论文库 经济学论文 会计学论文 金融学论文 统计学论文 管理学论文 毕业论文 经管之家

Linking customer-employee exchange and employee innovative behavior来源：人大经济论坛论文库　作者：Minglong Li, Cathy Hsu　时间：2016-06-22

1. Introduction

Employee innovative behavior, as a foundation for organizational innovation, is widely accepted as a key factor for service firms’ performance and long-term survival (Campo et al., 2014; Tajeddini and Trueman, 2012). It brings about new products for restaurants (Ottenbacher and Harrington, 2007), improves hotels’ service processes (Orfila-Sintes and Mattsson, 2009), and enhances service quality and customer satisfaction (Pivcevic and Petric, 2011). Compared with manufacturing companies, service firms suffer from difficulty in applying for patents and identifying infringement of intellectual property rights (Hipp and Grupp, 2005). One of the solutions is employee innovative behavior, which may erect barriers to duplication by competitors and maintain competitive advantage over others for hospitality firms (Ottenbacher, 2007). Hotels are traditionally not innovation oriented (Pivcevic and Petric, 2011). However, hotels nowadays highlight innovation as countermeasures to growing competition (Campo et al., 2014). Most innovation of hotels is based on technology application and service-oriented (Su, 2011). Thus, unlike manufacturing firms, innovation in hotels relies more on employees than professionals in research and development department, especially customer-contact employees (Ottenbacher, 2007).

Employee innovative behavior by its nature, brings actual benefits to hospitality firms because it is required to result in final outputs (Kim and Lee, 2013). This final outputs requirement is one of the differences between innovative behavior and creativity. Creativity refers to development of novel ideas, while innovative behavior involves not only idea generation but also idea implementation (Kim and Lee, 2013). To implement new ideas, employees must seek support and resources from others. Thus, although employee innovative behavior is a type of individual innovation, it still requires certain behaviors and resources from others such as customers (Foss et al., 2011). As a result, frequent information and emotion exchanges run through the process of idea implementation (Scott and Bruce, 1994). From a social perspective, creativity is an individual-level construct bringing about novel ideas, thus weak social ties are generally beneficial for it as weak ties foster autonomy, allowing employees making decisions that may be different from the approaches and views of their contacts (Perry-Smith, 2006). Yet weak ties do not foster innovative behavior because of the importance of support from others in idea promotion and realization (Janssen, 2000). In other words, employees’ exchanges with others are important factors influencing their innovative behaviors.

Exchanges with others may be potential innovation facilitators for employees due to the social network opportunities. Research found that employee innovative behavior is influenced by not only individuals’ cognitive skills, but also social capital, which refers to the potential benefits employees receive from the relationships with others (Shalley and Gilson, 2004). Employees’ cognitive skills facilitate innovative behaviors by discovering the connections among various similar yet different concepts, flushing out the old approaches or ideas and creating different responses to a single problem (Perry-Smith, 2006). Social capital, as opportunities for employees’ learning from others, would make this innovative cognitive process more successful because of the diversity of ideas and information provided by their social relationships (Zhou et al., 2009). Thus, much research has focused on the impact of employees’ relational exchanges on their innovative behaviors, such as leader-member exchange (LMX) and co-worker exchange (Scott and Bruce, 1994; Sparrowe, 1995). For example, Volmer et al. (2012) found that as long as job autonomy is provided, the higher level of LMX would lead to more employee innovative behaviors. To date, LMX and its influence on employee innovation in hospitality firms have been well studied (Sparrowe, 1995; Volmer et al., 2012). However, little research has been conducted to investigate customer-employee exchange and its impact on employee innovative behavior in services. This study attempts to fill this gap.

Customer-employee exchange (CEX) may influence employee innovative behavior differently from other exchanges (e.g., LMX). Customer-employee exchanges are essential parts of the services due to the inseparability of service production and consumption (Ma and Qu, 2011). They could enrich the service experience of customers and subsequently enhance customer satisfaction (Namasivayam and Mattila, 2007). In this way, customer-employee exchanges are important to hospitality firms who find themselves increasingly difficult to meet the expectations of customers with escalating demand (Lee et al., 2006). However, although customers are increasingly actively involved in services and become collaborators with employees nowadays, their exchanges with employees are different from that of employees’ leaders and co-workers. Customer-employee exchanges in hospitality services are indispensable, but also characterized by temporal duration of interaction (Solnet, 2007). Unlike leaders and co-workers, the customers employees serve may be constantly changing. This makes customer-employee relationship relatively unstable. Employees may thus use different behaviors as responses to customers (Sierra and McQuitty, 2005). In addition, leaders tend to dominate the LMX and influence employees through management actions and role expectation, essentially for Chinese culture, where employees are more likely follow their leaders (Scott and Bruce, 1994; Shao and Skarlicki, 2014). In contrast, CEX is more based on emotions and both parties exert influence on the exchange, sharing some responsibilities (Solnet, 2007). For example, restaurant employees in China may take managers’ casual advice as orders, while they may share their personal experiences with customers because customers may not determine their final income (i.e., Chinese customers do not give tips). Therefore, using the results based on LMX to explain the effect of CEX may not be appropriate. Furthermore, much research indicates that customers are more and more external resources for hospitality firms and they exert beneficial influence on service innovation (Duverger, 2011; Sjödin and Kristensson, 2012). These studies tend to regard customers as innovators or contributors to innovation directly. However, whether customers’ exchanges with employees could be facilitators for employees’ innovative behaviors still remains unexamined.

