Sustaining Affective Commitment and Extra-Role Service among Hospitality Employees: Interactive Effect of Empowerment and Service Training

The success of hospitality sector firms depends largely on the degree to which frontline employees offer a high-quality service to customers. In this context it is of vital importance to determine how to promote positive employees’ attitudes and behaviors through the human resource practices. This study analyzes the indirect effect of empowerment via affective commitment on the extra-role service. In addition, the moderating effect of service training on the empowerment–affective commitment–extra-role service relationship is explored. The data was collected from a sample of frontline employees working in three-to-five-star urban hotels in Santiago de Compostela (Spain). The results of the study suggest that for empowerment to lead to greater emotional attachment to the organization and extra-role customer service, it is necessary for frontline employees to perceive that they have received a moderated-high level of training in customer service. This finding is particularly interesting for frontline employees’ management in hotels sector.


Introduction
Nowadays, hospitality firms face to a complex and highly competitive environment together with more demanding tourists who call for personalized experiences. So, delivering a high-quality service generating a positive customer's experience is fundamental to their success and sustainability [1][2][3][4]. The frontline employees are the ones who with their behaviors promote the customers' perception about the service quality and achieve their satisfaction and loyalty [5][6][7]. Also, these employees play an important role in the innovation of services in hospitality firms [8].
In the service sector, such as hospitality, it is difficult to anticipate and formalize completely all of the frontline employees' behaviors that can generate a satisfactory experience for customers [9]. Therefore, it is required that frontline employees take the initiative to expand their roles, going beyond the formal requirements of their job and performing everything possible to solve customer's problems, adapting for the specific challenges and providing a high-quality service [4,7]. In other words, they need to perform an extra-role customer service.
Frontline employees' empowerment (i.e., "the freedom and ability to make decisions and commitments" [10] (p. 67)) plays a key role in facilitating the satisfaction of the customers' needs in the hospitality sector [2]. It is believed that empowered employees, due to their greater decision-making power, are more able and ready to offer a high-quality service to customers [11]. Empowering frontline employees gives them the flexibility for adapting the service to the customers' requirements without having to request the superior's approval [12,13], favoring customization [14]. Empowerment allows employees to quickly deal with service problems, respond to unanticipated demands [14][15][16] and live up to customers' expectations [5,17].
However, despite its potential benefits, it has been pointed out that empowerment is a challenging practice that does not always generate the expected outcomes [2,16]. Some empirical evidence shows inconsistencies or contradictory outcomes regarding the influence of empowerment on variables such as customer's complaint handling performance, service performance or customer's perceptions of service quality [18,19]. Consequently, new studies are needed to analyze how and when empowerment affects to outcomes.
Existing research in the hospitality field has found that human resource practices affect to employees' behavior through motivational and attitudinal variables (e.g., [7,20,21]). However, because of aggregating the practices in a single index, these studies cannot discern the unique effects of the different practices [22]. In this sense, recently, Kruja et al. [23] highlighted the scarcity of studies in the hospitality sector focused on empowerment and its consequences.
In recent years, employee well-being has been established as an important variable mediating the effects of human resource practices on employee performance [22,24], as well as an end in itself. Affective commitment-the emotional bond, the identification, and involvement of the employee with the organization [25]-constitutes a dimension of happiness at work, along with job satisfaction and engagement [26]. Framed in this line of research, and in order to try to fill this gap in the hospitality literature, this study analyzes the role of happiness well-being, depicted by affective commitment [26,27], as intervening variable between frontline employee empowerment and extra-role service.
It has been argued that direct employee-client interaction, characteristic of the hospitality sector [28], demands that employees have greater discretion and control, so that they can help meet customers' needs and quickly solve their problems. However, the effectiveness of empowerment continues to be a subject of debate [16]. Some authors have warned that empowered employees, who must make a great effort, could perceive greater workload, and consequently less well-being [18]. According to Marchington and Wilkinson, empowerment "may merely produce greater work intensification, increased stress levels, and redundancies" [29] (p. 399). Grant and Parker [30] mentioned that autonomy (similar to empowerment) may not generate the desired effects by placing pressure on employees. In the hospital industry, Chia and Chu [31] noted that empowered employees may face higher demands that will result in increasing levels of strain. The greater responsibility to be assumed by employees could cause anxiety if they do not possess the knowledge and skills to face it. Living up with customers' expectations is probably too demanding if employees don't have the necessary service training [1,32].
Although the importance of the training has been widely recognized and supported by research in hospitality and service sectors [1,8,33,34] and it has been argued that this is necessary to be able to realize the benefits of the empowerment [7,23], surprisingly, we have not found any study that have examined the interactive effect of both practices in the hospitality sector. In fact, we have only found two studies [9,35] examining the contingent role of training, but none in the hospitality industry.
To address the aforementioned research gap, our study proposes that the effect of empowerment on frontline employees' affective commitment and, consequently, on their extra-role service, will be moderated by service training. At the same time, we respond to Kostopoulos' [11] recent call for new studies that analyze the moderating role of training on the empowerment-performance relationship.
We believe that our research is significant for two reasons. On the one hand, by providing new evidence of the effect of empowerment on extra-role service behavior, via affective commitment, this work tries to address the lack of research pointed out by some scholars on empowerment and its individual outcomes in the field of hospitality. On the other hand, and perhaps more important, by explaining under which conditions empowerment is more effective and providing a more comprehensive view of the synergistic effect between empowerment and service training, this work try to offer interesting insights for frontline employees' management in hotel contexts. Although these two human resource practices are considered fundamental for the attitudes and performance of contact employees in the hotel sector [1,4,7,20,21,34], its interactive effect remains virtually unexplored.
The proposed model, which involves a moderated mediation process (see Figure 1), was tested on a sample of frontline employees from three-to-five-star hotels located in Santiago de Compostela (northwestern Spain).
Sustainability 2019, 11, x FOR PEER REVIEW 3 of 15 these two human resource practices are considered fundamental for the attitudes and performance of contact employees in the hotel sector [1,4,7,20,21,34], its interactive effect remains virtually unexplored. The proposed model, which involves a moderated mediation process (see Figure 1), was tested on a sample of frontline employees from three-to-five-star hotels located in Santiago de Compostela (northwestern Spain).

