Paternalistic leadership and subordinate responses: Establishing a leadership model in Chinese organizations

(1)

Asian Journal of Social Psychology (2004) 7: 89–117

Blackwell Science, LtdOxford, UKAJSPAsian Journal of Social Psychology1367-22232004 Blackwell Publishing Asia Pty Ltd with the Asian Association of Social Psychology and the Japanese Group Dynamics Association 2004 April 20047189117Original ArticlePaternalistic leadershipBor-Shiuan Cheng

et al.

Correspondence: Bor-Shiuan Cheng, Department of Psychology, National Taiwan University, #1 Sec, 4-Roosevelt Road, Taipei 106, Taiwan. Email: chengbor@ccms.ntu.edu.tw

REGULAR ARTICLE

Paternalistic leadership and subordinate

responses: Establishing a leadership model

in Chinese organizations

Bor-Shiuan Cheng, Li-Fang Chou and

Tsung-Yu Wu

Department of Psychology, National Taiwan University, Taipei, Taiwan

Min-Ping Huang

Department of Business Administration, Yuan Ze University, Taoyan, Taiwan

Jiing-Lih Farh

Department of Management of Organizations, Hong Kong University of Science and Technology, Hong Kong

Paternalistic leadership (PL) is the prevalent leadership style in Chinese business organizations. With an approach similar to patriarchy, PL entails an evident and powerful authority that shows consideration for subordinates with moral leadership. Although PL is widespread in Chinese business organizations, very few studies have focused on this leadership style and those that have were simply conceptual analyses and not empirical studies. We sampled 543 subordinates from local businesses in Taiwan to investigate PL, Western transformational leadership, and subordinate responses to these two leadership styles. Our hypotheses were as follows: (1) PL has a significant and unique effect on subordinate responses compared to Western transformational leadership; (2) there exists an interaction between the three elements of PL (benevolence, morality, and authoritarianism) and subordinate responses; and (3) the authority orientation of a subordinate’s traditionality has a moderating effect upon the relation between PL and subordinate responses. Statistical analyses generally supported these hypotheses. Directions for follow-up studies are offered and implications for leadership theory and practice are discussed.

Keywords: authoritarian leadership, benevolent leadership, moral leadership, paternalistic leadership, subordinate responses.

(2)

90 Bor-Shiuan Cheng et al.

Introduction

Along with the developing economies in East Asia and China, studies concentrating on leadership in Chinese organizations have been gaining more attention. Instead of adopting an etic approach, which transplants Western leadership theory to a local context, many researchers have chosen an emic (or indigenous) approach to explore leadership in Chinese organizations. Following Silin’s (1976) research on a large business in Taiwan, Redding (1990) and Cheng (1995a) also adopted an indigenous approach to explore medium- and high-level leadership issues within Chinese family businesses (CFB) throughout Hong Kong, Indonesia, Singapore, and Taiwan. Results indicated distinct and clear-cut features of leadership in CFB. This leadership style was labeled paternalistic leadership (PL) (Westwood & Chan, 1992).

Despite the prevalence of PL in Chinese business organizations, little is known about it. For example, a recent study (Farh & Cheng, 2000) indicated that unless major PL-associated issues are clarified, our knowledge of PL is incomplete, and precise predictions concerning its relationship with the organization, group, and subordinates cannot be made. These issues include: What are the theoretical constructs of PL? What is the best way to measure them? As compared to a typical Western leadership style, what are the unique effects of PL on subordinates? Is there contextual influence on these effects? What kind of association exists among the three components of PL and how do these associations affect subordinates’

responses? Currently, only the first two issues have been addressed (Cheng et al., 2000), while

the remaining issues still await further exploration. The intention of the present study is to clarify the following issues: (1) Compared to transformational leadership, the most popular leadership style in the West, does PL have a unique effect upon subordinate responses that reflects psychological reactions to the leader? (2) Do interactions among the three elements of PL (benevolence, morality, and authoritarianism) have any effect on subordinate responses? (3) Does the authority-related aspect of a subordinate’s traditionality have a moderating effect on the relations between PL and subordinate responses?

