Introduction
More supervisors are willing to establish extra-role relationships with their subordinates through off-the-job activities, such as after-work happy hours (Finkelstein, Protolipac, & Kulas, Reference Finkelstein, Protolipac and Kulas2000), in that they concur that this form of supervisor–subordinate relationship promotes employees’ positive behaviors (Farth, Tsui, Xin, & Cheng, Reference Farth, Tsui, Xin and Cheng1998; Chen, Friedman, Yu, Fang, & Lu, Reference Chen, Friedman, Yu, Fang and Lu2009). Guanxi, which refers to the quality of social interaction outside work, developing through social activities including visits, dinning, gift-giving, and reciprocating favors, is used to describe this form of relationship (Chen & Tjosvold, Reference Chen and Tjosvold2006). Although guanxi has been deemed as a critical feature of the Chinese management practice (Han & Altman, Reference Han and Altman2009), it is essential to understand the social interaction, leadership, and management in both Chinese and American societies (Hui & Graen, Reference Hui and Graen1997). Chen and Tjosvold (Reference Chen and Tjosvold2006) also argued that supervisor–subordinate guanxi was important within and across cultural boundaries. Supervisors interact with their subordinates on a daily basis by enacting formal and informal roles, leading their guanxi with subordinates to be an integral part of any work reality, and thus guanxi is regarded as a salient contributor to effective management of human resources (Han & Altman, Reference Han and Altman2009). Nevertheless, only a limited number of studies explored this issue in the workplace (Wong, Wong, & Wong, Reference Wong, Wong and Wong2010). For research and managerial understanding supervisor–subordinate guanxi deserves further exploration.
Two theories, including the attribution theory and psychological contract theory, underline the benefits of guanxi. The attribution theory, which is the major theoretical approach to social perception (Farmer & Aguinis, Reference Farmer and Aguinis1999), asserts that individuals tend to develop causal expectations for events experienced by themselves and others (Campbell & Swift, Reference Campbell and Swift2006). Attribution theory implies that individuals incline to form positive evaluations about others with whom they have good relationship. For example, when a supervisor has good guanxi with a subordinate, the supervisor may participate in the subordinate's private life (Zhai, Lindorff, & Cooper, Reference Zhai, Lindorff and Cooper2013). In this case, this supervisor shows more understanding to the subordinate and thus tends to forgive the subordinate's mistakes in works (Chen & Tjosvold, Reference Chen and Tjosvold2007; Cheung, Wu, Chan, & Wong, Reference Cheung, Wu, Chan and Wong2009).
The psychological contract describes an implicitly reciprocal exchange agreement between a supervisor and his/her subordinate (Lester, Turnley, Bloodgood, & Bolino, Reference Lester, Turnley, Bloodgood and Bolino2002). This implicit psychological contract may be created based on guanxi. For example, a supervisor may often treat his/her subordinates to dinner or helps them resolve problems in their life. These off-the-job favors help develop mutual psychological contract. The supervisor expects the subordinate returns his/her favors whereas the subordinate feels obligations to reciprocate the favor by means of working hard.
Based on these theories, guanxi which helps the supervisor and subordinate to form a friendship can benefit both parties (Bedford, Reference Bedford2011). Cheung et al. (Reference Cheung, Wu, Chan and Wong2009) concluded that both supervisors and subordinates could take advantage of the close personal friendship. Mao, Hsieh, and Chen (Reference Mao, Hsieh and Chen2012) found that workplace friendship would enhance perceived job significance and such enhancement did not vary across organizational levels. Based on guanxi, subordinates can seek work advices from their supervisors for facilitating job executions or better job-related returns, and supervisors can use informal channels to timely resolve non-routine work problems of their subordinates. Accordingly, the good guanxi aids job-related outcomes. However, the stream of research empirically explore guanxi's benefits often either focuses on the subordinate's view (e.g., Chen & Tjosvold, Reference Chen and Tjosvold2006; Ramasamy, Goh, & Yeung, Reference Ramasamy, Goh and Yeung2006) or leads to inconclusive findings (Cheung et al., Reference Cheung, Wu, Chan and Wong2009). There is a need to examine beneficial consequences of guanxi.
For extending the understanding of benefits of supervisor–subordinate guanxi, this study focuses on the effect of guanxi on job satisfaction rated by both the supervisor and subordinate. Job satisfaction reflects a positive emotional state resulting from an individual's appraisal of the job's characteristics and job experiences (Cheung & Wu, Reference Cheung and Wu2012). Cheung et al. (Reference Cheung, Wu, Chan and Wong2009) found that job satisfaction could mediate the beneficial influences of supervisor–subordinate guanxi on the subordinate's positive outcomes. Thus, job satisfaction is an appropriate job-related outcome variable due to its focal position in the organizational behavior literature, being considered as a proxy for an employee's well-being at work, and being determinant to various advantages (Heller, Judge, & Watson, Reference Heller, Judge and Watson2002; Carmeli, Reference Carmeli2003).
In addition to demonstrate the consequential effect of guanxi, it is necessary to explain how supervisors and subordinates can build up or improve guanxi with each other (Law, Wong, Wang, & Wan, Reference Law, Wong, Wang and Wan2000). Recently, the study on antecedents of guanxi quality receives increasing attention (Chen & Peng, Reference Chen and Peng2008; Bedford, Reference Bedford2011). Among others, personal antecedents deserve more exploration because the supervisor–subordinate relationship can signal fit between the individual's personality and the organization (Stevens & Ash, Reference Stevens and Ash2001). Although Zhai, Lindorff, and Cooper (Reference Zhai, Lindorff and Cooper2013) supported the notion regarding to a dispositional source of supervisor–subordinate guanxi, they also recognized that seldom research investigated potential dispositional antecedents of guanxi. Thus, personality traits are subsumed into the research theme for exploring their influences on the supervisor–subordinate relationship.
Among personality traits, core self-evaluation (CSE) has been found to positively associate with various job-related outcomes (Hiller & Hambrick, Reference Hiller and Hambrick2005). The benefit of CSE from a research perspective lies in its nature of self-evaluation, pivot of personality, and broadness in scope (Judge & Bono, Reference Judge and Bono2001). CSE refers to individuals’ appraisals of their own worthiness, effectiveness, and capability (Judge, Erez, Bono, & Thoresen, Reference Judge, Erez, Bono and Thoresen2003). It consists of four narrow traits including self-esteem, generalized self-efficacy, locus of control, and emotional stability (Judge, Van Vianen, & De Pater, Reference Judge, Van Vianen and De Pater2004). Individuals who rate CSE high see themselves as worthy, capable, in control of their lives, and freedom from anxiety (Hiller & Hambrick, Reference Hiller and Hambrick2005). In addition to the importance role of CSE playing in the job-related context, two reasons explain why this study chooses this personality as an antecedent to the supervisor–subordinate guanxi.