Theoretical Development
The self-determination theory (SDT, [36]) and the social exchange theory (SET, [37]) provide the theoretical frameworks that will guide the formulation of hypotheses.
From the SDT perspective, the satisfaction of the psychological needs of autonomy (sense of volition), competence (perceive oneself as capable) and relatedness (feeling of being connected to others) is key to autonomous regulation [38]. The concept of autonomous regulation overlaps considerably with that of affective commitment, since both represent the desire to follow a course of action [39,40]. Gagné et al. [39] demonstrated that when employees carry out an activity because they are autonomously motivated, over time they develop attachment to their organization. At least two studies [41,42] found positive associations between the satisfaction of psychological needs and affective commitment. Since empowerment and service training contribute to the satisfaction of the basic psychological needs (autonomy and competence), it is to be expected that they will facilitate the development of affective commitment, which in turn will encourage employees to perform an extrarole service [34].
Social exchange theory [37], and the norm of reciprocity [43], states that employees who perceive that the organization's actions are beneficial to them tend to feel obligated to reciprocate becoming more affectively committed to it. Empowerment and service training point out to employees that their organization considers them valuable resources for their survival and success [44,45]. As a result, employees will be more committed to their organization and willing to make a greater effort to benefit it [46][47][48].

Empowerment, Affective Commitment, and Extra-Role Customer Service
There are two main conceptions of empowerment. One considers empowerment as a human resource practice and another as a psychological state [11,49]. In this study, we consider empowerment as a management practice aimed at providing employees with the necessary authority to respond and satisfy possible customer demands [7,14,18,20].
Employee empowerment is considered one of the most relevant human resource practices promoting the necessary employees' behaviors for the successful running of service firms [9,11,15,23]. However, most empirical evidence shows that its influence on employee's behaviors occurs through happiness well-being variables, such as job satisfaction, affective commitment, or work engagement [2,5,7,12,20,50].
Karatepe et al. [20] pointed out that firms giving discretion to their employees enhance their feelings of autonomy. Need of autonomy, within the framework of the self-determination theory

