چرا رهبری تحول گرا برای گردش مالی کارمند مهم است؟چشم انداز مبادلات اجتماعی چند کانونی

Drawing on social exchange theory, the present study investigates the underlying mechanisms through which transformational leadership influences employee turnover. Leader–member exchange (LMX) and affective commitment (AC) are proposed as supervisor-based and organization-based social exchange mechanisms respectively, exemplifying how social exchange processes occur between an employee and his/her supervisor, and between the employee and his/her organization as a whole to underpin the effect of transformational leadership on turnover outcomes. Results of structural equation modeling on a sample of 490 full-time employees working in a large telecommunication company in the PRC provided support for the notion that transformational leadership is related to both social exchange mechanisms – LMX and AC – turnover intention and turnover behavior. Furthermore, the results revealed that AC rather than LMX mediated the link between transformational leadership and turnover intention. Turnover intention also only mediated the relationship between AC and turnover behavior over time.

Over the last few decades, a great deal of research attention has been devoted to exploring factors that influence voluntary employee turnover in organizations (Allen et al., 2003, Hancock et al., 2013 and Shaw et al., 2005). This widespread interest stems from the fact that employee turnover is detrimental and expensive (Mueller & Price, 1989). Recruiting, selecting and training new employees to cover production deficiencies, human capital development and employee well-being are very costly (Mossholder et al., 2005 and Shaw et al., 2005). Recent research has revealed that voluntary employee turnover is associated with a great variety of negative effects, including depressing financial performance, declining employee work attitudes and undermining workforce productivity (see Park & Shaw, 2013). An understanding of how to manage employee turnover will therefore provide organizations with valuable and unparalleled resources for operational effectiveness and employee well-being (Griffeth et al., 2000 and Hancock et al., 2013). Given the growing interest in utilizing work teams in organizations, effective leadership has become increasingly important (Tse & Chiu, in press). Work team supervisors are not only required to maximize individuals' contributions for organizational effectiveness, but must also retain their skills and capabilities for their organization's competitive advantage (Harris et al., 2011 and Waldman et al., in press). Indeed, a review of research has indicated that social exchange theory has been used to underpin the implication of transformational leadership for important work outcomes including job satisfaction, task performance, helping behavior, creativity, job-related stress and burnout (see Judge and Piccolo, 2004, Lowe et al., 1996 and Wang et al., 2011). This line of research has suggested that transformational leadership is effective in engaging subordinates in social exchange processes based on interpersonal trust, mutual loyalty, strong identification, and ongoing reciprocity with their supervisors. Subordinates thus feel indebted and obliged to repay their supervisors and organizations in kind over time (Walumbwa, Cropanzano, & Hartnell, 2009). Although this theoretical perspective of transformational leadership has been made explicit in the literature, it has not yet been used to explain whether transformational leadership can deter subordinates from forming intentions to leave and acting on those intentions (Martin and Epitropaki, 2001 and Waldman et al., in press). Research to date has focused on understanding the relationship between transformational leadership and employee turnover rather than providing an integrated framework to better understand the underlying process of the relationship (e.g., Bycio et al., 1995, Hughes et al., 2010 and Waldman et al., in press). It is therefore important to understand how and why social exchange theory is useful for exploring the underlying mechanisms through which transformational leadership induces retention of employees. Employees engage in social exchange relationship with their immediate supervisor and/or with the organization as a whole (Maertz, Griffeth, Campbell, & Allen, 2007). Strong identification, interpersonal trust, and mutual support between the employee and these two social entities may increase employees' propensity to stay in their organization. Building on the target-similarity framework (Lavelle, Rupp, & Brockner, 2007), we propose to test a multi-foci social exchange model of transformational leadership and employee turnover. Specifically, we propose two social exchange mechanisms that are significant to our inquiry: 1) leader–member exchange (LMX) (i.e., an individual's perception of the quality of the dyadic relationship he/she develops with his/her supervisor; see Dansereau, Graen, & Haga, 1975) is conceptualized as a supervisor-based social exchange mechanism that exemplifies how an employee engages in a social exchange with his/her immediate supervisor; and 2) affective commitment (AC) (i.e., an individual's perception of his/her emotional attachment and affective identification with his/her organization; see Meyer & Allen, 1991) is conceptualized as