First, CSE influences an individual's opinions about people and events related to oneself (Piccolo, Judge, Takahashi, Watanabe, & Locke, Reference Piccolo, Judge, Takahashi, Watanabe and Locke2005), and thus it determines an individual's perceptions about supervisors, subordinates or jobs. Second, Beehr, Glaser, Beehr, Beehr, Wallwey, and Erofeev (Reference Beehr, Glaser, Beehr, Beehr, Wallwey and Erofeev2006) stated that two sets of potential causes of satisfaction with subordinates: who we perceive them to be and what they do for us. CSE covers these two sets in that it is dominant in predicting perceptions about people and events related to oneself (Piccolo et al., Reference Piccolo, Judge, Takahashi, Watanabe and Locke2005) and about capability (Hiller & Hambrick, Reference Hiller and Hambrick2005). Specifically, CSE may influence one's ability to build a relationship and other's perception about oneself. However, there is a lack of research on the relationship between CSE and the supervisor–subordinate guanxi.
The objective of this study is to explore influences of CSE on the supervisor–subordinate guanxi and job satisfaction. However, the development of guanxi involving two parties does not rely on a single party's CSE without regarding to the other's CSE. Imagine that a subordinate prepares various social activities for his/her supervisor such as treating the supervisor a dinner, giving birthday gift, helping the supervisor to deal with private affairs. But the supervisor may refuse the subordinate's offer of favors in that the supervisor is not willing to have frequent interactions with this subordinate outside the office due to incompatible personalities. Supervisors seek right subordinates who fit their preferences. Thus there is a need to consider the fit of personality. Edwards (Reference Edwards1994) suggested that the (in)congruence effect was considered a predictor of job-related outcomes. Interpersonal interaction theory proposes that (dis)similarity between individuals governs their behaviors and attitudes toward each other (Glomb & Welsh, Reference Glomb and Welsh2005). The (dis)similarity of personality traits between individuals relates to the development of supervisor–subordinate guanxi. A harmonious supervisor–subordinate relationship may develop in this manner contingent on a match between the parties’ traits. However, very few studies have assessed whether the (dis)similarity in personality, especially CSE, leads to the interpersonal relationship (Klohnen & Luo, Reference Klohnen and Luo2003). Thus, another objective of this study is to clarify whether the essence of fit reflects guanxi and job satisfaction derived from the (dis)similarity of CSE between a supervisor and his/her subordinate. In sum, this study attempts to link the theoretical gap by empirically exploring the personal antecedent and the beneficial effect of supervisor–subordinate guanxi, and by offering a fresh idea to the fit theory. For clarifying the objectives of this study, Figure 1 lays out the research framework.
Figure 1 The research framework
LITERATURE REVIEW
Supervisor–subordinate guanxi
Supervisor–subordinate guanxi refers to the strength of the off-the-job interaction or informal connection between a supervisor and a subordinate developed from off-work-related social activities inside and outside working hours such as gift presenting, family visits, and resolving personal problems (Chen & Tjosvold, Reference Chen and Tjosvold2006; Cheung et al., Reference Cheung, Wu, Chan and Wong2009; Wei, Liu, Chen, & Wu, Reference Wei, Liu, Chen and Wu2010; Liu, Hui, Lee, & Chen, Reference Liu, Hui, Lee and Chen2013). The most frequently mentioned components of supervisor–subordinate guanxi include affective ties, instrumental ties, personal-life inclusion, trust, and respect (Chen & Peng, Reference Chen and Peng2008; Chen et al., Reference Chen, Friedman, Yu, Fang and Lu2009; Han & Altman, Reference Han and Altman2009; Lin & Ho, Reference Lin and Ho2010; Bedford, Reference Bedford2011; Zhai, Lindorff, & Cooper, Reference Zhai, Lindorff and Cooper2013)Footnote 1.
While affective ties means emotional connection, understanding, and willingness to care for one another (Chen et al., Reference Chen, Friedman, Yu, Fang and Lu2009; Zhai, Lindorff, & Cooper, Reference Zhai, Lindorff and Cooper2013), instrumental ties describe the reciprocal exchange of favors, implying an informal social obligation (Chen & Tjosvold, Reference Chen and Tjosvold2007; Wong, Wong, & Wong, Reference Wong, Wong and Wong2010; Cui, Wen, Xu, & Qin, Reference Cui, Wen, Xu and Qin2013). Personal-life inclusion reflects the extent to which the supervisor and subordinate involve in each other's private or family lives (Chen et al., Reference Chen, Friedman, Yu, Fang and Lu2009). Example of the personal-life inclusions include after office hours activities (e.g., home visits), other social functions (e.g., dining together), or participating in other's personal and family events (Zhai, Lindorff, & Cooper, Reference Zhai, Lindorff and Cooper2013). Trust is defined as recognition of the other party's integrity and credibility (Bedford, Reference Bedford2011) whereas respect refers to the subordinate's degree of obedience and devotion toward the supervisor (Chen et al., Reference Chen, Friedman, Yu, Fang and Lu2009).
For achieving some certain means, individuals require the exchange of favors and personal-life inclusions to develop or improve a guanxi including affect, trust, and respect (Chen & Tjosvold, Reference Chen and Tjosvold2007). During a continuous socializing process, the connection between the supervisor and subordinate gradually turns into a friendship (Bedford, Reference Bedford2011). Consequently, the supervisor and subordinate with guanxi are granted a high level of affect, favors, personal interaction, trust, and respect. Based on the high-quality guanxi, individuals are regarded as in-groups (Hom & Xiao, Reference Hom and Xiao2011).
Note that the concept of supervisor–subordinate guanxi is similar to the concept of leader-member exchange (LMX) in some respects but they are quite different (Law et al., Reference Law, Wong, Wang and Wan2000; Cheung et al., Reference Cheung, Wu, Chan and Wong2009; Wei et al., Reference Wei, Liu, Chen and Wu2010). Both guanxi and LMX are grounded in social exchange theory, highlight the importance of the quality of relationship between a supervisor and subordinate, and allow individuals to assign others as in-groups and out-groups (Chen et al., Reference Chen, Friedman, Yu, Fang and Lu2009; Han, Peng, & Zhu, Reference Han, Peng and Zhu2012; Jiang, Chen, & Shi, Reference Jiang, Chen and Shi2013). But they are different in terms of the relationship's foundation and reciprocity rule (Chen et al., Reference Chen, Friedman, Yu, Fang and Lu2009). While LMX is restricted to work-related exchanges, guanxi comes from social interactions both inside and outside working hours (Chen et al., Reference Chen, Friedman, Yu, Fang and Lu2009; Cheung et al., Reference Cheung, Wu, Chan and Wong2009; Wei et al., Reference Wei, Liu, Chen and Wu2010). In addition, the reciprocity rule of LMX focuses on equal exchange of working-related benefits, but people with guanxi tend to reciprocate favors that are social and economic in nature without returns of equal value (Law et al., Reference Law, Wong, Wang and Wan2000; Chen et al., Reference Chen, Friedman, Yu, Fang and Lu2009; Hom & Xiao, Reference Hom and Xiao2011). For example, a supervisor may often treat his/her specific subordinate a dinner for initiating a development of guanxi, and he/she would not like to equally exchange his/her treat for the subordinate's next treat. The subordinate may reciprocate the kind treatment by preparing a surprising birthday party for the supervisor. The spirit of guanxi lies in the continuous and harmonious friendship.