Theory and hypotheses

Paternalistic leadership and its cultural roots

In the 1960s while Taiwan’s economy was still developing, Silin (1976) found that the leadership concepts and behavioral styles of business owner/managers in Taiwan were greatly different from those in the West. His conclusion was based on a year’s observation and many interviews conducted in a large private enterprise in Taiwan. The differences were displayed via various behaviors such as didactic leadership, moral leadership, centralized authority, maintaining social distance with subordinates, keeping intentions ill-defined, and implementing control tactics (Silin, 1976). Inspired by the phenomenal success of overseas CFB, in the 1980s, Redding (1990) identified, through in-depth interviews with 72 owner/ managers, a distinct brand of economic culture labeled as Chinese capitalism, in which paternalism is a key element. Although Redding’s depiction of PL is similar to Silin’s,

Redding noted one additional finding - benevolent leadership. He also observed that Chinese

societies have a strong legacy of personalism, a tendency to allow personal factors to be included in decision-making. Therefore, he suggested that the leader’s authoritarianism and benevolence may not be extended to all subordinates uniformly but to different subordinates in varying degrees.

(3)

Paternalistic leadership 91

© Blackwell Publishing Ltd with the Asian Association of Social Psychology Building on the work of Silin (1976) and Redding (1990), Westwood (1997) proposed a model of paternalistic headship for CFB and suggested that paternalistic leadership is manifest in a context characterized by centralization, low formalization, harmony building, and personalism via the following nine stylistic elements of paternalistic headship: (i) didactic leadership; (ii) non-specific intentions; (iii) reputation building; (iv) protection of dominance; (v) political manipulation; (vi) patronage and nepotism; (vii) conflict diffusion; (viii) aloofness and social distance; and (ix) dialogue ideal. Cheng (1995a, 1995b, 1995c, 1996; 1997) used a series of case studies, in-depth interviews, and quantitative data to explore leadership in Taiwan’s family businesses, high-tech industry, schools, and army, and confirmed that PL is widespread in various Chinese organizations.

But, what is paternalistic leadership? What is its main construct? After Farh and Cheng (2000) reviewed all the research conducted since Silin’s (1976), they defined PL as a style that combines strong discipline and authority with fatherly benevolence and moral integrity couched in a personalistic atmosphere. Based on this definition, PL consists of three important elements: authoritarianism, benevolence, and moral leadership. Authoritarianism refers to a leader’s behavior that asserts absolute authority and control over subordinates and demands unquestionable obedience from subordinates. Benevolence means that the leader’s behavior demonstrates individualized, holistic concern for subordinates’ personal or familial well-being. Moral leadership can be broadly depicted as a leader’s behavior that demonstrates superior personal virtues, self-discipline, and unselfishness.

Based on this definition, a survey tool has been developed for measuring the constructs of PL. According to Farh and Cheng’s (2000) conceptual analysis and past research, Cheng

et al. (2000) amended the dyadic leadership questionnaire created by Cheng (1996) and then

established a new three-dimensional PL scale. Moreover, they also explored the construct

validity of the three dimensions of PL -shi-en (granting favors), li-wei (inspiring awe or fear)

and shuh-der (setting a moral example). Using exploratory and confirmatory factor analyses,

the leadership styles of supervisors or principals as perceived by mid-level supervisors, employees, and teachers across various types of businesses and schools were found to consist of three subtypes: benevolent, moral, and authoritarian leadership. Benevolent leadership

contained shi-en behaviors (favor granting), such as ‘individualized care’ and ‘understanding

and forgiving’. Moral leadership entailed shuh-der (setting an example) behaviors, such as

‘integrity and fulfilling one’s obligations’, ‘never taking advantage of others’ and ‘selfless

paragon’. Authoritarian leadership comprises five types of li-wei (awe-inspiring) behaviors:

‘powerfully subduing’, ‘authority and control’, ‘intention hiding’, ‘rigorousness’ and ‘doctrine’. The PL scale had excellent construct validity as well as superior internal consistency.

