Social exchange spillover in leader–member relations: A multilevel model

• DOI: http://doi.org/10.1002/job.2030
• Author: Zhiqiang Liu,Xiujuan Zhang,Lifeng Zhong,Ying Chen,Zhen Xiong Chen,Jooyeon Son
• Published date: 01 July 2015
• Date: 01 July 2015

Social exchange spillover in leader–member

relations: A multilevel model

YING CHEN

1

*, ZHEN XIONG CHEN

2

, LIFENG ZHONG

3

, JOOYEON SON

1

,

XIUJUAN ZHANG

4

AND ZHIQIANG LIU

5

1

School of Labor and Employment Relations, University of Illinois at Urbana-Champaign, Champaign, IL, U.S.A.

2

Research School of Management, Australian National University, Canberra, Australia

3

School of Business, Renmin University of China, Beijing, China

4

School of Business, Sun Yat-sen University, Guangzhou, China

5

School of Management, Huazhong University of Science and Technology, Wuhan, China

Summary Drawing on role theory and the cultural theory of collectivism, we developed and tested a multilevel model of

social exchange spillover in leader–member relations in the Chinese context. In Mplus analyses of a sample of

213 subordinates from 47 groups, we found that, at the individual level, a dimension of leader–member guanxi

(LMG), leader–member personal life inclusion (LMG-P), which is defined as the extent to which leaders and

members include each other in their personal or family lives, can spill over to affect subordinates’contextual

performance (i.e., interpersonal facilitation and job dedication); furthermore, this effect was moderated by sub-

ordinates’horizontal collectivism orientation, such that LMG-P spilled over to affect contextual performance

only for those who were low in horizontal collectivism orientation. At the group level, the variance of LMG-P

within a group, which is referred to as LMG-P differentiation, was related negatively to group performance

when the supervisors had a low horizontal collectivism orientation. At the cross level, LMG-P differentiation

moderated the relationship between LMG-P and job dedication, such that the relationship was positive only

when LMG-P differentiation was low. Copyright © 2015 John Wiley & Sons, Ltd.

Keywords: leader–member guanxi (LMG); leader–member exchange (LMX); collectivism; multilevel;

performance

Leader–member relationships have been demonstrated to play a very important role in influencing members’out-

comes in Chinese organizations (e.g., Farh, Tsui, Xin, & Cheng, 1998; Hui & Graen, 1997; Law, Wong, Wang,

& Wang, 2000). The current research on leader–member relationships conducted in Chinese organizations has been

based primarily on the construct of leader–member exchange (LMX), which was developed initially in the West.

However, because of the implicit cultural assumptions and limitations of the LMX construct, current LMX theory

cannot capture the indigenous domain of leader–member relations in Chinese organizations fully. By adopting a

contextualized research strategy in this study (Tsui, 2006, 2012), we examined a unique aspect of leader–member

relationships, which we term leader–member personal life inclusion (LMG-P: cf., Chen, Friedman, Yu, Fang, &

Lu, 2009), and developed a multilevel model of social exchange spillover from the private to the work domain in

leader–member relations in the Chinese context.

Leader–member personal life inclusion, which is one of the three dimensions of leader–member guanxi (LMG), is

defined as the extent to which leaders and members include each other in their family or personal lives (Chen et al.,

2009). As indicated in its definition, LMG-P is clearly private domain-specific and thus fits our theoretical inquiry.

However, the othertwo dimensions of LMG in Chen et al.’s (2009) study,affective attachment (LMG-A)and deference

to supervisor (LMG-D),are not domain-specific. As such, we treatedthese two dimensions as controls in our analyses.

Leader–member exchange refers to the quality of the exchange relationship between supervisors and subordinates

(Graen & Scandura, 1987). It reflects the Western cultural preference that business relationships should not be

*Correspondence to: Ying Chen, School of Labor and Employment Relations, University of Illinois at Urbana-Champaign, 504 East Armory

Avenue, Champaign 61820, IL, U.S.A. E-mail: ychen01@illinois.edu

Copyright © 2015 John Wiley & Sons, Ltd.

Received 09 February 2014

Revised 07 May 2015, Accepted 07 May 2015

Journal of Organizational Behavior, J. Organiz. Behav. 36, 673–697 (2015)

Published online 15 June 2015 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/job.2030

Special Issue Article

intertwined with personal life, and therefore, it focuses implicitly on working relationships. In contrast, LMG is

modeled after family relationships, and the quality of LMG refers to the extent to which a leader–member

relationship has become a pseudo-family relationship (Chen et al., 2009).