CSE, supervisor–subordinate guanxi, and job satisfaction
Four traits embody the concept of CSE including the self-esteem, self-efficacy, locus of control and emotional stability. Understanding of these traits helps capture the concept of CSE. While self-esteem is the valuation and perception of self-worth, self-acceptance, self-liking, and self-respect, generalized self-efficacy describes an individual's own estimate of ability to activate the motivation, resources, and courses of action needed to successfully execute tasks (Judge, Locke, Durham, & Kluger, Reference Judge, Locke, Durham and Kluger1998). Locus of control pertains to both the extent to which individuals believe that they control the occurrence of events and the attribution for (un)successful outcomes (Avery, Reference Avery2003). People who have internal locus of control (i.e., internals) ascribe outcomes to themselves, whereas people who have external locus of control (i.e., externals) ascribe outcomes to luck or the environment. Emotional stability captures an individual's tendency to present good emotional adjustments and to focus on positive affects of selves rather than neuroticism (Judge et al., Reference Judge, Erez, Bono and Thoresen2003). In sum, people with a high CSE display self-liking, being confident of capability, internal locus of control, and positive affects of selves.
This study proposes that people who have high CSE tend to build good guanxi with others due to their willingness and ability. In terms of the willingness to build guanxi, individuals with high CSE view themselves as worthy and appealing to others (Bernerth, Armenakis, Feild, Giles, & Walker, Reference Bernerth, Armenakis, Feild, Giles and Walker2007) and thus they are more willing to approach people. Anderson (Reference Anderson2009) also suggested that individuals who desired to create close bonds with others were less likely to experience negative relationship. Individuals with CSE are more willing to engage in off-the-job activities with others and to develop the guanxi. In terms of the ability to build guanxi, Varma, Stroh, and Schmitt (Reference Varma, Stroh and Schmitt2001) suggested that a subordinate's ability was beneficial to develop a supervisor–subordinate relationship. Subordinates with high CSE are confident in their abilities and so incline to develop supervisor–subordinate guanxi. Supervisors often form good impressions of capable subordinates and thus develop good relationships with those subordinates, because these subordinates with high CSE usually display behaviors of focused attention, persistence in pursuing goals, and overcoming obstacles (Huang & Weng, Reference Huang and Weng2012).
Although few studies identify the relationship between CSE and supervisor–subordinate guanxi, some studies confirming the benefit of one of CSE traits document indirect evidence of this hypothetic relationship. For example, Ng, Sorensen, and Eby (Reference Ng, Sorensen and Eby2006) posited that internals were more possible to develop and maintain positive relationships with their supervisors. Internals conceive that they can control events and so they feel comfortable in social experiences. In addition, individuals with emotional stability are effective in socializing (Bernerth et al., Reference Bernerth, Armenakis, Feild, Giles and Walker2007) or are open to interacting with others (Porter, Wrench, & Hoskinson, Reference Porter, Wrench and Hoskinson2007) in that they do not perceive others as detrimental to themselves (Judge et al., Reference Judge, Locke, Durham and Kluger1998). Based on the foregoing logics, this study develops the following hypotheses:
Hypothesis 1a : Supervisor's CSE positively associates with supervisor–subordinate guanxi rated by supervisor.
Hypothesis 1b : Subordinate's CSE positively associates with supervisor–subordinate guanxi rated by subordinate.
Although the link between CSE and job satisfaction receives lot supports from previous research (Chang, Ferris, Johnson, Rosen, & Tan, Reference Chang, Ferris, Johnson, Rosen and Tan2012), its mediation still seems to be vague. Prior studies attempted to identify the mediating mechanisms behind the relationship between CSE and job satisfaction (Srivastava & Adams, Reference Srivastava and Adams2013). Based on Baron and Kenny (Reference Baron and Kenny1986), it is necessary to examine the direct relationship between the predictive variable and outcome variable before investigating the mediating effect. Therefore, this study includes the discussion about the relationship between CSE and job satisfaction and tests this relationship for providing a comprehensive conception. In addition, seldom studies examined this relationship as considering perspectives of both supervisors and subordinates simultaneously.
Empirical evidences reveal that both respective traits (e.g., Ng, Sorensen, & Eby, Reference Ng, Sorensen and Eby2006; Porter, Wrench, & Hoskinson, Reference Porter, Wrench and Hoskinson2007) and overall CSE positively relate to job satisfaction (e.g., Judge et al., Reference Judge, Locke, Durham and Kluger1998, Reference Judge, Erez, Bono and Thoresen2003). Judge, Bono, Erez, and Locke (Reference Judge, Bono, Erez and Locke2005) suggested that an individual with a high CSE tended to develop personal happiness by choosing self-concordant work goals. Also, an individual with a high CSE excels in translating advantages into subsequent achievements and so sees jobs in a positive manner (Judge & Hurst, Reference Judge and Hurst2007). Moreover, based on self-verification motivation, an individual with a high CSE tends to maintain a positive self-concept by focusing on positive features of the work environment (Wu & Griffin, Reference Wu and Griffin2012). The positive influence of CSE on job satisfaction holds across cultures (Piccolo et al., Reference Piccolo, Judge, Takahashi, Watanabe and Locke2005).
Hypothesis 1c : Supervisor's CSE positively associates with job satisfaction.
Hypothesis 1d : Subordinate's CSE positively associates with job satisfaction.
Consequences of supervisor–subordinate guanxi
Social exchange theory argues that both supervisors and subordinates can benefit from high-quality supervisor–subordinate relationships (Chen, Friedman, Yu, & Sun, Reference Chen, Friedman, Yu and Sun2011; Farr-Wharton & Brunetto, Reference Farr-Wharton and Brunetto2012). Taken from the standpoint of the supervisor, guanxi may lead to increased job satisfaction of the supervisor because the supervisor is satisfied with the subordinate's positive behaviors and the leadership effectiveness. First of all, guanxi is positively associated with the subordinate's organizational commitment (Farth, Tsui, Xin, & Cheng, Reference Farth, Tsui, Xin and Cheng1998), organizational citizenship behavior (Wong, Ngo, & Wong, Reference Wong, Ngo and Wong2003; Lin & Ho, Reference Lin and Ho2010; Liu & Wang, Reference Liu and Wang2013), and reduced turnover intentions (Farth, Tsui, Xin, & Cheng, Reference Farth, Tsui, Xin and Cheng1998; Wong, Ngo, & Wong, Reference Wong, Ngo and Wong2003). The subordinate's positive behaviors can ease the supervisor's working loadings. Second, subordinates tend to trust in their supervisors based on high-quality guanxi (Farth, Tsui, Xin, & Cheng, Reference Farth, Tsui, Xin and Cheng1998; Wong, Ngo, & Wong, Reference Wong, Ngo and Wong2003; Wong, Wong, & Wong, Reference Wong, Wong and Wong2010; Han, Peng, & Zhu, Reference Han, Peng and Zhu2012). In this case, supervisors can use their personal power to influence these subordinates’ behaviors (Wei et al., Reference Wei, Liu, Chen and Wu2010). Also supervisors have beliefs in subordinates’ reciprocity for the favorable treatment and assistance (Liu & Wang, Reference Liu and Wang2013). Consequently, supervisors may produce high job satisfaction due to the benefits from guanxi.