To examine the effect of CEX on employee innovative behavior, the climate for innovation cannot be ignored. Climate describes employees’ perceptions of service settings where they work in terms of psychological interpretation (Schneider et al., 1996). It is found to mediate the relationship between LMX/team member exchange and employee innovative behavior, as high quality of LMX/team member exchanges makes employees perceive that they are in a positive and supportive climate, which may further encourage employees’ innovative behaviors (Schneider et al., 1996; Scott and Bruce, 1994). This climate involves support for innovation and resources supply (Scott and Bruce, 1994). Similarly, customer-employee exchanges involve interdependence between the two parties and may thus result in closer relationships between them (Kim and Cha, 2002). Better relationships make employees perceive that their decisions and behaviors are more likely to be supported by customers (Sigala, 2005). The support for innovation and resources provided by customers are important as customers are the final evaluators of some innovation outcomes (such as new services). However, climate for innovation created by customers receives little attention from researchers. Thus, this study adopted social psychological climate, which was originally used for firms, to investigate the role of support and resources from customers in the relationship between CEX and employee innovative behavior. Based on this, two objectives were set for this study: to examine the effect of CEX on employee innovative behavior and to investigate the role of social psychological climate (customer support and resource supply) in the influence of CEX on employee innovative behavior. 

2. Theory and hypothesis development

2.1 Customer-employee exchange and employee innovative behavior

Customer-employee exchange (CEX) relates to employees’ behaviors essentially. CEX involves both information and emotional interactions between the two parties (Ma and Qu, 2011). It defines how a service is transacted, which is very important for customers’ service perceptions as well as employees’ well-being (Groth and Grandey, 2012). Hospitality services are highly interactive and it is the main job of employees to serve customers (Victorino et al., 2005). The process involves frequent exchanges, which are found to influence employees’ job stress, job satisfaction and turnover intentions (Karatepe, 2009). Customers are external to hotels, thus their exchanges with employees may exert influence outside the service setting and also affect employees’ other behaviors such as organizational citizenship behavior (Groth and Grandey, 2012; Ma, 2013). Customers as external factors and employees’ turnover make customer-employee exchanges in hotels dynamic (Duverger, 2011) Yet hotels are increasingly paying attention to customer loyalty as retaining existing customers costs much less than finding new customers (Agarwal et al., 2003), thus customer-employee exchanges in hotel services are expected to be more stable (Agarwal et al., 2003). On the other hand, customers nowadays are no longer passive service recipients. They more actively participate in hotel services and are viewed as partial employees (Bendapudi and Leone, 2003). Therefore, although CEX is different from LMX and co-worker exchange, its influence on employees’ behavior should not be underestimated.

CEX may stimulate employees’ motivation and provide inspiration for innovative behaviors. Customer and employee frequent exchanges in hospitality service transactions may improve the relationships between the two parties (Ma and Qu, 2011). The emotional components of customer-employee exchange, such as the politeness of the two parties, not only play an important part in creating a successful experience, but also make each party have more reliance on the other (Lerman, 2006). According to the social exchange theory, employees who get benefits from customers have great intentions to return back (Lawler, 2001). Employee innovative behaviors, most of which are related to customers’ experience in hospitality firms, may thus be driven to improve service efficiency and effectiveness for customers (Chen, 2011). In addition, the interdependence between customers and employees caused by CEX may bring about information sharing and knowledge transfer, which are facilitators for employee innovative behaviors (Foss et al., 2011; Teece, 1994). Customer-employee exchanges may be favorable for employee innovative behavior because employees may capture customers’ internal demand, information and knowledge through frequent exchanges (Grissemann et al., 2013). These exchanges show in technical quality (what is being exchanged) and functional quality (the way it is being exchanged) (Gremler and Gwinner, 2000). If the technical quality of customer-employee exchanges is high, more information is exchanged between customers and employees. This information could be regarded as outward resources for employees, influencing their views and innovative behaviors (Sridhar and Srinivasan, 2012). Social influence due to exchanges is unavoidable and employees may combine their opinions and the views and information from others in their decision-making and behavior processes (Sridhar and Srinivasan, 2012). Similarly, high functional quality of customer-employee exchanges may build employee confidence, with which they are more likely to exhibit risk-taking behaviors (Ali and Ndubisi, 2011). Essentially, customers share some responsibilities in their exchanges with employees, and this tends to reduce the risk of employees’ proactive behaviors and encourage innovative behaviors (Lawler, 2001). Therefore, the following hypothesis is proposed.

Hypothesis 1. Customer-employee exchange is positively related to employee innovative behavior.