Theoretical Development
The self-determination theory (SDT, [36]) and the social exchange theory (SET, [37]) provide the theoretical frameworks that will guide the formulation of hypotheses.
From the SDT perspective, the satisfaction of the psychological needs of autonomy (sense of volition), competence (perceive oneself as capable) and relatedness (feeling of being connected to others) is key to autonomous regulation [38]. The concept of autonomous regulation overlaps considerably with that of affective commitment, since both represent the desire to follow a course of action [39,40]. Gagné et al. [39] demonstrated that when employees carry out an activity because they are autonomously motivated, over time they develop attachment to their organization. At least two studies [41,42] found positive associations between the satisfaction of psychological needs and affective commitment. Since empowerment and service training contribute to the satisfaction of the basic psychological needs (autonomy and competence), it is to be expected that they will facilitate the development of affective commitment, which in turn will encourage employees to perform an extra-role service [34].
Social exchange theory [37], and the norm of reciprocity [43], states that employees who perceive that the organization's actions are beneficial to them tend to feel obligated to reciprocate becoming more affectively committed to it. Empowerment and service training point out to employees that their organization considers them valuable resources for their survival and success [44,45]. As a result, employees will be more committed to their organization and willing to make a greater effort to benefit it [46][47][48].

Empowerment, Affective Commitment, and Extra-Role Customer Service
There are two main conceptions of empowerment. One considers empowerment as a human resource practice and another as a psychological state [11,49]. In this study, we consider empowerment as a management practice aimed at providing employees with the necessary authority to respond and satisfy possible customer demands [7,14,18,20].
Employee empowerment is considered one of the most relevant human resource practices promoting the necessary employees' behaviors for the successful running of service firms [9,11,15,23]. However, most empirical evidence shows that its influence on employee's behaviors occurs through happiness well-being variables, such as job satisfaction, affective commitment, or work engagement [2,5,7,12,20,50].
Karatepe et al. [20] pointed out that firms giving discretion to their employees enhance their feelings of autonomy. Need of autonomy, within the framework of the self-determination theory (SDT), refers to the sense of being able to direct and determine one's own behavior [36]. According to Van den Broeck et al. [51], this need can be fulfilled by allowing employees to make personal choices or by the internalization of external choices. Since empowerment allows employees to take the initiative to respond to issues that arise in their workplace, it fosters a sense of autonomy.
According to the SDT, since empowerment contributes to the satisfaction of this basic psychological need it should lead to greater affective commitment. That is, when employees perceive that the organization gives them power in decision making, they will feel that their basic need is met and develop a greater attachment to their organization [41,42]. Greguras and Diefendorff [41], in a heterogeneous sample, showed the positive influence of the autonomy need satisfaction on affective commitment. For their part, Marescaux et al. [42] corroborated the positive impact of the satisfaction of the needs for autonomy and relatedness on the affective commitment.
Moreover, empowering frontline employees means valuing their contribution to the firm [44], while indicating the intention to establish a long-term relationship with them. Additionally, by providing empowerment to their frontline employees, the firms are giving them the opportunity of using their skills and capabilities [5]. Therefore, from the SET perspective [37], empowerment initiates a process of social exchange in which employees respond reciprocally by increasing their affective commitment. This positive attitudinal response, in turn, implies a positive behavioral response. According to Kehoe and Wright [48], the influence of affective commitment on behaviors is due to, at least, two reasons. First, the attitudinal response alone may not compensate for the treatment received by the organization. Second, employees who are affectively committed to their organization are more likely to act in ways that benefit their organization. One of these ways is to expand its role to perform a high-quality service in order to solve the customers' demands. Kazlauskaite et al. [2], in a sample of customer-contact employees in the hospitality sector, showed that the positive effect of authority granting on customer-oriented behavior was via affective commitment. Based on the above arguments and evidence, it is proposed the following hypothesis: Hypothesis 1. Empowerment is indirectly related to extra-role customer service via affective commitment.

