness and willing facilitating resources to provide employee to control and perform their roles and employees feel satisfied with HRM practices, they will reciprocally act to organization through positive behaviors as engagement.
Nowadays, the increasing numbers of women occupying professional positions prompted social scientists, psychologists, and employers to consider how the changing status of women affected how they experienced their work. The inconclusiveness of studies examining the function of biological gender, marital status in determining job engagement led researchers to look beyond this demographic variable to the psychological characteristics and behaviors that distinguish men from women in a given society. This study proved an empirical evidence about the moderating roles of gender and marital status. As the result indicated that male frontlines employees heightened the positive relationship between HRM practices and employee engagement, but this positive relationship was weakened by married employees. Electronic copy available at: https://ssrn.com/abstract=3655357

Managerial implications
The findings of this study contributed to the understanding of HRM practices, which play very important roles in frontline employees' working behaviors (e.g., employee engagement). Based on the outcomes of this research, the authors concentrate on suggesting several practical implications for managing effective plans of HRM within banking industry through four key dimensions of HRM practices. First, the results showed that employee participation had the strongest impact on job engagement of frontline employees. Bank managers should consider proactively managing and developing HRM practices and employee engagement. They should provide opportunities for frontline employees to participate in decision-making activities. Employee participation in the workplace will boost feelings of involvement and it is through this mechanism that employee engagement is improved. In other words, when employees have more involvement in their organizations, they will be promoted to fulfill higher levels of their duties. Second, information sharing had a second impact on employee engagement. To enhance the level of job engagement of frontline employees thought information-sharing practices, the banks should apply suitable strategies in information sharing policies. Banks should encourage employee sharing experience and helping each other to solve their problems. In addition, a good treatment of supervisors, co-worker may have a favorable exchange relationship between employee and organization. Increasing the relationship between supervisor and employee is an important element that creates the cohesion of working place. Staff and managers will have a strong connection with each other; by that, problems of employee will lead to problems of manager and the supporting of managers for employee will be increased strongly. Establishing outdoor activities for relaxing, these activities also help staff, and supervisor gets along with each other. In addition, it is necessary for employees and managers to spend time to arrange the knowledge relates to work, manager can Notes: ***=p < 0.001; **= p < 0.01; *= p < 0.05 Standardized coefficients are reported Solid line represents the relationship that is statistically significant share their experiences and mistakes for employee. Through this time, the relationship between employees and supervisors will closer and learning more experience in their work. Next, reward and recognition had the third impact on employee engagement. Bank managers need to set up a fair and effective reward and recognition practice so that they could promote their employees to work hard. When employees perceive reward and recognition as justice in their organization. If they feel there is fairness in how they perform in their roles, they will be more engaged in their roles. But if there is lack of justice, they will withdraw them out the engagement. Additionally, Bank needs to add more bonus programs to promote employee. Sales officers are grouped in team. Each team is responsible for a specific piece of work. Every employee in a team will work together for finding and maintaining their customers. After a period, the team with the highest growth index on customers and loans will have a bonus. This program can create the ability to work in team of employee. Another type to encourage workers to devote their energy to tasks is recognition. To create a recognition award, each department should have a proclamation board to praise employee for their good performance. At the end of every month or every quarter, employee who has the best performance will be honored in front of staffs and receive a present to encourage for their hard-working.
Fourth, one method to enhance the engagement of employees is changing and restructuring the training and development program. Training is considered as the fourth factor that enhancing employee engagement. For new employees, updating the orientation, adding more organization information, which can help to explain organization policy and procedure, describe how organization operates for newcomers. In addition, through the orientation, new employees can understand clearly about organization's target and know how they can contribute their performance to achieve organization's mission. For current employees, maintaining training is necessary to review their knowledge relating to work. Moreover, it also helps staff to enhance their skills to get high productivity at work. However, in the present situation, there are many complaints about the quality of training, it takes too much time for training in a working day and employees usually neglect in class. To get high benefits from training investments, banks should explore digital technology in training. It is a good idea to establish digital classes that satisfying both the need for improving knowledge and convenience in time. This study recommends creating online classes, using technology to offer self-spaced which employees can learn at home or learn in their free time. These classes not only help to reduce the cost of training but also help employee to manage their time and enhance their knowledge. In addition, thus every position has its own specialized knowledge, it is encouraged that each department sets up their own training time for maintaining and consolidating their employee skills and knowledge by sharing experience and learning together. Consequently, the engagement of employees to the organization is increased.
Finally, this study suggested that gender role identification moderated, or influenced the strength and direction, of the HRM practices and employee engagement relationship. The findings of this study provided evidence that masculinity is an important moderator of HRM practices in affecting overall employee engagement. Besides, this study also indicated that married employees tended to weaken the positive effect of HRM practices on employee engagement. With more women in the workplace, activities typically thought of as women's responsibility are now being carried out by men, and vice versa. The nature of gender roles is that they change slowly, if at all. While it may be a necessity to perform tasks typical of the opposite gender, our socialization to identify with our own gender's role oftentimes makes this behavior feel unnatural and stressful. At home, men are now taking on more typically feminine activities, such as caring for children. To succeed in the workplace, banks should encourage women behaving in ways that contradict socially accepted feminine traits, such as when the situation calls for them to be assertive to get their own needs met but are socially inclined to attend to the needs of others. The legacy of most workplaces to reward more masculine behaviors places more strain on those socialized and expected to portray feminine traits whether they are men or women. The masculinity of most work settings combined with the similarity of men and women's job engagement suggest that it is the identification with the masculine gender role, and not being a man, that is advantageous to feelings of congruence between one's own tendencies and the work environment.

Limitations and recommendations for future research
The current research has been conducted with certain limitations. First, the study has applied a convenient sampling method. Besides, the research only investigated the frontline employees in the Vietnamese banking industry. Further research should investigate general employees in other sectors by using probability sampling. They also compare between those different sectors to add more understanding of HRM practices in Vietnam. Second, HRM practices and employee engagement are large fields. HRM practices have been selected with 7 basic components but there are various dimensions that could be chosen to analysis and examine their roles. Besides, further research should consider some outcomes of employee engagement, such as individual performance, organizational citizenship behaviors, deviance.