CEX involves three aspects including solidarity, harmonization and information exchange, according to an influential study by Keith et al. (2004). Solidarity of CEX refers to the extent to which an exchange is considered as important and ongoing (Keith et al., 2004). It indicates the degree to which customers and employees expect the relationships to continue in the long term and to remain beneficial for both parties (Keith et al., 2004). Harmonization describes the level of trust between the two parties and the ability to resolve conflict based on the relationship, while information exchange involves information contents in the exchanges (Castellanos-Verdugo et al., 2009; Keith et al., 2004). The three dimensions of CEX are also widely accepted, and cited with high frequency in academic publications. Additionally, this three-dimensional construct well reflect the interaction between customers and employees in hospitality services: high quality of exchanges are characterized by beneficial, sustainable, and involving much information and knowledge exchange (Chathoth et al., 2013). Thus, this study adopted the three factors of CEX to more specifically investigate hypothesis 1. Such sub-hypotheses testing is commonly used in hospitality innovation research (e.g., Nieves et al., 2014).

The three factors of CEX by Keith et al. (2004) may also have impact on employee innovative behavior. Customer-employee exchanges involve reciprocity, where both customers and employees benefit from each other and feel obliged to return good deeds to the other party (Blau, 1964). If the results of customer-employee exchanges make employees see their relationships with customers as important (i.e., solidarity), they tend to give customers their good deeds (Blau, 1964). Driven by the desire to return back the benefits to the customers, employees are more likely to exhibit innovative behaviors (Ottenbacher, 2007). In addition, employee innovative behaviors are risk-taking behaviors that may fail without support from others (Clegg et al., 2002). If employees perceive their relationships with customers as reliable and can be relied on (i.e., harmonization, with high level of trust), they may exhibit high level of creative self-efficacy and low level of risk, which may lead to more innovative behaviors (Clegg et al., 2002; Tierney and Farmer, 2011). Furthermore, information exchange between employees and customers is important external resources for employees (Keith et al., 2004). It may not only bring about creative new ideas in the “opportunity exploration” stage of innovation, but also lead to employees’ better understanding of customers’ needs and the services, which facilitates employees’ idea application (Foss et al., 2011; Kleysen and Street, 2001, p.285). Based on the analysis above, Hypothesis 1 can be further divided into the following:

Hypothesis 1a. Solidarity is positively related to employee innovative behavior.

Hypothesis 1b. Harmonization is positively related to employee innovative behavior.

Hypothesis 1c. Information exchange is positively related to employee innovative behavior.

2.2 Customer-employee exchange, social psychological climate, and employee innovative behavior

Researchers have traditionally investigated climate for innovation when examining the effect of contextual factors on employee innovative behavior. Climate refers to employees’ collective perceptions of their work (intangible) environment and service settings in terms of facilitating quality, rewards, support and encouragement of excellence (Andrews and Rogelberg, 2001). It is a type of employees’ perceived psychological condition reflecting their workplace (Ahmed, 1998). Scott and Bruce (1994) argued that the climate for innovation included two dimensions as support for innovation and resources supply from organizations; they also tested the mediating role of climate in the relationship between LMX/co-worker exchange and employee innovative behavior. The climate by Scott and Bruce (1994) is actually organizational psychological climate. Another type of climate for innovation is social climate (Ahmed, 1998). While organizational climate is inferred by members through organization practices, social climate reflects more about the emotional perceptions of employees’ other external social relationships (Ahmed, 1998). Thus, the support and resources employees need for innovation could be also provided by customers outside of service firms. These support and resources may relate to employees’ innovation objectives, participative safety (low risk of innovation), task orientation (concern for excellence) and support for innovation (Mathisen et al., 2004). The current study focuses on social psychological climate for innovation, describing employees’ collective perceptions of the service environment related to customers in terms of customers’ challenging requirement, support for risk-taking behaviors for service improvement and potential resource supply.

It is reasonable to propose that social climate for innovation may mediate CEX and employee innovative behavior, as organizational climate has been found mediating LMX/co-worker exchange and employee innovative behavior (Scott and Bruce, 1994). As a matter of fact, a probable result of high level CEX is the high quality customer-employee relationship, which is characterized by mutual obligation and emotional commitment (Kanagal, 2009). In a relationship with mutual trust, employees are more likely to obtain support or necessary resources from customers. That means high level of customer–employee exchange is likely to lead to high level of social psychological climate. Studies such as Foss et al. (2011) have found that employees perceive the environment as more supportive when customers are more willingly to interact with them. Thus, the following hypothesis is proposed.

Hypothesis 2. Customer-employee exchange is positively related to social psychological climate.

Same as Hypothesis 1, Hypothesis 2 is divided into three sub-hypotheses:

H2a. Solidarity is positively related to social psychological climate.

H2b. Harmonization is positively related to social psychological climate.

H2c. Information exchange is positively related to social psychological climate.