an organization-based social exchange mechanism that reflects how an employee engages in a social exchange with his/her organization as a whole (for evidence supporting this framework, see Lavelle et al., 2007 and Rupp and Cropanzano, 2002). We contribute to the transformational leadership and turnover research by addressing the repeated calls for exploring the underlying mechanisms through which transformational leadership is effective in managing turnover processes (Griffeth et al., 2000, Hancock et al., 2013 and Park and Shaw, 2013). This study is the first to adopt social exchange theory as an overarching framework to examine LMX and AC as the supervisor-based and organization-based social exchange mechanisms linking transformational leadership to employee intention and turnover, and turnover behavior over time (Colquitt et al., 2013 and Waldman et al., in press). Our study also provides insights into the relative importance of supervisor-based and organization-based social exchange mechanisms on the transformational leadership-turnover relationship. It remains largely unclear in the literature whether transformational leaders reduce employee turnover through enhancing LMX or AC or both. On the one hand, transformational leaders are able to induce employee staying by expressing individualized consideration to develop a strong personalized exchange relationship with their employees (Hughes et al., 2010). On the other hand, transformational leaders can transcend employees' self interests to organizational interests by inducing a social exchange between the organization and employees to mitigate their turnover intention and behavior (see Martin & Epitropaki, 2001). We therefore attempt to empirically test the relative strengths of the mediating roles of LMX and AC in order to advance a more nuanced understanding of the specific influencing processes of transformational leadership. Moreover, little research on transformational leadership and employee turnover has been conducted in non-western contexts, such as China (Waldman et al., in press). This study thus intends to increase the external validity of the implications of transformational leadership for turnover processes across cultures, according to the global management perspective. An understanding of the applicability of transformational leadership in different cultural contexts could help develop universal practice for global leadership training (Kirkman, Chen, Farh, Chen, & Lowe, 2009). In the following section, we provide a rationale underlying our model development, and develop theoretical arguments supporting each of the hypothesized relationships. We begin by discussing why and how social exchange theory can be used as an overarching framework in this study, based on a review of turnover literature. We then explain how LMX and AC can be conceptualized as the supervisor-based and organization-based social exchange mechanisms which influence the relationship between transformational leadership and employee turnover intention. Finally, we present arguments explaining the theoretical basis of the mediating role of turnover intention in the relationships between both social exchange variables and turnover behavior over time.

Previous studies have suggested that transformational leadership has a negative impact on employee turnover intention and behavior. However, we do not know how transformational leaders actually retain employees and reduce their turnover rate. This study advances our understanding of the underlying mechanisms of transformational leadership by testing whether transformational leaders retain employees through triggering a high level of supervisor-based social exchange, organization-based social exchange with their subordinates, or both. We found that the negative relationship between transformational leadership and employee turnover intention was mediated by AC, which captures organization-based social exchange, rather than by LMX, which represents supervisor-based social exchange. Consistent with the results of past studies, we also found that turnover intention mediates the link between AC and turnover behavior at both Time 2 and Time 3 (Griffeth et al., 2000 and Martin and Epitropaki, 2001). Although we did not hypothesize the differential effects of LMX and AC on the relationship between transformational leadership and turnover intention, our findings showed that AC is a stronger social exchange mechanism, translating the leadership effect into turnover intention. Our findings lend support to a key feature of transformational leadership, in that such leaders are likely to induce employees to stay, not because they build high-quality personalized exchange relationships with subordinates, but because they are able to inspire employees to transcend their individual interests and orient themselves to the collective interests of the organization and to make a high level commitment to their organization. One possible explanation for these findings is that employees who do not experience high-quality LMX relationships with their leader may seek to transfer to another work team instead of leaving the organization outright. The leader in another work team may develop and maintain strong supervisor-based and organization-based social exchange relationships with employees who may in turn be more likely to maintain their organizational membership. 