In terms of the standpoint of the subordinate, two theories, including the social identity theory and inducement and contribution theory, can justify why the subordinate can produce job satisfaction resulting from his/her guanxi with the supervisor. Subordinates in good guanxi with supervisors obtain greater benefits than their colleagues (Law et al., Reference Law, Wong, Wang and Wan2000; Zhai, Lindorff, & Cooper, Reference Zhai, Lindorff and Cooper2013). Based on social identity theory, supervisor–subordinate guanxi helps the subordinate to develop a group identity and perceived in-group status (Han, Peng, & Zhu, Reference Han, Peng and Zhu2012; Wang & Kim, Reference Wang and Kim2013), and thus the subordinate internalizes into the evaluation that he/she is worthy of the organization's investment (Liu et al., Reference Liu, Hui, Lee and Chen2013). The subordinate may find it easier to be satisfied with the working conditions or environment due to this perception of social identity.
Based on inducement and contribution theory, those subordinates who receive more inducements from their supervisors, such as valuable resources, premium assignments, and promotion opportunities, tend to contribute more to their works (Cheung et al., Reference Cheung, Wu, Chan and Wong2009; Wei et al., Reference Wei, Liu, Chen and Wu2010). Supervisors may be impressed by these contributions. Thus, the supervisors often give positive evaluations of task performance to their subordinates in good guanxi (Wang & Kim, Reference Wang and Kim2013). In addition, these subordinates often can produce high job performance due to good-quality guanxi. Specifically, a guanxi between the supervisor and subordinate facilitates knowledge sharing (Wang, Tseng, & Yen, Reference Wang, Tseng and Yen2012). The subordinate can seek advice from the supervisor and thus meet job requirements easily (Cheung et al., Reference Cheung, Wu, Chan and Wong2009). Evaluations of task performance from both the supervisor and subordinate make the subordinate feel satisfied with the job.
Guanxi relates positively to the subordinate's fairness perception, which further enhances his/her job satisfaction (Chen et al., Reference Chen, Friedman, Yu and Sun2011). Some studies recognized the positive effect of guanxi on the subordinate's job satisfaction rather than the supervisor's (Cheung et al., Reference Cheung, Wu, Chan and Wong2009; Zhai, Lindorff, & Cooper, Reference Zhai, Lindorff and Cooper2013). This study considers standpoints of both supervisors and subordinates and proposes the following hypotheses.
Hypothesis 2a : Supervisor–subordinate guanxi rated by supervisors positively associates with their job satisfaction.
Hypothesis 2b : Supervisor–subordinate guanxi rated by subordinates positively associates with their job satisfaction.
Supervisor–subordinate guanxi as a mediator
The previous hypotheses, when combined, set up a logical question related to supervisor–subordinate guanxi potentially mediating the relationships between CSE and job satisfaction. While some research has previously shown that personality traits lead to positive job-related outcomes (e.g., Piccolo et al., Reference Piccolo, Judge, Takahashi, Watanabe and Locke2005), this study further attempts to document that one of mechanisms through which it does so relates to the ability of certain personality traits to develop good supervisor–subordinate guanxi. Seldom studies regarded the supervisor–subordinate relationship as a useful mediation in the relationship between CSE and job satisfaction, except for Srivastava and Adams (Reference Srivastava and Adams2013) who used the supervisor support rated by subordinates as the mediator. Nevertheless, the authors only considered the single view of subordinates and found a partial mediating effect of supervisor support. Thus, there is a need to examine the mediating effect of guanxi in the relationship between CSE and job satisfaction.
In terms of the mechanism of guanxi, social exchange theory and LMX theory act as guiding frameworks. Social exchange theory concludes that personal traits influence the quality of exchange relationship, and both parties experience certain benefits from the better quality of relationship (Harris, Harris, & Eplion, Reference Harris, Harris and Eplion2007). LMX theory asserts that supervisors incline to categorize some subordinates with whom they have good relationships as in-group memberships (Suazo, Turnley, & Mai-Dalton, Reference Suazo, Turnley and Mai-Dalton2008), and the good relationships result in reduced psychological distance for in-group dyads (Campbell & Swift, Reference Campbell and Swift2006). Grodzicki and Varma (Reference Grodzicki and Varma2011) found that supervisors would give higher performance ratings to in-group subordinates.
More specifically, an individual with a high CSE is capable of developing off-the-job guanxi, and further capable of obtaining job-related outcomes from the in-group membership. Some research documented the mediating role of guanxi in the relationships between the traits and job satisfaction. For example, Cheung et al. (Reference Cheung, Wu, Chan and Wong2009) suggested that a subordinate who had a good guanxi with a supervisor may be believed to be more important than others, and this prestigious feeling can result in a positive evaluation about jobs. Harris, Harris, and Eplion (Reference Harris, Harris and Eplion2007) argued that subordinates high in traits related to CSE were more likely to develop high-quality relationships with their supervisors, which in turn result in better job-related outcomes. Based on the foregoing logic, this study provides the supervisor–subordinate guanxi as an explanatory mechanism through which CSE is associated with job satisfaction.
Hypothesis 3 : Supervisor–subordinate guanxi mediates the relationship between CSE and job satisfaction.
Supervisor–subordinate fit
Fit theory divides into two perspectives on this issue. Similarity perspective advocates that similarity of relational demography between supervisors and subordinates can enhance subordinates’ satisfactions (Green, Whitten, & Medlin, Reference Green, Whitten and Medlin2005). Conversely, the dissimilarity perspective argues that dissimilarity of characteristics is advantageous to the supervisor–subordinate dyad (Glomb & Welsh, Reference Glomb and Welsh2005). Nevertheless, both perspectives rarely focus on CSE in the supervisor–subordinate dyad. Although Bernerth et al. (Reference Bernerth, Armenakis, Feild, Giles and Walker2007), who explored the respective influences of the big five traits on a subordinate's perception of LMX quality, did not particularly aim at fit theory, their findings could offer reference for this study, in that CSE consistently relates to extraversion and conscientiousness (Judge et al., Reference Judge, Erez, Bono and Thoresen2003). Bernerth et al. found that both the supervisor and subordinate's conscientiousness positively related to the subordinate's perceptions of the LMX quality, but not extraversion. Their findings suggest that neither of the perspectives is overwhelmingly conclusive. Accordingly, I adopt two different perspectives to elaborate on this issue.
Similarity-attraction paradigm suggests that individuals who share some characteristics in common will be attracted to each other (Varma, Stroh, & Schmitt, Reference Varma, Stroh and Schmitt2001) due to the reduced psychological distance between individuals (Campbell & Swift, Reference Campbell and Swift2006). The similar conclusion is reached with LMX theory and moral exclusion theory. LMX theory posits that LMX status signifies the reduced psychological distance for in-group dyads and greater psychological distance for out-group dyads (Graen & Uhl-Bien, Reference Graen and Uhl-Bien1995; Campbell & Swift, Reference Campbell and Swift2006). The moral exclusion theory implies that individuals tend to categorize others as similar or dissimilar and show favoritism toward similar ones and derogation toward dissimilar ones (Tepper, Moss, & Duffy, Reference Tepper, Moss and Duffy2011).