The importance of social psychological climate to employee innovative behavior is recognized by researchers. Climate is the result of customers’ expectations, from which employees may predict how customers would respond to their behaviors (Ottenbacher and Gnoth, 2005). Innovative behaviors are risk-taking behaviors, which may encounter other people’s resistance to the outcomes due to the challenges caused by innovation (Clegg et al., 2002). One source of resistance could be customers’ poorer performance on the same activities due to the new situations after innovation (e.g., self-ordering in restaurants after a new technology is adopted) (Ram and Sheth, 1989). If employees have sought support or resources from customers before innovation, that means customers already have some information about the innovation outcomes and may adjust to the new services better (Dorenbosch et al., 2005). Another source of resistance to innovation would come from conflicts between traditions and innovation (Ram and Sheth, 1989). Social psychological climate with high level of customer support could relieve the tension caused by these conflicts (Dorenbosch et al., 2005), thus may reduce resistance to innovation and encourage employees to innovate. Besides employees’ intention to innovate, they need to have the abilities to turn new ideas into reality. The resources provided by customers, such as time, demand information and advice, are significant to employee innovative behavior (Shalley and Gilson, 2004). Based on these, this study proposes the following hypothesis.

Hypothesis 3. Social psychological climate is positively related to employee innovative behavior.

Employee innovative behavior is a type of extra-role workplace behavior, by which employees voluntarily do more than required (Dorenbosch et al., 2005). Employees are not required to innovate unless they have intentions to do so. As a result, CEX may not inevitably lead to employee innovative behavior unless they feel that certain support and resources would be provided by customers (Geng et al., 2014). In other words, the strength of the relationship in Hypothesis 1 depends on the nature of social psychological climate. In services, the quality of customer–employee exchange is central to customers' evaluation and future behaviors, thus important to employees’ decision-making or behaviors (Gremler and Gwinner, 2000). CEX is bidirectional; if customers trust employees in their relationship development, employees may also trust customers (Keith et al., 2004).  Thus, if employees regard the relationships with customers as important (solidarity) and see customers as trustworthy (harmonization), they are more likely to turn customers’ input and resources into innovative behaviors (Ulwick, 2002). In this way, high level of CEX and social psychological climate together may lead to more employee innovative behaviors. Based on this, and Hypotheses 2 and 3, the following hypothesis is proposed.

Hypothesis 4. Social psychological climate mediates the relationship between customer-employee exchange and employee innovative behavior.

This hypothesis is further divided into three sub-hypotheses (see Figure 1).

Hypothesis 4a. Social psychological climate mediates the relationship between solidarity and employee innovative behavior.

Hypothesis 4b. Social psychological climate mediates the relationship between harmonization and employee innovative behavior.

Hypothesis 4c. Social psychological climate mediates the relationship between information exchange and employee innovative behavior.

(Insert Figure 1 Here)

3. Methods

3.1 Questionnaire development

The authors treat employee innovative behavior as the endpoint of the study; thus it is important to have a baseline understanding of the extent of innovativeness of the hotels where target employees worked. If employee innovative behavior is not regarded as important in a hotel, surveying employees in the hotel is meaningless. Thus, three screening questions were designed to enquire about the importance of innovation, support for new ideas and treatment of risk-taking behaviors in the hotel (Shalley and Gilson, 2004). Respondents were asked to rate the degree of their agreement on these three statements, from strongly agree (7) to strongly disagree (1). If the mean scores of the three questions based on participants in a hotel are all higher than 5 (=“Slightly agree”), innovative behaviors are regarded as important in the hotel. Otherwise, the questionnaires collected from the hotel would be removed.

Employee innovative behavior was measured by the scale developed by Janssen (2000), which is widely accepted and has been confirmed in a hotel context (Janssen, 2005; Slåtten and Mehmetoglu, 2011). The scale involves 3 factors as idea generation, idea promotion and idea realization; each factor includes 3 items. Respondents were asked to assess how often they perform innovative behaviors from “7=always” to “1=never”. In previous studies, employee innovative behaviors were rated by supervisors or by employees; in both circumstances, the scale was found reliable in some studies (Janssen, 2005). Thus, the measurement scale was assessed by employees in this study, because surveying the manager-employee pairs would require employees to reveal their identity, which may influence their responses, although ideally surveying both supervisors and employees could help alleviate potential common method bias.

Although the importance of customer-employee exchange (CEX) in services is widely accepted, not many studies have empirically tested the measurement of this construct. Among them, Keith et al. (2004) developed a 15-item scale based on previous research. This measurement scale involves three dimensions (solidarity, harmonization and information exchange), with each dimension including 5 items. This scale was confirmed in their study with high reliability and validity (α=0.8495, AVE=0.5857) (Keith et al., 2004). Meanwhile, this scale was accepted by many other researchers (Dampérat and Jolibert, 2009) and adopted in the hospitality context (Chathoth et al., 2013). In hotels, customers need to build at least a temporary relationship with employees for service transactions to be completed (Ma et al., 2013), although the relationship may end as soon as they leave the hotel (for one time guests). Even so, the scale provided by Keith et al. (2004) could be adopted in this study. As many customers may have only temporal exchanges with employees, it is more appropriate to survey employees who have more accumulative customer employee interactions than customers. Thus, some minor adjustments were made for the CEX measurement based on employees’ perspectives. Actually, Keith et al. (2004) mentioned the necessity of revising the measurement scale for other studies in the final section of their paper. After the item adjustments, a 7-point Likert scale was used, with 7 representing “strongly agree” and 1 “strongly disagree”.