5.1. Theoretical contributions The findings of the current study have several implications for theory. First, although only a few studies have been conducted into the link between transformational leadership and turnover intention in organizational settings (see Bycio et al., 1995, Hughes et al., 2010 and Martin and Epitropaki, 2001), little attention has been directed toward the role of transformational leadership in the actual turnover process (Griffith, 2004 and Waldman et al., in press). We attempt to address this issue by including both turnover intention and turnover behavior in our model. While we did not hypothesize the main effect of transformational leadership on turnover outcomes, our findings do show that transformational leadership has a negative impact on turnover intention at Time 1 and on turnover behavior at Time 2 and Time 3 over a period of a year and a half. These findings are consistent with past research (e.g., Griffith, 2004 and Waldman et al., in press), supporting the notion that transformational leadership is effective in accounting for variance in actual turnover over time after controlling for turnover intention. Second, our study shed new light on why and how exercising transformational leadership can help deter followers from forming intentions to leave and then acting on those intentions. Underpinned by the multi-foci social exchange model, our findings suggest that, although transformational leadership can facilitate both supervisor-based and organization-based social exchanges, it is the organization-based social exchange (AC) that translates the leadership effect into turnover intention. This lends support to key features of transformational leadership, in that such leaders are likely to induce employees to stay, not because they can build a high-quality of personalized relationship with subordinates (LMX), but because they are able to inspire employees to transcend their individual interests and to orient themselves to collective interests, arousing a high level of affective commitment to their organization (Avolio, 1999 and Bass, 1985). Subordinates are thus more likely to respond to transformational leaders by taking the organization's collective interests into account, instead of simply their leaders' interests and goals. This echoes the notion of whether supervisors are perceived to represent or personify the organization (Levinson, 1965). The findings here imply that subordinates tend to direct their reciprocating attitudes and behaviors toward the target from which benefits originate, even though their supervisors are likely to be seen as agents representing the organization translating its benefits, resources and support by displaying transformational leadership (Lavelle et al., 2007 and Maertz et al., 2007). Subordinates may thus continue their reciprocal transactions with the organization by seeking to transfer to another work team in order to withdraw from a low-quality exchange relationship with their leader. Third, the current study informs the existing cross-cultural leadership research about the universal impact of transformational leadership in organizational context worldwide (Kirkman et al., 2009). Although we did not collect data for any cultural variables to test whether power distance and collectivism potentially influence the hypothesized relationships using the Chinese sample in this study, our findings reveal that the mediating effect of AC was stronger than LMX on the transformational leadership-turnover relationship. This should not be the case because (guanxi:LMX) —personal relationship plays a dominant role in the Chinese culture (see Xin & Pearce, 1996). Our findings may provide support for other studies confirming that transformational leadership influences important work outcomes similarly in culturally distinct countries such as PRC and US (see Avolio et al., 2004, Kirkman et al., 2009 and Wang et al., 2005). Finally, our findings also extend Van Breukelen et al.'s (2004) study and research on the intention–behavior relationship by theorizing that organization-based exchange and supervisor-based exchange were exemplified by AC and LMX as the external factors that predict turnover behavior through turnover intention. This in turn determines turnover behavior between two points in time 1 year apart. Although our results reveal that only the organization-based exchange (AC) was significant in determining turnover behavior via turnover intention, these results are still important because Van Breukelen and colleagues failed to provide support for the role of AC in their model. The current study further increases our understanding of TPB to confirm its applicability in the context of turnover by showing that turnover intention is mediated by the AC-turnover behavior relationship. In addition, our results further indicate the unexpectedly weak and inconsistent relationship between turnover intention and turnover behavior (Griffeth et al., 2000, Hom and Griffeth, 1995 and Park and Shaw, 2013). 