A couple of studies have examined the cultural roots of PL. Farh and Cheng (2000) suggested that authoritarian leadership stems from the cultural tradition of Confucianism and Legalism. Under the influence of Confucianism, the father–son cardinal relationship is always considered paramount and supersedes all other social relations. In a father–son relationship under Confucian ethics, a father has authority over his children and all other family members

and possesses absolute power and legitimacy (Cheng et al., 2000). Meanwhile, the concepts

of control and political manipulation are highly valued in Legalist school and were emphasized and developed during imperial China. Consequently, authoritarian leadership was encouraged, and leaders maintained strong authority and subordinates were obligated to be obedient.

The cultural roots of benevolent leadership originate from the Confucian ideal of the five cardinal relationships and the norm of reciprocity. The five cardinal relationships are the

(4)

benevolent ruler with the loyal minister, the kind father with the filial son, the righteous husband with the submissive wife, the gentle elder brother with the obedient younger brother, and the kind elder with the deferent junior. Ideally, according to the five cardinal relationships, a father should be kind to his children and the children should show respect and filial piety to their father. Benevolence on the part of the superior generates indebtedness on the part of the inferiors, who try to reciprocate earnestly. This reciprocity may be driven by genuine gratitude, personal loyalty, or obedience to and compliance with the superior’s requests, even beyond what is normally required by the subordinate role (Yang L. S., 1957). Particularly in a system ruled by people than by laws and regulations, the government guarantees less protection of people’s rights, and the virtue of government officials or superiors is essential in order to protect the inferiors. Accordingly, in Chinese organizations, characterized by underdeveloped institutional structures and personalism, moral leadership is not only rooted in Confucianism, but is also expected by subordinates for their well-being.

Paternalistic leadership and Western transformational leadership theory

Paternalistic leadership, in indigenous leadership in Chinese societies, should demonstrate

certain unique features as compared to Western leadership concepts. Morris et al. (1999)

developed a framework utilizing both emic and etic approaches for theory development, and by comparing Western leadership concepts with PL, it may be possible to integrate indigenous and Western leadership concepts for the development of universal leadership models.

Traditional leadership theory tends to interpret the leader’s role as transactional - simply

to manage subordinates by clarifying their job roles and tasks to achieve goals. In transformational leadership, leaders not only direct subordinates to complete tasks, but also encourage them via individualized consideration, charisma, intellectual stimulation, and inspirational motivation to strive for accomplishments exceeding others’ expectations (Bass, 1985). In a comprehensive review of contemporary leadership theories in the West, transformational leadership was identified as the most highly valued (Yukl, 2002), and many researchers have explored the relationship between transformational leadership and subordinate performance (Bass & Avolio, 1990). In a real business environment, a focal concern is how to train and foster a transformational leader for the company (Bass, 1998). Transformational leadership in the West and paternalistic leadership in Chinese societies originated from distinct leadership theories bred in different cultural contexts with dissimilar fundamental assumptions about the rights and obligations of the superior and the inferiors. They each demonstrate a manifest leadership pattern and specific individual styles as perceived by subordinates. For instance, a paternalistic leader displays authority, control, and image building (Farh & Cheng, 2000), while a transformational leader shows personal charisma and intellectual inspiration (Bass, 1985). These two leadership styles do have something in common. Transformational leadership includes a style of individualized care that is similar to benevolent leadership. Some researchers (Parry & Proctor-Thomson, 2002) have recently started to explore the display of integrity by transformational leaders, which is similar to moral leadership.

Consistent with an old Chinese saying: ‘Similarity exists in dissimilarity and vise versa’, transformational leadership from the West and Chinese paternalistic leadership may contain general transnational behaviors that are applicable across cultures, as well as emic behaviors that are unique and are only applicable in a particular cultural setting (Yang K. S., 2000). Hence, it is essential to understand which are universal and which are culture specific when

(5)

© Blackwell Publishing Ltd with the Asian Association of Social Psychology transplanting a Western leadership model to Chinese business organizations. Those leadership behaviors with cultural generality will be more effective, while those reflecting indigenous Western behaviors, less effective. In other words, leadership behaviors conforming to Chinese cultural characteristics will be more effective, and those conflicting with Chinese cultural features, less effective.