Several studies appeared to support the notion of similarity perspective. Sears and Holmvall (Reference Sears and Holmvall2010) found that a similarity of emotional intelligence between a supervisor and a subordinate was associated with their relationship. Bernerth et al. (2008) confirmed that similarity in emotional stability, intellectual openness, agreeableness, and conscientiousness positively related to perceptions of LMX. Likewise, Beehr et al. (Reference Beehr, Glaser, Beehr, Beehr, Wallwey and Erofeev2006) documented that the similarity between supervisors and subordinates would relate to supervisors’ satisfactions with subordinates. Additionally, Bernerth et al. (Reference Bernerth, Armenakis, Feild, Giles and Walker2007) mentioned that supervisors favored subordinates who were evaluated as capable, while Porter, Wrench, and Hoskinson (Reference Porter, Wrench and Hoskinson2007) found that subordinates favored supervisors who were evaluated as capable. Although Bernerth et al. (Reference Bernerth, Armenakis, Feild, Giles and Walker2007) and Porter, Wrench, and Hoskinson (Reference Porter, Wrench and Hoskinson2007) did not investigate the fit theory, these studies appear to support similarity perspective. Specifically, these studies imply that both of parties prefer capable ones. Individuals with high CSE are often considered capable (Judge et al., Reference Judge, Van Vianen and De Pater2004). Based on Bernerth et al. (Reference Bernerth, Armenakis, Feild, Giles and Walker2007) and Porter, Wrench, and Hoskinson (Reference Porter, Wrench and Hoskinson2007), it is concluded that both parties tend to favor each other when they are evaluated as capable or when both of them have high CSE.
Hypothesis 4 : Similarity of CSE between supervisor and subordinate positively associates with supervisor–subordinate guanxi and job satisfaction.
There is evidence that might conflict with the notion of personality similarity (Blass & Schwarcz, Reference Blass and Schwarcz1982). For example, Brown, Poole, and Rodgers (Reference Brown, Poole and Rodgers2004) found that a complementary dyad in which one party was dominant and the other was submissive tends to exhibit a high level of trust. Likewise, Glomb and Welsh (Reference Glomb and Welsh2005) partially supported that a dissimilarity of control traits ranging from dominance to submission in a supervisor–subordinate dyad enhanced the subordinate's satisfaction with the supervisor. Kausel and Slaughter (Reference Kausel and Slaughter2011) suggested that organizations adopting a recruiting strategy based on a similarity in personality might not succeed in attracting the most preferred employees. However, prior research studied on linking personality dissimilarity to work-related outcomes produced mixed and contradictory results (Guillaume, Brodbeck, & Riketta, Reference Guillaume, Brodbeck and Riketta2012). Particularly, few studies focused their attentions on the dissimilarity between a supervisor and a subordinate in the fit theory. It is necessary to examine this issue to link the literature gap.
Complementary perspective may offer the justification for the positive effect of dissimilarity in CSE. Complementary hypothesis posits that individuals should be attracted to those who have personality traits that are complementary to their own (Klohnen & Luo, Reference Klohnen and Luo2003). This is because individual's dissimilarity in personality may complement those of other colleagues and thus produce some benefits. For example, individuals whose personality traits are dissimilar to other peers can contribute their unique attributes that are necessary for the group to be successful (Liao, Joshi, & Chuang, Reference Liao, Joshi and Chuang2004). Personality dissimilarity allows individuals to learn from others and to generate new ideas by combining different perspectives (Molleman, Reference Molleman2005), and thus may aid in personal development, improve problem solving and information sharing (Glomb & Welsh, Reference Glomb and Welsh2005). In addition, dissimilarity in personality may decrease the likelihood of comparison, resulting in reduced contrast process (Ruys, Spears, Gordijn, & de Vries, Reference Ruys, Spears, Gordijn and de Vries2006). People are likely to initiate the development of a relationship when they do not engage in the comparison and contrast process.
Some studies provide the indirect evidence to support the positive effect of dissimilarity of CSE though they are rare. Bregman and McAllister (Reference Bregman and McAllister1982) revealed that individuals with low self-esteem liked others with high and low self-esteem rather than neutral self-esteem. In addition, Liao, Joshi, and Chuang (Reference Liao, Joshi and Chuang2004) supported that extraversion dissimilarity positively predicted coworker satisfaction. Since self-esteem is one of CSE traits and relates to extraversion, which characterizes individuals as sociable and outgoing (Judge et al., Reference Judge, Erez, Bono and Thoresen2003), this study expects that dissimilarity in CSE between a supervisor and subordinate positively associates with guanxi and job satisfaction based on Bregman and McAllister (Reference Bregman and McAllister1982) and Liao, Joshi, and Chuang (Reference Liao, Joshi and Chuang2004).
Hypothesis 5 : Dissimilarity of CSE between supervisor and subordinate positively associates with supervisor–subordinate guanxi and job satisfaction.
Methodology
Sampling plan
Smith and Canger (Reference Smith and Canger2004) argued that it was appropriate to study the effects of each supervisor's traits on the outcomes of ten subordinates, rather than individual supervisor–subordinate dyads. Accordingly, I asked one supervisor and five of their subordinates in one company to fill out the questionnaires. I confined the scope to the banking industry in Taiwan to eliminate the influence of industry types. In the banking industry employees and their immediate supervisors have frequent interactions. Besides, Taiwan is a place that fuses Chinese and Western cultures. Those in organizations adopt traditional Chinese cultural values while adopting Western management practices. The choice of banks in Taiwan therefore represents a valid source of respondents.
I randomly picked out 16 sample banks in 36 domestic banks from the list of financial institutions offered by the Central Bank of Taiwan. I then obtained contact information on each bank's branches from its official website, and contacted with senior managers in the bank branches by means of a personal letter inviting their participations and assuring confidentiality. After obtaining branches’ approval to participate in the survey, I sent confirmation letters to contacts of sample branches in order to ask for lists of members in the firm. In this process, some branches that refused to offer lists dropped out of the survey. Based on the list of members offered by each branch, five subordinates for a supervisor in each branch are randomly designated as potential respondents by this study. I further contacted with each of potential respondents for obtaining personal approval of participating and mail addresses. Then I sent a survey instrument, including a cover letter addressed personally to the participant, a questionnaire, and a postage-paid return envelope to each participant.
I mailed a total of 900 questionnaires to 150 bank branches. A supervisor and his/her immediate five subordinates in each bank branch were asked to fill out questionnaires. Supervisors needed to respond self-administered items of CSE and job satisfaction. Also, they received a questionnaire listing their subordinates’ names that participated in the survey and were asked to rate their guanxi with these subordinates. Specifically, supervisors were invited to use the same guanxi items to score each subordinates listed on the questionnaire. In addition, subordinates needed to respond items of CSE, guanxi with their supervisors, and job satisfaction based on their own perceptions. In order to pair subordinates with their supervisors, I will pre-code each questionnaire with an identification number before sending out to subordinate-level participants. After completing questionnaires, respondents can seal the document for fear of leaking private evaluation. Four weeks after initial mailing, I directly contacted participants that still did not return questionnaires to me by telephone and visited them in person to bring them replacement questionnaires. Eventually, I received 526 completed questionnaires, a response rate of 58%, that split 89 from supervisors and 437 from their subordinates, from 89 bank branches belonging to 11 banks. On average, the sample included eight branches for each bank among 11 banks (see Table 1 for detailed respondent information).