5.2. Practical implications This study has several practical implications. First, it builds upon social exchange theory by determining that subordinate–supervisor social exchange and subordinate–organization exchange are the underlying bases for explaining why and how transformational leadership reduces employee turnover. Our findings reveal that the effect of transformational leadership on turnover intention is indirect, mediated through organization-based exchange (AC) rather than supervisor-based exchange (LMX). The findings suggest that transformational leadership is effective in managing employee turnover, enabling subordinates to internalize thane organization's values and mission, and thus encouraging them to be proud of their organizational membership (Bass and Riggio, 2006 and Shamir et al., 1993). Effective leadership training programs should be developed to make sure that transformational leaders are able to facilitate organization-based social exchanges at work. This leadership behavior helps build employees' emotional attachment to their organization, which in turn creates the desire to retain membership of that organization (Bass, 1985). Supervisors thus need to pay attention to their subordinates' perceptions of their organizational mission, vision, goals, and objectives. Effective transformational leaders should be aware of the need to adopt different practices in order to facilitate a sense of affective commitment among their subordinates. Second, this study was conducted in a large call center of a telecommunication company in a Chinese setting. Our findings may inform western practitioners or managers about the practical implications of transformational leadership in the PRC, which is considered to be one of the world's fastest developing countries (Waldman et al., in press). Based on our findings, the variables included in this study appear not to be culturally specific: our results relating to the hypothesized relationships are similar to the findings reported in western studies (e.g., Bycio et al., 1995 and Hughes et al., 2010). This implies that transformational leadership plays an important role in employee turnover processes across cultures (Waldman et al., in press and Wang et al., 2011). Universal practice can thus be formulated to promote global leadership development for expatriates working in leadership positions in China (Kirkman et al., 2009). 5.3. Limitations and future research directions The findings of the present study should be considered in light of several limitations, each of which should be addressed by future research. First, our study sample consisted of employees working in one call center of a single telecommunication company, and the majority of the employment population in the call center industry is young and female, with a relatively low tenure. This could have introduced gender, age and tenure biases into our findings. For example, if older longer tenured male employees had a high turnover rate, failing to include an equal number of older longer tenured male employees may have affected the hypothesized relationships. Thus, the findings may have been different if there had been more older, male and longer tenured employees included in our sample. Although failing to include a sufficient number of older, male and longer tenured employees in this study may have led to higher or lower means within the study variables, it is arguable that this may have affected only the mean values. However, the associations between the variables would have been of similar magnitude, regardless of means. To clarify this sampling issue, we include gender, age, and tenure as control variables to test their effects on turnover behavior, and the results show that gender, age, and tenure were not found to be related to turnover intention (Time 1) or turnover behavior (Time 2), and only tenure was somewhat associated with turnover behavior (Time 3). The results suggest that imbalanced data for gender, age, and tenure did not have a strong negative impact on the findings of this study. Nevertheless, further research should be conducted to replicate this study using a sample more balanced in terms of gender, age, and tenure in different organizational settings in order to improve confidence in the findings and their generalizability. A second problem associated with the sample may be the potential lack of a strong generalizability of our findings to other, western countries. The sample in our study was obtained from the PRC, which is collectivistic and scores high on power distance (Hofstede, 2001). These cultural characteristics may have an effect on the extent to which supervisor-based and organization-based social exchanges are perceived in work teams. We believe, however, that the current findings could still help inform the social exchange literature and target-specificity model about the implication of transformational leadership for turnover in western contexts because the Chinese sample rendered our analyses more conservative. Specifically, Chinese culture values personalized exchanges (i.e., LMX and guanxi) over other forms of social exchanges, such as AC (see Xin & Pearce, 1996). This suggests that supervisor-based social exchange (LMX) should have played a stronger mediating role than organization-based social exchange (AC) in the transformational leadership-turnover relationship. However, in this cultural context, we did not find LMX to have a significant mediating effect in the relationship. This implies that the effect of cultural variation may not have been as salient as it should have been in our study. Future research can continue to explore cultural variables (such as