The social psychological climate for innovation was measured using Scott and Bruce’s (1994, p.593) scale, originally for measuring organization’s (including staff members) support and resources, with some adjustments, as social psychological climate is a parallel concept to organizational psychological climate (Ahmed, 1998). Employees may form general impression of customers through constant exchanges with different customers, although customers may be changing (Chathoth et al., 2013). In addition, the participative customers as “partial employees” share some characteristics with organization members (Bendapudi and Leone, 2003). Thus, this scale adoption is reasonable; only some items need to be revised so that they can be used in the customer context. Items such as “this organization can be described as flexible and continually adapting to change” and “personnel shortages inhibit innovation in this organization” were removed, with 14 items remaining. Others are adjusted if necessary for this study. For example, the original item “creativity is encouraged here” is replaced by “innovation is encouraged by my customers”. These items were also measured with a 7-point Likert scale where 7=strongly agree and 1 = strongly disagree.

All the measurements adopted from previous studies were designed in English. As the target respondents are Chinese, the questionnaire was translated into Chinese using the back-translation technique to make sure that the Chinese version of the questionnaire is comparable to the English version in meaning. Also, measures were taken to reduce the common methods bias, such as making sure that respondents are anonymous, separating the different constructs on three separate pages and not indicating any relationships among the constructs in the survey instruction.

3.2 Pilot study

A pilot study was carried out in Shenzhen, China, to test the content validity and reliability of the survey instrument as well as to evaluate the readability and translation adequacy. A convenience sample of customer contact employees in three hotels (Kempinski, InterContinental and Ritz-Carlton) was recruited and a total of 70 questionnaires was collected. Some items in the scales of CEX and social psychological climate were negatively worded, such as “The information provided to me by customers is often inadequate”. Before the data analysis, scores of these items were reversed (e.g., the original 2 was replaced by 6). After that, Exploratory Factor Analysis (EFA) was conducted for the three constructs to investigate the underlying factors. The reliability statistics and correlation analysis were also conducted to refine the measurements. The variables with factor loadings lower than .6 were to be removed, as the scales were previously developed with high reliability and validity (Hair et al., 2009). The Cronbach’s α of a construct should be higher than 0.7 and the mean inter-item correlations be more than 0.4 (Cortina, 1993). An item was to be removed if its corrected item-total correlation (CITC) was lower than 0.3 and deleting the item can increase the reliability of the measurement (Cortina, 1993). Actually, the α values of all constructs were higher than 0.7 and CITC of all items were higher than 0.3.

EFA on CEX showed that three factors had eigenvalues greater than 1 and the total variance explained was higher than 63%. The three factors were consistent with the original scale developed by Keith et al. (2004). Two items (“Customers and I want to cultivate a good working relationship” and “I expect my relationship with customers to last a long time”) that loaded on Factor 1 (“solidarity”) with factor loadings lower than .6 (0.527 and 0.574, respectively) were removed. The removal was supported by Solnet (2007) that working relationships and long-lasting relationships are not common in customer-employee exchanges, characterized by the temporary interaction for service transaction and constant adjustments to different customers. Also removed was another variable (“Customers keep me informed to help me plan for their needs”) loaded on Factor 3 (“information exchange”) due to the low factor loading (.569). This item actually measures the exchange before the services, which may not reflect the meaning of CEX in the present study. After the three items were removed, EFA was conducted again. The results showed that the number of factors and the items loaded on them did not change. In addition, after these three items being removed, the Cronbach’s α of solidarity and information exchange became higher (solidarity: from 0.82 to 0.84; information exchange: from 0.84 to 0.85). Thus, it is reasonable to remove these items and the refined scale was used in the main survey.

The EFA on employee innovative behavior (KMO=0.952, Bartlett’s test p<0.001) did not support the three–factor structure (idea generation, idea promotion and idea realization) by Janssen (2000) and suggested only one factor, with eigenvalue of 7.42, and 82.44% of the variance explained. Thus, this study treated employee innovative behavior as a unidimensional construct. As for social psychological climate, the two–factor structure (“support for innovation” and “resource supply”) was also not supported. The data showed good fit for the factor analysis (KMO=0.980, Bartlett’s test p<0.001), yet the results of EFA suggested only one factor. Thus, customer related climate for innovation was treated as a unidimensional construct. Meanwhile, two items (“Customers give me positive responses to encourage innovation” and “Customers enjoy some flexibility and continually adapt to change”) were removed because of the low factor loadings (0.532 and 0.511, respectively) and removing them increased the reliability of the construct.

3.3 Sample and procedure

The main survey was carried out in Shenzhen, China, from February to March in 2015. The hotels in Shenzhen are regarded as serving customers well and also innovative, and employees of these hotels come from diverse areas of China (China Tourist Hotel Association, 2014), thus surveying hotel employees in Shenzhen may provide implications for other areas of China. The researchers contacted the managers of Front Office and/or Food and Beverage (F&B) departments in ten four- and five- star hotels, which are regarded as more innovative than low-star hotels (China Tourist Hotel Association, 2014). Managers from seven hotels agreed to support this study and arranged for the data collection. Convenience sampling was used to select employees who were available for the survey, with a plan to recruit half of the respondents from Front Office and half from F&B departments. Questionnaires were distributed to employees face-to-face. The researchers are unacquainted with all the respondents. During the survey process, the researchers were not present while respondents completed the questionnaires (we distributed the questionnaires, left enough time for employees to fill out the questionnaires, and collected the questionnaires  at a later time) to avoid social desirability issues. Respondents were asked to return the questionnaires to the designated service counter in their hotel. Altogether, 200 questionnaires were distributed. The sample size was determined following the rule-of-thumb of 5, where the sample size is based on 5 times the number of variables (33*5=165) (Westland, 2010).Although 189 questionnaires were collected, 9 questionnaires with too many missing values were discarded. Thus, 180 questionnaires were retained, involving 76 front line employees from Front Office and 104 customer-contact employees in F&B departments.