The Moderating Effect of Service Training
Another practice with a relevant role in promoting employees' positive attitudes and behaviors is service training [1,7,20]. Indeed, in the field of hospitality, training has been linked to greater employee satisfaction, commitment, and positive behaviors, such as role and extra-role performance, and innovative service behavior [1,8,17,34,52,53]. Training contributes to the sustainability of firms in general [54] and of the hospitality sector, in particular [55].
According to Brown and Sitzmann [56], the benefits of training derive both from its contribution to the expansion of employees' skills and knowledge and from the increase in their effort and commitment to reciprocate the organization.
Service training promotes the frontline employees' affective orientation to the organization, and, through it, influences their behaviors aimed at achieving customer satisfaction and loyalty [1,12,50]. Service training provides frontline employees with the skills and abilities required to perform successfully their job, satisfy the customers' demands and deliver a high-quality service [7,12,15,57].
Therefore, from the perspective of the SDT, service training is expected to promote frontline employees' competence feelings [1,20,42] and their "perception of being effective in the interaction with the environment" [45] (p. 372). Haccoun and Saks [58] suggested that training reinforces the self-efficacy beliefs. When frontline employees feel confident and capable, they will be willing to take on new responsibilities and challenges [59].
Contrary, the lack of service training programs can lead to feelings of incompetence and frustration. According to the SET, service training can be perceived by the frontline employees as an investment and as a signal of the firm's commitment with them [12,44,57]. Consequently, frontline employees reciprocate that investment showing a higher level of affective commitment and a high-quality service [1,46].
Several authors have pointed out the synergic effect of empowerment and training [7,12,20]. These practices separately do not achieve a significant impact on the happiness, well-being, and the behaviors of the employees, as well as on the organizational effectiveness. In the words of Karatepe: "Empowering employees without training programs would not produce the intended outcomes or training employees without empowerment for effective complaint management would be useless" [7] (p. 133).
Moreover, employees' motivation will depend on the belief that their actions will produce results that they consider valuable and attractive [60]. When frontline employees consider that they have the necessary knowledge and skills to satisfy the customers' demands, they will feel valued, perceive empowerment more positively and, consequently, experience greater affective commitment. Training also helps employees perceive that they can achieve their career goals [8,34,45], which will motivate them to take on new challenges.
On the contrary, if frontline employees are empowered but they consider that they lack the skills, abilities and knowledge to perform their job role, their well-being could be negatively affected. Training is usually considered under the discretionary control of the firm and, therefore, an indicator of organizational support perceived by employees [61]. When training opportunities are scarce, frontline employees may consider that they do not receive enough support from the firm, while increasing their responsibilities. For example, Chan and Lam [18] found that empowered employees perceived greater workload, a construct related negatively with employee's affective commitment [62]. On the basis of the aforementioned arguments, the following hypothesis is formulated: Service training moderates the positive relationship between empowerment and affective commitment.
Based on previously hypothesized relationships, it can be expected that the indirect effect of empowerment on extra-role customer service through the employee's affective commitment will be moderated by the service training. When frontline employees perceive that they have received service training, empowerment will have a more beneficial effect on the affective commitment in such a way that they will perform discretionary behaviors tending to deal with the customers' demands and complaints. However, when service training is low, frontline employees receive inconsistent or contradictory messages (on the one hand, discretion and responsibility are encouraged; on the other hand, they are not provided with the necessary resources). In this situation, their emotional attachment to the organization is likely to weaken and, as a result, they are less willing to go beyond the call of duty. Therefore, the following hypothesis is proposed: Hypothesis 3. Service training moderates the indirect effect of empowerment on extra-role customer service (via affective commitment). Specifically, this positive relationship will be stronger when employees perceive that they have received higher levels of service training than lower.

Sample
The tourism sector is one of the most important for the Spanish economy. In fact, the latest data available from the INE (Instituto Nacional de Estadística-National Statistics Institute) reports a contribution of tourism sector to GDP of 11.7% and employment of 12.8%. With respect to Galicia, tourism contributed 10.4% to GDP and 11% to employment.
We chose to carry out the study in Santiago de Compostela, because although it is a small city (approx. 100,000 habitants), it receives a high number of tourists each year. Declared a World Heritage City by UNESCO in 1985, this city is the main urban destiny in the region of Galicia (Northwestern Spain). The city stands out for being one of the three main centers of pilgrimage of Christianity. The number of overnight stays registered in 2017 amounted to 1,444,919 (INE, Instituto Nacional de Estadística-National Statistics Institute).
In order to control the possible effects of different sectors, this research focused on hotel frontline employees [2]. Since service orientation is influenced by the hotel category [63], all the three-to-five-star hotels in the city were selected. In December 2017, according to the data of the Tourist Information Registry of Galicia, Santiago de Compostela had 4 five-star, 10 four-star, and 21 three-star hotels. In addition, it was argued that human resource practices may vary depending on the group of employees they are targeting [64]. In this study we focused on frontline employees, i.e., those in direct contact with the customer.
Initially, the management of all the three-to-five-star hotels that make up the sampling frame was personally contacted by one of the researchers (by phone or face to face) to explain the purpose of the investigation and request their participation. We obtained permission to conduct the survey in the 4 five-star hotels, 7 of the four-star hotels and 5 of the three-star hotels.
One of the researchers, with the support of the hotel administrative staff, conducted the data collection. Employees were asked to complete a self-administrated questionnaire. Once the pencil-and-paper based questionnaire was completed, employees handed it in a sealed envelope to the support staff. One of the researchers picked them up personally at the hotel.
At the beginning of the questionnaire, the objective of the study was briefly explained, indicating that participation was voluntary, and that the data will be treated confidentially.
Of the 122 questionnaires carried out, 19 had to be eliminated because they were incomplete, being the final sample of 103 questionnaires. The maximum estimated sampling error with a sample size of 103 was ±7.33% (p < 0.05). This value is similar to that of other studies (e.g., [20]). Respondents' average age was 35.46 years (SD = 10.09) and the 52.4% were women. In addition, the 54.4% of sample had higher education and the average organizational tenure was 69.83 months (SD = 94.42).