Table 1 Sample description
Notes. aThe sum of each category does not equal to sample size because of missing values.
bThe number of supervisors equals to the number of bank branches.
Construct measurement
Measures of the research constructs draw on the extant literature. CSE comprises four narrow traits of self-esteem, generalized self-efficacy, locus of control, and emotional stability (Judge et al., Reference Judge, Van Vianen and De Pater2004). Judge et al. (Reference Judge, Erez, Bono and Thoresen2003) developed a direct measure of CSE. The instrument uses three items to measure each trait and it has proven valid in cross-cultural contexts (Judge et al., Reference Judge, Van Vianen and De Pater2004). Examples for the self-esteem include ‘I am confident I get the success I deserve in life’ and ‘Sometimes when I fail I feel worthless (reverse item).’ Examples for the generalized self-efficacy include ‘I complete tasks successfully’ and ‘I am capable of coping with most of my problems.’ Examples for the locus of control include ‘I do not feel in control of my success in my career (reverse item)’ and ‘I determine what will happen in my life.’ Examples for the emotional stability include ‘Sometimes I feel depressed (reverse item)’ and ‘There are times when things look pretty bleak and hopeless to me (reverse item).’
Guanxi refers to the off-the-job activities in a dyadic relationship between a supervisor and a subordinate. Wong, Ngo, and Wong (Reference Wong, Ngo and Wong2003) scale focused on measurement of off-the job activities in the development of guanxi and so was considered appropriate for this study. Following Wong, Ngo, and Wong (Reference Wong, Ngo and Wong2003) scale, subordinates completed eight items to reflect their perceptions about guanxi with their supervisors. Sample items include: ‘I have frequent interactions with my immediate supervisor after work,’ ‘I have a high degree of mutual interest with my immediate supervisor after work.’ For assessing supervisors’ perceptions about guanxi, I created a parallel version of scale. Minor changes (e.g., changing supervisor to subordinate) were made to make items appropriate for supervisor responses. Sample items include: ‘I have frequent interactions with my immediate subordinate after work,’ ‘I have a high degree of mutual interest with my immediate subordinate after work.’
Job satisfaction is a self-evaluation of job-related outcomes. Based on Price's (Reference Price1997) comments on several measures of job satisfaction, this study employed Weiss, Davis, England, and Lofquist (Reference Weiss, Davis, England and Lofquist1967) short form of the Minnesota satisfaction questionnaire. This measure has been validated in diverse contexts and is a composite measure of different facets (cf. Price, Reference Price1997: 480). Examples of 20 items in total to capture the concept of job satisfaction include the following: ‘I am satisfied with the competence of my supervisor (my subordinate) in making decisions’ and ‘I am able to do things that don't go against my conscience.’
All items were measured on a 7-point Likert scale. I designed initial version of questionnaire and then submitted them for a panel discussion with three professors in the organizational behavior domain to discuss the appropriateness of items. Various amendments were made through pretesting with 92 part-time graduate students.
A two-stage analytic process, containing exploratory factor analysis and second-order confirmatory factor analysis (CFA), was applied to confirm the dimensionality, validity, and reliability of the constructs. In addition, the second-order CFA helps confirm whether the research construct as one-factor measure is robust. Table 2Footnote 2 shows that the factor loadings and Cronbach's αs for all constructs cross the required thresholds and thus dimensionality and reliability of all constructs are acceptable (Hair, Black, Babin, & Anderson, Reference Hair, Black, Babin and Anderson2010).
Table 2 Factor analysis for mixed sample
Notes. aItems which failed to meet the selection criteria of factor analysis were deleted.
bA second-order CFA for supervisors produced a chi-square of 530.31 (df = 243, p < .001) and satisfactory fitness indices (GFI = 0.88, CFI = 0.89, RMSEA = 0.07). A second-order CFA for subordinates produced a χ2 of 1090.51 (df = 287, p < .001) and satisfactory fitness indices (GFI = 0.89, CFI = 0.90, RMSEA = 0.07).
***p < .001.
Second-order CFA models produced satisfactory fitness indices (GFI = 0.88, CFI = 0.89, RMSEA = 0.07 for supervisors; GFI = 0.89, CFI = 0.90, RMSEA = 0.07 for subordinates). Table 2 displays standardized loading estimates exceeded 0.50 and the variance explained and reliability value for each factor are >0.50 and 0.70. These results indicate an adequate convergent validity and reliability for each factor. Discriminant validity is adequate given each of the variance-extracted estimates from Table 2 is greater than the corresponding interfactor squared correlation estimates (i.e., values above the diagonal) in Table 3 (Hair et al., Reference Hair, Black, Babin and Anderson2010).
Table 3 Means, standard deviations, correlations among research variables
Note. Values below the diagonal are correlation estimates. Values above the diagonal are squared correlations.
Results
Hypotheses tests
A series of structural equation modeling (SEM) was conducted to examine the hypothetic relationships: (a) a unconstrained model with all samples (the unconstrained model allows parameters of potential relationships in SEM model to be estimated free), (b) a constrained model with all samples for further testing the mediations, (c) unconstrained and constrained models with multi-groups (including supervisors and subordinates).
Table 4 displays the standardized estimates of paths and Table 5 summarizes model fitness indices for the SEM models. Hypothesis 1 through Hypothesis 2 concern the influences of CSE on guanxi and job satisfaction and the influence of guanxi on job satisfaction from supervisors’ and subordinates’ views, respectively. The results in relations to Hypothesis 1 through Hypothesis 2 are shown in the first column in the block of unconstrained model (gray zone). As shown in Table 5, GFI and AGFI exceed 0.90, and the value of RMSEA is reasonably small (Hair et al., Reference Hair, Black, Babin and Anderson2010), indicating an adequate fit of the data to the hypothesized models, including the full model and competing model with multi-groups.
Table 4 Results of SEM
Notes. aConstrained models are specified with zero direct effects of CSE on job satisfaction.
–denotes zero direct effects.
**p < .010; ***p < .001.
Table 5 SEM model fitness comparisons
Table 4 shows that CSE positively relates to guanxi (supervisors: β = 0.52, p < .001; subordinates: β = 0.29, p < .001), and the influence of CSE on guanxi between supervisors and subordinates is indifferent (t = −0.72, p > .05). Hypothesis 1a and Hypothesis 1b are supported. CSE positively relates to job satisfaction for supervisors (β = 0.45, p < .01) and subordinates (β = 0.22, p < .001), which is consistent with Hypothesis 1c and Hypothesis 1d. The coefficient for supervisors does not differ significantly from that for subordinates (t = 0.45, p > .05). This shows that CSE is antecedent of job satisfaction. Guanxi positively relates to job satisfaction (supervisors: β = 0.55, p < .01; subordinates: β = 0.45, p < .001), and its coefficients are indifferent between supervisors and subordinates (t = 0.35, p > .05). Hypothesis 2a and Hypothesis 2b are supported.