A total of 42 missing values was found among the 33 variables of the construct measurement, accounting for less than 1% of all values. Thus, missing values were replaced with the mean values of each variable (Hair et al., 2009). Tests of data normality were conducted for all the variables and results showed that all variables were approximately normally distributed. Descriptive statistics were analyzed for all survey items with IBM SPSS Statistics 20.0. AMOS version 20.0 was employed to conduct Confirmatory Factor Analysis (CFA) for the measurements and Structural Equation Modelling (SEM) for the hypotheses testing (Hair et al., 2009).

4. Results and discussion

Among the 180 respondents, 56.1% were females and 38.3% were males, with 5% unknown (see Table 1). Over forty percent (43.9%) of the respondents were aged 18-25, followed by the age of 26-35 (37.8%). Less than 3% of the respondents were over the age of 45, and the other 8 were missing values. Most of the respondents had a bachelor’s degree or higher (59.4%) and 31.7% completed education in secondary/high schools. The monthly income of respondents mainly ranged from ￥2,000 to 2,999 (32.2%) and from ￥3,000 to 3,999 (36.1%).

(Insert Table 1 Here)

Most respondents agreed that their hotel regarded employee innovation as important (Mean=5.82, see Table 2). Meanwhile, managers in the hotel rewarded those who brought new ideas to work (M=5.57) and showed understanding of failure in their risk-taking behaviors (M=5.26). Therefore, the importance of employee innovative behaviors is well accepted in the participating hotels. Means and standard deviations of all construct measurement items are also reported in Table 2.

(Insert Table 2 Here)

4.1 Reliability, validity, and measurement model

Confirmatory Factor Analysis (CFA) was carried out to assess discriminant validity of the constructs. As shown in Table 3, the Cronbach’s alpha values showing the composite reliability of the multi-item scales all exceed 0.7, the recommended cut-off point (Tavakol and Dennick, 2011), and an acceptable level of reliability for each construct. In addition, all average variance extracted (AVE) of the constructs are higher than 0.5, suggesting high convergent validity (Fornell and Larcker, 1981). Discriminant validity is further confirmed by the fact that AVE of each construct is higher than its squared correlation coefficients for inter-constructs (Fornell and Larcker, 1981). For example, the highest squared correlation coefficient of employee innovative behavior is .648 (=.805², coefficients are listed in Table 4), lower than .67 (Table 3). Also, CFA results showed that the factor loadings for indicators are significant, with p< .01. Therefore, the reliability and validity of the constructs are acceptable. Meanwhile, the R squares for the three models are .532, .730 and .743, respectively, indicating significant relationships (Hair et al., 2009).

(Insert Table 3 Here)

(Insert Table 4 Here)

Meanwhile, the goodness-of-fit indices from the CFA model were obtained. The values of χ2 (= 872.8) and df (= 485) in the model indicated significance at 0.01 level of probability. Both NNFI (= .94) andCFI (= .95) were higher than .9, the suggested cut-off point by Kline (2011). RMSEA (=.067) lay between .05 and .08, suggesting acceptable fit (Kline, 2011). All these indices show that the CFA model fit the data well. Therefore, the measurement model was confirmed. As the overall measurement model turned out to be reliable and valid, the structural models were tested and the main results are shown in Figure 2, which will be further analyzed in the next section.

(Insert Figure 2 Here)

4.2 Hypotheses testing

First-order SEM models were used to test the hypotheses (Kline, 2011). In other words, three factors of customer-employee exchange (CEX) were analyzed separately. To test the mediating effect of social psychological climate, three models were tested (Baron and Kenny, 1986) with the regressions from predictor to outcome (Model 1), from predictor to mediator and from mediator to outcome (Model 2), and from both predictor and mediator to outcome (Model 3). The model fit indices of the three models are listed in Table 5 (RMSEA= .067< .08, NNFI= .94 > .9, CFI= .95> .9). They indicate that Model 2 and Model 3 have good model fit, while Model 1 fits the data relatively poorer. The RMSEA for Model 1 (=.083) is slightly higher than .08, but far away from .1, suggesting mediocre fit (Kline, 2011). NNFI (= .898) is slightly lower than .9. Considering the relatively small sample size of this study (n= 180< 200), the authors regard Model 1 as acceptable (Hair et al., 2009).