Measures
The constructs were assessed with a five-point Likert scale (1 = strongly disagree, 5 = strongly agree). The measurement instruments were original written in English, so a parallel back-translation procedure had to be followed [65]. The items used to measure the constructs are detailed in the Appendix A.
Since employee attitudes and behaviors are largely linked to how employees experience human resources practices [48,66], in line with other studies [2,7,12,20,45,50] we analyze employee's perceptions of empowerment and service training.
In particular, empowerment was assessed with three slightly modified items of the Hayes' [67] scale. An example item is "I do not have to get management's approval before I handle customer problems".
Service training was measured with three items adapted from Boshoff and Allen [15]. An example is "I receive training on how to serve customers better".
Affective commitment was measured with three items of the scale used by Rhoades et al. [68]. An example item is "I feel personally attached to my organization".
Finally, extra-role service was assessed with three items of Bettencourt and Brown [69]. An example is "I voluntarily assist customers even if it means going beyond job requirements".
Additionally, employee age, gender, educational level, and organizational tenure were included as control variables because they can affect affective commitment and extra-role service [9,22,57].

Analytic Method
The study aim, of an exploratory nature, is to improve the understanding of how and when empowerment affects affective commitment and the subsequent extra-role service of frontline employees.
Analysis of the data was carried out with SPSS Statistics 24 and SPSS AMOS 23. Firstly, a confirmatory factor analysis (CFA), using AMOS, was conducted to analyze the psychometric properties of the scales. Secondly, to contrast the hypotheses a mediation analysis and a moderated mediation analysis were carried out using the PROCESS macro for SPSS [70]. This method is especially appropriate when, in addition to analyzing the mechanisms through which one variable influences another variable, it is also important to know the conditions in which this relationship occurs. Moreover, the PROCESS macro for SPSS allows to use bootstrap, a procedure recommended to avoid problems arising from the non-normal distribution of data. Particularly, the use of bootstrap confidence intervals for hypothesis testing is especially recommended to avoid possible problems related to a small sample size [71]. To facilitate the interpretability of the coefficients, the variables involved in the interaction term were mean-centered [70]. Table 1 shows descriptive statistics and correlations. These are in line with the proposed hypotheses. In addition, the Cronbach's alphas values are showed in the diagonal.
Then, the scales' reliability and validity were examined. As can be seen in Table 2, the reliability of the scales was supported since the composite reliability (CR) and the average variance extracted (AVE) of every scale are higher than 0.6 and 0.5, respectively [72].
In addition, all the factor loadings (ranging from 0.737 to 0.951) were significant and higher than ideal value of 0.7 (see Appendix A), providing support for the convergent validity of the constructs. Moreover, the discriminant validity between constructs was also corroborated as the correlation confidence intervals (CI) do not include the unit, and their squared correlation were lower than the AVE [72].

Harman's One-Factor Test
Since all the data came from the same source, to check if the common method bias could influence the results, we used Harman's one-factor test [73]. As the one-factor model did not fit the data well (χ2 (54) = 610.221; χ2/df = 11.300; CFI = 0.416; TLI = 0.287; RMSEA = 0.318; SRMR = 0.201), it is to be expected that the common method bias is not a matter of relevance in this study.