Hypothesis 3 concerns the mediating effect of guanxi on the relationship between CSE and job satisfaction. For examining the mediations, this study fixes the direct paths from CSE to job satisfaction to the value of zero, and compares this constrained model with the unconstrained model (see Table 5). The unconstrained model with the direct relationships has a significant decrease in χ2 (full model: Δχ2 = 28.07, df = 1, p < .001; Competing model with multi-groups: Δχ2 = 29.03, df = 2, p < .001), and the model fitness indices for the unconstrained model are slightly better than those for the constrained model. According to Hair et al. (Reference Hair, Black, Babin and Anderson2010), the result suggests that there is not complete mediation.
In terms of supervisors, CSE significantly relates to guanxi, and guanxi relates to job satisfaction based on the unconstrained model. However, the relationship between CSE and job satisfaction remains significant as guanxi is included in the model. Accordingly, the partial mediation of guanxi in the relationship between CSE and job satisfaction is concluded. The indirect effect (0.52 × 0.55) is smaller than the direct effect (0.45), making this a partial mediation situation. Likewise, the unconstrained model in Table 4 shows that subordinate CSE positively relates to guanxi that further relates to job satisfaction. The inclusion of guanxi does not alter the significant relationship between CSE and job satisfaction. The indirect effect (0.29 × 0.45) is smaller than the direct effect (0.22), indicating a partial mediation situation. Consequently, these findings partially support Hypothesis 3.
Supervisor–subordinate fit
This study is interested in how fit supervisors with different personality traits and subordinates with different personality traits are. In preparation I paired responses from supervisors and subordinates in the sample. Subordinates in the same company have information related to supervisors in common in that they share the same supervisor. Note that 10 subordinates were deleted because most data of their supervisors were missing. After pairing, the total sample of supervisor–subordinate dyads in the fit analyses equals to 427.
To examine Hypothesis 4 and Hypothesis 5, this study adopted polynomial regression analysis and three-dimensional surface plot analysis, given the approaches help better understand the precise nature of the relationship between the (dis)similarity of CSE and outcome variables (i.e., supervisor–subordinate relationship and job satisfaction) (Edwards & Parry, Reference Edwards and Parry1993; Edwards, Reference Edwards1994). Following Edwards and Parry (Reference Edwards and Parry1993) and Shanock, Baran, Gentry, Pattison, and Heggestad's (Reference Shanock, Baran, Gentry, Pattison and Heggestad2010) procedures, a separate polynomial regression was run for each of traits and each of the dependent variables.
The polynomial regression involving the interaction term and the squared terms may cause multicollinearity problem. To reduce the potential for multicollinearity and facilitate interpretation of the graphs, all predictor variables were scale-centered by subtracting a value of four from each score because CSE was measured on a 7-point Likert-type scale (Edwards, Reference Edwards1994).
Table 6 presents the results of the polynomial regression analyses, and Figure 2 shows selected surface plots. In the graph, the fit line runs across the floor of the graph from front to back, and the line of misfit runs across the floor of the graph from left to right. The results indicate that the amount of variance explained in supervisor–subordinate guanxi and job satisfaction by the combination of five CSE terms is statistically significant (R 2 = 0.16~0.46, p < .001). Thus, the results of the polynomial regression are evaluated with regard to four surface test values.
Table 6 (Dis)similarity of CSE as predictor of supervisor–subordinate relationship
Notes. n = 427. Unstandardized regression coefficients are reported.
CSE1: Supervisor's CSE; CSE2: Subordinate's CSE.
Guanxi1: Supervisor's perception toward guanxi; Guanxi 2: Subordinate's perception toward guanxi.
JS1: Supervisor's job satisfaction; JS2: Subordinate's job satisfaction.
*p < .05; **p < .01; ***p < .001.
Figure 2 CSE predicting guanxi and job satisfaction. CSE=core self-evaluation
The significant and positive slope along the fit line in Table 6 indicates that congruence at high levels of CSE results in higher outcomes, including supervisor's guanxi (0.85, p < .01), supervisor's job satisfaction (0.92, p < .001), subordinate's guanxi (1.14, p < .05), and subordinate's job satisfaction (0.90, p < .001), than congruence at low levels. In addition, the slope along the line of misfit for supervisor's job satisfaction is significantly positive (0.79, p < .001), indicating that when a supervisor's CSE level is higher than his/her subordinate's, his/her job satisfaction increases more sharply than it does when the supervisor's CSE level is lower than the subordinate's. Figure 2b also exhibits a shift toward the region where supervisor's CSE (CSE1) is greater than subordinate's CSE (CSE2).
Taken together, the (dis)similarity of CSE conditions cannot always apply to all of outcome variables. Accordingly, the research findings do not entirely support or reject Hypothesis 4 and Hypothesis 5. Nevertheless, the similarity effect is mostly prevalent for guanxi and job satisfaction. To clarify, Table 7 summarizes the conditions leading to high outcomes.
Table 7 The (dis)similarity of CSE condition leading to higher supervisor−subordinate guanxi and job satisfaction
Note. CSE1: Supervisor's CSE; CSE2: Subordinate's CSE.
Discussion
Conclusions
This study attempts to extend theoretical understanding by empirically investigating the antecedent of supervisor–subordinate guanxi and its influences on job satisfaction. I also seek to understand whether similarity or dissimilarity between a supervisor and a subordinate facilitates guanxi and job satisfaction. By examining the antecedent I find that CSE is a stronger antecedent to predict the supervisor–subordinate guanxi and job satisfaction for both supervisors and subordinates. The similarity in the CSE between the supervisor and the subordinate can augment both parties’ guanxi and job satisfaction.
The positive path from guanxi to job satisfaction has a highest estimate among all paths. This finding suggests that guanxi is strongly dominant in predicting job satisfaction. This is consistent with previous studies. Liden and Graen (Reference Liden and Graen1980) suggested subordinates who reported high-quality relationships with their supervisors assumed greater job responsibility and contributed more to their units. Henderson and Argyle (Reference Henderson and Argyle1986) advanced that better social relations led to better communications and more cooperation. These studies construct the reasons explaining the beneficial effect of guanxi on the job satisfaction.
Fit theory
In terms of the fit theory, the research findings suggest that neither of the (dis)similarity perspective is overwhelmingly conclusive. But the similarity perspective gains more supports from most of the research findings. Specifically, a congruence at high levels of CSE results in higher levels of the supervisor and subordinate's guanxi than a congruence at low levels. The CSE concept represents the fundamental assessments that people make about their worthiness, competence, and capabilities (Judge et al., Reference Judge, Bono, Erez and Locke2005). This congruence effect of CSE is consistent with the similarity-attraction paradigm that people who have certain characteristics in common elicit positive responses in one another and thus form positive relationships (Varma, Stroh, & Schmitt, Reference Varma, Stroh and Schmitt2001). Accordingly, the high-high CSE condition leads to a liking for each other and the liking in turn helps develop a better vertical guanxi.