Model 1 indicates that both solidarity and harmonization of CEX positively influence employee innovative behavior, while there is no significant relationship between information exchange and employee innovative behavior (see Table 5). Thus, hypothesis 1 is partially supported, with H1a and H1b being supported but H1c not. As information exchange does not significantly affect employee innovative behavior, hypothesis 4c (the mediating role of social psychological climate between information exchange and employee innovative behavior) is not supported.

Model 2 shows that the paths to social psychological climate from both solidarity (β = .42, t = 5.28) and harmonization (β = .43, t = 5.77) are significant, but from information exchange (β = .11, t = 1.70< 1.96) is not. Thus hypotheses H2a and H2b are supported, while H2c is not. Meanwhile, social psychological climate positively and significantly affect employee innovative behavior (β = .87, t = 12.01). Therefore, Hypothesis 3 is supported.

As indicated by Models 1 and 2 (see Table 5), solidarity of CEX is significantly associated with employee innovative behavior (β = .36, t = 4.01) and social psychological climate (β = .42, t = 5.28). However, in Model 3, when both CEX and social psychological climate are regarded as predictors of employee innovative behavior, solidarity no longer significantly affect employee innovative behavior (β = .06, t = .78). As both the paths from solidarity to social psychological climate  and from social psychological climate to employee innovative behavior are significant, this suggested that social psychological climate may be a perfect mediator of employee innovative behavior in this case. Sobel test results confirmed that social psychological climate mediated the relationship between solidarity and employee innovative behavior (z = 4.61, p<.01). Therefore, Hypothesis 4a is supported.

(Insert Table 5 Here)

Results also show that the harmonization dimension of CEX significantly influences employee innovative behavior, in both Model 1 (β = .46, t = 5.25) and Model 3 (β = .16, t = 2.35). Nevertheless, this effect is weaker in Model 3 (when social psychological climate is included in the regression analysis) than in Model 1 (both β and t decrease). In addition to the findings that both the effect of harmonization on social psychological climate and the impact of social psychological climate on employee innovative behavior are significant and positive, it can be concluded that social psychological climate mediates the relationship between harmonization and employee innovative behavior (Sobel test z = 4.35, p < .01). As a result, hypothesis 4b is also supported.

5. Conclusion and implications

This study attempts to link customer-employee exchange (CEX) and employee innovative behavior in a hotel context, bringing a multidisciplinary contribution to service marketing and innovation research. The results indicate that high level of CEX (solidarity and harmonization) may lead to more employee innovative behaviors, although information exchange does not significantly affect employee innovation. Also, social psychological climate for innovation is found as an important facilitator for employee innovative behavior. In addition, climate mediates the relationship between solidarity/harmonization of CEX and employee innovative behavior. The findings could provide some implications theoretically and practically.

5.1 Theoretical implications

This study adopted the measurement scales from previous studies and refined them in a hotel context. The scale of employee innovative behavior developed by Janssen (2000) was found valid and reliable to assess employee innovative behaviors in the hospitality industry. Nevertheless, although employee innovative behavior is viewed as a multiple-stage process consisting of idea generation, idea promotion and idea realization (Janssen, 2000; Kleysen and Street, 2001), the data of hotel employees in Shenzhen only supported a unidimensional construct. A possible reason could be that the boundaries among stages of employee innovation in hotels are blurred (Martĺnez-Ros and Orfila-Sintes, 2009), so that one underlying construct is sufficient to explain the data.

Another key construct is customer-employee exchange. From the service point of view, CEX is an important part of service production and delivery, and these exchanges involve more than just transaction of services (Grandey et al., 2012). The exchanges between customers and employees may not only affect customer related outcomes such as service quality and experience (Gremler et al., 2001), but also facilitate employee innovative behavior. Three aspects of CEX in hotels are confirmed in this study. High level of customer-employee exchanges in hotel services is characterized by both parties’ viewing the relationship as important and ongoing, by involving mutual commitment and resolving the conflicts based on their relationships rather than a third party, and the exchange of various forms of information.

This study contributes to the understanding of employee innovative behavior formation and the role of customers’ influence. Previous research examined employees' personality, leadership and organization climate and their influence on employee innovative behavior (Shalley and Gilson, 2004); however, much fewer studies focused on customer-related factors. This study confirmed the link between CEX and employee innovative behavior. Customers may not only bring innovation to hotels (Tajeddini and Trueman, 2012), but also influence employees’ innovative behaviors via their exchanges with employees, as found in this study. Of course, the three factors of CEX have different effect on employee innovative behavior. Harmonization of CEX positively influence employee innovative behavior; solidarity affect employee innovative behavior via social psychological climate; but information exchange does not exert any influence. This conclusion is different from leader-member exchange (LMX), with all dimensions having significant and direct effect on employee innovation (Sparrowe, 1995). The possible reasons may lie in the differences between CEX and LMX. For example, LMX may involve management actions and constant feedback, while CEX is highly focusing on service experience (Ma and Qu, 2011; Scott and Bruce, 1994). Of course, the exact reasons need further investigation, such as examining the effect of LMX and CEX on employee innovative behaviour at the same time in a hotel context.