Hyphoteses Testing
The presence of a mediating effect requires that two conditions to be fulfilled [74]: the independent variable (empowerment) must be significantly related to the mediating variable (affective commitment), and this must be significantly related to the dependent variable (extra-role service).
To test the first hypothesis, we employed Model 4 [70]. Table 3 presents the nonstandardized regression estimates for the mediation model. As seen, empowerment was positively associated with affective commitment (a 1 = 0.308; p = 0.001) which, in turn, was positively associated with extra-role service (b 1 = 0.452; p < 0.001). This gives initial support to the indirect effect of empowerment on the extra-role service. Moreover, the confidence interval for the indirect effect (95% CI = 0.049; 0.253) estimated by bootstrap exclude the zero value. The latter gives definitive support to the mediation hypothesis (Hypothesis 1).
Then to examine the moderation and the moderated mediation hypotheses we used Model 7 [70]. As show in Table 4, service training moderates the relationship empowerment-affective commitment (a 3 = 0.202; p = 0.018), confirming Hypothesis 2. Also, the confidence interval for the effect of the interaction term estimated by bootstrap (95% CI = 0.040; 0.353) provides additional support. In addition, Figure 2, plotted using the process by Dawson [75], shows that the effect of empowerment on affective commitment is not significant when employees perceive that they have received a low level of service training.  In addition, Figure 2, plotted using the process by Dawson [75], shows that the effect of empowerment on affective commitment is not significant when employees perceive that they have received a low level of service training. Finally, as shown in Table 5, the indirect influence of empowerment on extra-role service depends on the service training level. Moreover, the confidence interval for the index of moderated mediation do not include zero (95% CI = 0.018; 0.179), providing additional support for hypothesis 3 [76].
For moderate (the mean value) and high levels of training (one standard deviation above the mean), empowerment indirectly influences, through affective commitment, extra-role service. When the perceived level of service training is low (one standard deviation below the mean), the indirect relationship of empowerment on extra-role service is not significant (CI includes zero). It is therefore confirmed that for empowerment to indirectly influence extra-role service it is necessary for employees to perceive that they have received a moderated-high level of training. Finally, as shown in Table 5, the indirect influence of empowerment on extra-role service depends on the service training level. Moreover, the confidence interval for the index of moderated mediation do not include zero (95% CI = 0.018; 0.179), providing additional support for hypothesis 3 [76]. For moderate (the mean value) and high levels of training (one standard deviation above the mean), empowerment indirectly influences, through affective commitment, extra-role service. When the perceived level of service training is low (one standard deviation below the mean), the indirect relationship of empowerment on extra-role service is not significant (CI includes zero). It is therefore confirmed that for empowerment to indirectly influence extra-role service it is necessary for employees to perceive that they have received a moderated-high level of training.