traditionality, collectivism, and power distance) as important contingent boundaries that would potentially influence the relationships between transformational leadership and turnover. A third limitation is related to the sample, and might also affect the generalizability of our findings to other call centers and organizational settings in different industries. Our sample was drawn from employees in a large telecommunication company and their supervisors. Research undertaken using a single organization always limits the findings' representativeness. However, the results presented in our study are similar to those obtained in other organizational settings, including hospitals (Waldman et al., in press), schools (Griffith, 2004), and aerospace design and manufacturing (Hughes et al., 2010). This suggests that our findings may still be generalizable to other organizational settings. Nevertheless, future research should replicate and extend our study using a more representative sample drawn from companies in different industries. A fourth limitation concerns the common-method bias that might potentially inflate the observed relationships in this study (Podsakoff et al., 2003). Although common-method bias was not identified as a major problem in the prediction of turnover behavior, which was provided independently by the company, the method bias might still apply because a common survey method was employed for data collection (Liao & Chuang, 2007). We thus attempted to address this concern by conducting CFAs to test the distinctiveness of the variables and by undertaking correlation analysis to examine their discriminant validity. In addition, although our data came from different sources (i.e., subjective data from employees and objective data from the organization's human resources department), we continued to adopt several procedural and statistical remedies suggested by Antonakis, Bendahan, Jacquart, and Lalive (2010), Podsakoff et al. (2003), and Richardson, Simmering, and Sturman (2009) to minimize potential bias. First, the anonymity and confidentiality of responses were guaranteed to reduce respondents' evaluation apprehension. Second, a psychological separation was constructed in the survey through the use of different instructions and by interspersing the variables throughout the survey, mixed with multiple filler items. This procedure helped lower respondents' perceptions of any direct connection between the variables. Third, we calculated the variance explained by the method factor (13.6%), and the result was much lower than the 25% suggested by Williams, Cote, and Buckley (1989, pp. 29–30). Nonetheless, future research should consider validating the findings of this study by collecting data from different sources and employing different types of research design (Podsakoff et al., 2003). A fifth limitation relates to the research design of this study. Turnover behavior was collected over two six-month intervals after we had administered the survey to obtain information on transformational leadership, LMX, AC, and turnover intention. Consistent with turnover research, our study had a less than ideal increase in turnover rate over 12 months (Griffeth et al., 2000 and Shaw et al., 2005). As a result, the turnover behavior at Time 2 and Time 3 (i.e., 6% at 6 months and 14% at 12 months) did not show substantial changes leading to a restriction of range in the outcome variables. This could have caused the observed relationships among the study variables to be weaker than when they are tested with a higher percentage of current turnover rates in Time 3. Future research should attempt to obtain information on the actual turnover rate over a longer period (e.g., 2 years or more) in order to increase the effect size for relationship testing. Another promising avenue for future research attention is to examine a more comprehensive social exchange-based model that incorporates other potential mediating variables, such as perceived organizational support (POS) and perceived organizational supervisor support (PSS), to influence the hypothesized relationship between transformational leadership and turnover behavior. Researchers such as Maertz and colleagues (2007) found support for the mediating effects of PSS and POS on turnover cognitions, and their interactive effects on turnover behavior. This study reflects the need for future research to look at these social exchange-based mediators of the leadership-turnover relationship. In conclusion, the present study represents a first attempt to explore the underlying process by which transformational leadership relates to turnover. We tested a social exchange model in which LMX and AC were theorized as supervisor-based and organization-based social exchange mechanisms, respectively, exerting their effect linking transformational leadership and turnover outcomes. Overall, SEM results show that transformational leaders can reduce turnover intention and actual turnover through enhancing subordinates' emotional attachment to, and affective identification with their organization, rather than by developing high-quality personalized exchange relationships with them. We hope that our findings inspire researchers to look at other possible mediating and moderating variables underpinning the social exchange model of transformational leadership and turnover in the near future.