The congruence effects of CSE between the supervisor and the subordinate on both parties’ job satisfaction support the similarity perspective. That is job satisfaction levels are higher when both the supervisor and subordinate are higher in CSE than when both are lower in traits. As stated earlier, CSE describes the positive evaluation of one's own worthiness and capability. The result implicates that individuals include evaluations of the capabilities of either their supervisors or subordinates as appraising their jobs. This, in turn, predicates their self-evaluated job satisfaction.
Not only a congruence at high levels of CSE but also a condition in which the supervisor's CSE is greater than the subordinate's result in a high supervisor's job satisfaction. Compared with subordinates, supervisors need to perceive more positive assessments about their worthiness and capability from others because of their positions in the higher level of hierarchy. Although most findings support the similarity perspective, the findings are not contradictory to the notion of complementarity in the fit literature. The idea of complementarity of dissimilar personalities was applied to outcomes related to organizational performance and personal development, rather than job satisfaction (Glomb & Welsh, Reference Glomb and Welsh2005). Besides, the findings related to fit analysis also reveal an idea that the research focusing on the fit between a supervisor and subordinate dyad should take the hierarchical role into account.
Research contributions
This study attempts to make several contributions. First, this study adds an important nuance to the acclaimed benefits of role of guanxi in workplace. While LMX indicates the extent to which the supervisor and subordinate exchange resources and support beyond the formal employment contract in terms of the working context (Graen & Uhl-Bien, Reference Graen and Uhl-Bien1995), guanxi describes the off-the-job interaction and friendship between the supervisor and subordinate without regarding to their working relationship. People can work together (i.e., LMX) but cannot necessarily be friends (i.e., guanxi). Likewise, a good friendship is not necessarily beneficial to work. I document that the supervisor–subordinate guanxi, initiated and developed by extra-role activities, conveys benefits to in-role activities.
Second, I identify the antecedents of the supervisor–subordinate guanxi. Thus, the findings can offer implications for management: how to select an appropriate employee who is comparatively fit with management. In order to recognize who fit with each other, this study intends to find out one pair, including a supervisor and a subordinate with certain levels of CSE, leads to an optimal supervisor–subordinate guanxi and job satisfaction. Third, this study shifts the focus on the impact of comparative personality traits in supervisor–subordinate relationship dyads and offers alternate view to the fit theory. The findings can bridge the gap of literature.
Research implications
The results make some practical implications. First, the off-the-job supervisor–subordinate relationship favors the job-related outcome. The positive effect of guanxi on the job satisfaction reflects the mood-congruent theory. The mood-congruent theory suggests that individuals often make mood-congruent judgments (Xin & Pelled, Reference Xin and Pelled2003). Good off-the-job interactions facilitate producing good moods, and thus lead individuals to make mood-congruent judgments on work-related matters. The fine friendship that is promoted by off-the-job social activities can help build a happy working environment or atmosphere. Thus, the guanxi eventually enhances the job satisfaction.
Second, the managers need to create opportunities for cultivating a good guanxi by encouraging some off-the-job activities, such as visits, dinners, gift-giving (Chen & Tjosvold, Reference Chen and Tjosvold2006). Finkelstein, Protolipac, and Kulas (Reference Finkelstein, Protolipac and Kulas2000) also stated that both non-work activities sponsored by organizations, such as involvement in wellness programs, outdoor treks, or company sports teams, and activities not organizationally sponsored, such as after work happy hour, were needed to develop extra-role relationships. Proactive socialization behaviors have been found to be positively associated with beneficial outcomes for supervisors and subordinates (Wang & Kim, Reference Wang and Kim2013). These social activities strengthen the supervisor's and subordinate's perceptions of their bonds to each other in that a supervisor–subordinate relationship is not only built upon works but also personal interactions. Furthermore, according to the beneficial effect of guanxi on job satisfaction, the managers have to use guanxi for instrumental purposes to accomplish tasks (Chen et al., Reference Chen, Friedman, Yu and Sun2011).
Third, the rationale behind the relationship between CSE and job satisfaction can be explained by the guanxi. Accordingly, for increasing job satisfaction, the personality can be a signal for selecting members who can get along with supervisors. When a subordinate is more aligned with a supervisor at a high level of CSE, the levels of subordinate-rated and supervisor-rated job satisfaction are higher. In this regard, individuals’ CSE and their guanxi should be taken into consideration as forming a work group. A leadership relationship exists between supervisors and subordinates. The supervisor's leadership effect can be activated and further trigger the subordinate's positive behaviors when both of the supervisor and subordinate have high levels of CSE. Specifically, the supervisor with a high CSE tends to employ the transformational leadership skill (Hu, Wang, Liden, & Sun, Reference Hu, Wang, Liden and Sun2012), and the subordinate's CSE facilitates transferring the leadership effect to the subordinate's positive behaviors (Zhang & Peterson, Reference Zhang and Peterson2011; Hu et al., Reference Hu, Wang, Liden and Sun2012; Nübold, Muck, & Maier, Reference Nübold, Muck and Maier2013). That is why the supervisor prefers the subordinate who has a high CSE. So the subordinate's CSE motivates the supervisor to build guanxi with the subordinate. Both the supervisor and subordinate with high levels of CSE become satisfied with their jobs because they can work with favorable partners and can effectively deal with tasks.
Limitations and future research suggestions
This study has limitations and these suggest avenues for future research. First, I confined our sample to banks to control the influence of industry type. To ensure the wide applicability of our findings, future research should test this related issue with a sample of firms active in a variety of industrial sectors. Second, this study focuses on Chinese people's guanxi as most studies of guanxi did. Future research can explore the impacts of guanxi between a foreign supervisor and a Chinese subordinate.
Third, this study aims on the fit personality traits. However, the (in)congruence effect of supervisor–subordinate relationship on job-related outcomes reveals another interesting angle of research. In effect, both parties in a relationship do not necessarily have the same evaluations of the relationship. One party who reports highly of the relationship is more likely to do favors for others. Although the data in this study suffices for examining the (in)congruence effect of supervisor–subordinate relationship, this study is short of space to deal with this big issue. Future research can explore this issue based on the sound theoretical basis.
Fourth, this study only focuses on the bright side of the supervisor–subordinate relationship and ignores the dark side such as the vertical conflict. Conflict is delineated the task conflict from the relationship conflict and is a significant feature of the supervisor–subordinate interaction (Xin & Pelled, Reference Xin and Pelled2003). It is interesting to treat the vertical conflict as the mediator in the relationship between CSE and job satisfaction, and to compare its mediating effect with guanxi's mediating effect. In addition, the mediating effect of vertical conflict may vary with the type of conflict and the position in the organization (e.g., supervisor and subordinate). This issue deserves to attention from future researchFootnote 3.
In sum, this study explores the individual and congruence effects of CSE on guanxi, and outlines the beneficial consequence of guanxi on the job satisfaction. An individual with a high CSE can gain the high job satisfaction because he/she can enjoy the benefit resulting from the vertical guanxi. In addition, a supervisor and a subordinate can get along with each other when both have high CSE. The findings offer implications for managing human resources.
Acknowledgement
Funding of this research was provided by the National Science Council, Taiwan, under grant NSC98-2410-H-468-016.