Discussion and Implications
A large part of the success in the hospitality sector lies in frontline staff, as they are responsible for providing service to customers and therefore meeting their demands [6,7,28]. Since frontline employees know the customers' wishes and expectations, they are in the best position to decide on appropriate actions and responses [15]. That is why proper management of this group of employees becomes a strategic variable to create and sustain a competitive advantage [3,69].
Studies in the hospitality sector have point out empowerment as one of the main human resources practices to improve customer-contact employees' performance [7,11,20]. Frontline employees have to deal with unpredictable customers' needs, so they require individual discretion to give them a prompt response [59,77,78]. However, inconclusive results provided by some studies (e.g., [18,19]) suggest the need for new research to clarify how and when the empowerment-performance relationship occurs. Consequently, the main objective of this paper was to improve the understanding of the mediating and moderating mechanisms on the relationship between empowerment and extra-role service. Specifically, this study analyzed the mediating role of affective commitment and the moderating effect of service training on the empowerment-affective commitment relationship and, consequently, on the mediated relationship.
The results support the hypotheses. Empowerment influences extra-role service behavior through its effect on frontline employee's affective commitment. The mediated relationship found in this study is in line with the findings of previous studies which demonstrated that the effect of human resource practices on employee's behaviors was mediated by happiness well-being variables [2,7,12,20,48,50].
In their review of human resources practices in the hospitality sector, Kusluvan et al. [79] wrote: "Employee involvement and empowerment techniques can be used to conquer the hearts and minds of employees and help generate employee commitment, satisfaction, and performance" [79] (p. 198). According to results of the moderated mediation model, for this relationship to occur, frontline employees need have been trained to deal with customers' demands. It is evidenced that employees' empowerment affects extra-role service behavior via affective commitment, but only when they perceive to have received a moderate or high level of service training.
The findings are in consonance with the SDT [36] and the SET [37]. On the one hand, human resource practices that allow employees the opportunity to have the discretion to perform their jobs and provide them the required skills and knowledge fulfill the autonomy and competence basic needs that promote the internalization and integration, generating positive attitudes and improving the well-being of frontline employees [20,42,80]. On the other, empowering frontline employees, along with providing the necessary level of service training, signals employees that their contributions and well-being are considered worthwhile. This strengthens their affective bond with the organization which, in turn, increases their willingness to go beyond the requirements of their job to provide a high-quality service that addresses customers' demands. These volunteer behaviors have been linked to more customer satisfaction and organizational outcomes [3].
When frontline employees perceive that the firm does not care about improving their skills and knowledge (low-service training), empowerment does not generate a greater extra-role service through affective commitment. In this condition, frontline employees may believe that the organization is simply trying to increase their demands and responsibilities without providing them with the necessary resources. At low levels of perceived training, frontline employees may feel that the organization does not invest enough in them, generating insecurity, and weakening their sense of belonging and well-being.
Several studies have shown that giving employees more power of decision and discretion to perform tasks and providing them training lead to higher satisfaction and affective commitment [12,20,50]. However, the most common way of representing this synergistic effect has been through the aggregation of both practices into a composite index. Unlike these previous studies, we consider the interactive effect of both practices on the affective commitment. In fact, to the best of our knowledge, only two studies have analyzed the contingent role of training on the relation between empowerment and the attitudes or behaviors of employees. The study by Ellinger et al. [9] examined the interaction between empowerment and formal training on performance and organizational citizenship behaviors in a sample of frontline logistics service employees. Very recently, Manzoor et al. [35] corroborated, in a sample of 130 employees of two public universities in Pakistan, that training moderated the effect of empowerment on job performance. Our study deepens in this research line by supporting a model of moderate mediation in which empowerment and service training interact to affect the frontline employee well-being (affective commitment) and the resulting extra-role service in hospitality firms.
From a practical perspective, the results of this study underline that, in order to capitalize on the advantages of empowerment, hotel management must provide a moderate-high service training level for frontline employees. Empowerment is one of the core practices in the hospitality sector, especially for employees who are in contact with customers [15,20,78]. However, without an adequate training level in customer service, empowerment will not be able to sustain frontline employee's commitment and extra-role service. While frontline employees need to have the discretion to make decisions that allow them to respond immediately to customer requirements, they also need to possess high technical and interpersonal skills to feel able to make those decisions.
Regarding the study limitations, we should point out the followings. Firstly, the main limitations of the study are the sample size and that the data were taken in a specific context and geographic location (frontline employees from three-to-five-star hotels in Santiago de Compostela). Although the sample size is comparable to other studies in the field of human resource management (e.g., [7,35,81]) and it fulfills the requirements regarding the minimum sample size [72], it may be not possible to generalize the results to other hospitality firms. In order to corroborate ours results, additional research should test the model proposed in a larger sample, incorporating hotels located in other cities.
Secondly, as the study design is cross-sectional to support the causality of the relations future longitudinal studies are required. Thirdly, the use of self-reports could result in common method bias. However, we followed the recommendations for the design of the survey form by Podsakoff et al. [73] to try to minimize this bias. Additionally, the results of Harman's one-factor test suggest that common method bias was not a significant issue in the present study. Also, there is evidence that common method bias rarely affects interaction tests [82].
Despite these limitations, we believe that the results of this study, besides providing interesting insights for frontline employees' management in hotel contexts, can serve as insights for further development of research.

Conclusions
This investigation has revealed that when frontline employees are given authority and freedom to use their own judgment in solving the customer's problems but perceive that the organization does not care enough about their training, they will not feel part of the organization and will not engaged in discretionary behaviors. This is in line with the postulate by Boshoff and Allen, that "in order to provide a consistently high level of service, it takes more than just having the right people in the right job and empowering them. They must be trained to deal with customers" [15] (p. 68).
In summary, the sustainability of hospitality firms requires cultivating and sustaining employees' commitment and well-being. The joint implementation of human resource practices (empowerment and training), which reinforce each other, is beneficial for both employees and organizations. The employees' well-being will increase, and the subsequent extra-role service will result in more satisfied and loyal customers.