Information exchange, one component of CEX, was found not to be significantly associated with employee innovative behavior, which means that CEX facilitates employee innovative behavior through emotional engagement rather than through technical information sharing. The information or even ideas provided by customers, as argued by Ulwick (2002), may be nothing new and just based on their experiences in other firms or their personal needs. Therefore, the importance of customers to employee innovative behavior lies in developing deeper emotional bonds and goodwill to enhance employee motivation to innovate and creative self-efficacy instead of simple information sharing during their exchanges with employees. Even with supportive psychological climate, the effect of information exchange on employee innovative behavior is still not significant (see Model 3 in Table 5). One possible reason could be that the measurement items of information exchange reflect more about the amount of the information (e.g., “frequently”, “inadequate”) than the quality, as the effect turn negative in Model 3 (Table 5). The qualityf information needs to be considered in future research.

Finally, the mediating effect of social psychological climate has been partially supported in this study, which enhanced the understanding of the influence mechanism of customer-related factors on employee innovation. Solidarity and harmonization of CEX were found positively related to psychological climate. Meanwhile, the positive relationship between social psychological climate and employee innovative behavior has been confirmed. The mediating role of social psychological climate indicates that customer support and resource supply for innovation are also important for hospitality firms. This expands the meaning of climate for innovation to include factors outside of the organizations.

5.2 Managerial implications

Relational exchanges between customers and employees should be valued and encouraged in service firms. Among the three aspects of CEX, the harmonization of exchange is a key factor in fostering employee innovative behavior. In other words, if employees and customers have trustworthy relationships and they could resolve any conflict or problem on their own (Keith et al., 2004), employees are likely to exhibit more innovative behaviors. Thus, firms should emphasize the value of the exchanges between customers and employees, create positive environment for the development of mutual trusting relationships between customers and employees, and give employees more autonomy to deal with service and customer related issues, which lead to high level of relational exchange (Keith et al., 2004). Harmonization of CEX also affect employee innovative behavior partially via psychological climate. In addition, results of the study showed that high level of solidarity of CEX could result in more employee innovation indirectly, through the mediation of psychological climate. Thus, high level of solidarity may represent high relationship quality, but it does not inevitably lead to employee innovative behavior, unless social psychological climate exists. In practical terms, to facilitate employee innovative behavior, hotels could nourish psychological climate, an even more direct approach (than CEX), by involving more customers in some participative programs (e.g., tea ceremony or cooking classes).

Creating social psychological climate is utmost important for employee innovative behavior in hotels, as social psychological climate has a stronger effect on employee innovative behavior than CEX. Therefore, hotel managers need to provide some incentives to customers for employees to obtain customers’ support and resources for innovation. For example, in hotel restaurants, customers can be offered newly developed menu items on a complementary basis to solicit their views and provide opportunities for employees to interact with customers. Hotels could also develop customer engagement activities, such as games and competitions relating to creativity, to cultivate customers’ openness to new ideas. Encouraging customers to more actively participate in service production and delivery is another way to gain some resources from customers (Victorino et al., 2005), although some of these resources may not be directly related to innovation. Customers could also be invited to nominate employees who have engaged in innovative behaviors so that customers are aware of the innovative culture of the organization. These activities encourage more exchanges between customers and employees as well as gain support from customers for innovation. These deeper relationships and understanding could eventually derive innovative behaviors of employees. Some of the suggestions to managers could be challenging; however, those who can conquer these challenges will earn their competitive advantages and outperform others in the crowded marketplace.

5.3 Limitations and future research directions

There exist several limitations in this study, which also pave ways for future research. First, the sampling in four- and five-star hotels and only in Shenzhen may weaken the generalizability of the research findings. The results based on employees in Shenzhen may not be applicable to other less developed areas of China. Also, other types of hotels such as economy hotels, which develop quickly in China, allow customers to perform some services themselves and are regarded as models for hotel innovation, could be the focus in future research. Another limitation on sampling is the small sample size (180). Some results may be different if more questionnaires were collected. In future research, a larger sample size could be adopted (e.g., based on rule of 10), to purify the measurement scales further in a hospitality context.

Secondly, this study examined the customer-employee exchange only from the perspective of employees. Meanwhile, social psychological climate was also assessed by employees only and there may be gaps between employees’ perceptions and the actual behaviors of customers. Future research could involve customers as respondents and measure the impact of CEX on employee innovative behavior using experimental design, where customers and employees are matched. The data collection could also be extended over a longer period of time as innovative behaviors are not daily activities.

Information exchange is found not significantly associated with employee innovative behavior. This seems to contradict with the observations that information and knowledge sharing in hospitality firms lead to employee innovative behavior (Hu et al., 2009). One possible reason could be that the quality of the information exchanged is neglected in this study. Future research could specifically investigate customer-employee information exchange in terms of amount and contents.

Social psychological climate could vary largely among diverse characteristics of customers, such as different personality types and frequencies of visit (first time customers vs. repeat customers), which were not considered in the current study. Future research could test the influence of personality or characteristics of customers on the relationship between CEX and employee innovative behavior.

Service transactions involve more than just the exchanges between customers and employees. Future research could also investigate other customer-related constructs, such as customers’ participation in service production and delivery, and customers’ feedback and their possible impact on employees. Co-innovation of customers and employees in service settings may also be a direction for future research.