We acknowledge that LMX is a global construct, the prevalence of which has been confirmed in a recent

meta-analysis (Rockstuhl, Dulebohn, Ang, & Shore, 2012). At the same time, we contend that this does not mean

that LMX theory has captured all the aspects of leader–member relationships from an international perspective.

Leader–member relationship research should explore both etic (transferable) and emic (culture-specific) aspects of

those relationships (Rockstuhl et al., 2012). Recently, researchers surveyed the three dimensions of LMG (Chen

et al., 2009) in Taiwan, Singapore, and six non-Chinese cultures (Brazil, India, Russia, Saudi Arabia, Turkey, and

the UK) and found that two subscales—affective attachment and deference to supervisor—demonstrated full metric

invariance across all eight samples, whereas LMG-P appeared to be a Chinese culture-specific construct (Smith

et al., 2014).

The uniqueness of the LMG-P phenomenon is that it allows or even facilitates interpenetration between the

private and work domains. Although previous research on LMG has studied the relationships between dyadic

LMG and “self”-oriented outcomes, such as promotion opportunities (e.g., Law et al., 2000), little research has been

conducted to explore whether or not socializing with leaders in the private domain is beneficial to other-oriented

work outcomes, such as contextual performance. Furthermore, even though Chinese are very sensitive to informa-

tion about who has been included in their leaders’personal or family lives, and scholars have speculated about the

possible negative externality effects of guanxi practices on group and organizational outcomes (Chen & Chen,

2009), a recent review of guanxi research revealed that group-level guanxi research is “almost completely absent”

(Chen, Chen, & Huang, 2013, p. 195). Hence, we know little about how the composition of high-quality or

low-quality LMG-P affects group effectiveness. The boundary conditions between LMG-P and contextual and

group performance are also unclear.

Drawing on role theory and the cultural theory of relational collectivism, the purpose of this study was to develop

and test a multilevel model of social exchange spillover in leader–member relations. Role theory suggests that role

extension leads to the expansion of role behavior requirements (Kahn, Wolfe, Quinn, Snoek, & Rosenthal, 1964;

Katz & Kahn, 1978). The theory suggests further that organizations are essentially a set of work roles, and thus, role

differentiation and coordination are critical for collective success (Katz & Kahn, 1978). Looking at LMG-P from the

perspective of role extension, we argue that, at the individual level, a subordinate who has high LMG-P with his or

her supervisor in the private domain will tend to display higher contextual performance in the work domain.

Contextual performance is defined as “a set of interpersonal and volitional behaviors that support the social and

motivational context in which organizational work is accomplished”(Van Scotter & Motowidlo, 1996, p. 525).

Based on the cultural theory of relational collectivism (Brewer & Chen, 2007), we posited further that, at the

individual level, subordinates’horizontal collectivism orientation can moderate the spillover effect of LMG-P on

contextual performance. At the group level, we proposed that a supervisor’s horizontal collectivism orientation

moderates the relationship between LMG-P differentiation, defined as the variance of LMG-P at the group level,

as well as group performance. Moreover, LMG-P differentiation moderates the spillover effects of LMG-P on

contextual performance. Figure 1 illustrates our hypothesized theoretical framework.

We attempted to provide three theoretical extensions to the LMG and LMX literatures. First, research has

demonstrated a positive relationship between LMG and work outcomes that benefit LMG beneficiaries, such as

promotion opportunities (Law et al., 2000). However, we know little about whether LMG in the private domain

is related to work outcomes that can benefit coworkers and the organization. Studying the relationship between

LMG-P and contextual performance can help solve this puzzle. LMG-P also enriches current LMX theory by

demonstrating that an indigenous construct (i.e., LMG-P) can contribute to the development of a more global

leader–member relationship theory. Second, no prior studies have examined the LMG-P spillover effects

concurrently at both the individual and group levels of analysis. The multilevel social exchange model provides

an integrated picture that enables us to understand more clearly how LMG-P affects work outcomes across levels,

thus deepening our understanding of the complexity of the LMG phenomenon in the Chinese context. Third, we

674 Y. CHEN ET AL.

Copyright © 2015 John Wiley & Sons, Ltd. J. Organiz. Behav. 36, 673–697 (2015)

DOI: 10.1002/job