The inference above is supported by Chen and Farh’s research (1999) examining the suitability of transformational leadership principles to businesses in Taiwan using local employees as participants. They found that in light of the relationship orientation of the Chinese cultural tradition, transformational leadership behaviors relating to relationship-oriented behavior (such as promoting cooperation, providing individual support, acting as an exemplar to subordinates) had a stronger positive effect on subordinates’ job performance and organizational citizenship behaviors. In contrast, the behaviors not related to relationship-oriented activities (such as vision building, intellectual stimulation) showed only a weaker or insignificant effect.

With regard to PL, although it is a leadership concept couched within the Chinese cultural tradition, it is highly relevant in modern organizations. All leaders’ in modern organizations have almost identical goals: to command, to organize, to influence, to decide, and to control, which are independent of cultural influence. Thus, some aspects of PL may be applicable only to social organizations with Chinese culture, while others may be universalistic and applicable to many other cultures. In concrete terms, the universalistic aspect of PL will result in an overlap with concepts in transformational leadership, and the emic aspect of PL will result in a significant and unique effect of PL on subordinate responses, even when the overlap of PL with Western leadership concepts are removed. Specifically, we examined in the present study whether PL has unique explanative power for subordinate responses while controlling for the effects of transformational leadership. Three dimensions were extracted from transformational leadership, including individualized consideration, high performance emphasis, and modeling, which can be viewed as corresponding to benevolent leadership, authoritarian leadership, and moral leadership, respectively. These three Western notions were used to remove the universalistic aspects of PL to reveal its emic effects in the present study.

Individualized consideration indicates that a leader respects subordinates, cares for them,

satisfies their individual feelings and needs, and gives them appropriate support. This is similar to the individualized care of benevolent leadership, although some differences remain. First, benevolent leadership is not confined to just being magnanimous on the job, but extends

to a subordinate’s personal issues. Second, the benevolent shi-en (favor granting) behaviors

to a subordinate are long-term oriented. Third, benevolent leadership includes behaviors of both forgiving and protection. Fourth, and most important of all, individualized consideration in the Western tradition is generally displayed in the context of equal treatment and equivalent

rights between the superior and inferior. In contrast, the shi-en behaviors of benevolent

leadership are enacted in a situation with a large difference in authority and power distance between superiors and inferiors. The leader constantly reminds the subordinates who is really in charge in his/her benevolent behaviors.

High performance standard denotes the expectation shown by a leader for a subordinate

to display a strong performance with excellent quality. These characteristics are similar to rigorousness and doctrine in authoritarian leadership. However, just as benevolent and moral leadership are based on large power distance, the legitimacy of authoritarian leadership is also derived more from the subordinate’s internalization of an inferior role than from accepting the superior’s assignment of job responsibilities. Consequently, an authoritarian

(6)

leader will try to control and promote a subordinate’s performance via personal preferences and threat.

Modeling in the Western tradition implies that the leader will stress a higher level of moral

values, such as equity and justice, and will lead by example (Burns, 1978; Bass, 1985, 1996;

Podsakoff et al., 1990; Podsakoff et al., 1996). Modeling is comparable to moral leadership.

For Western organizations, which have smaller power distance, researchers did not focus on the leader’s virtue until recently, as business ethics in Western societies have become more of a concern (Kanungo & Mendonca, 1996). Recent studies have proposed the concept of modeling. Modeling means the leader’s manifest behavior does not hinder the subordinates’ rights and development and harm the organization, but instead provides an example for subordinates (Parry & Proctor-Thomson, 2002). However, researchers of leadership behavior in Chinese organizations have long noted that leaders display this sort of moral behavior (Silin, 1976; Cheng & Chuang, 1981; Ling, 1991). Generally, modeling through transformational leadership stresses that a leader will lead by example to set up an integrity paradigm (Parry & Proctor-Thomson, 2002). In contrast, modeling through moral leadership places emphasis on unselfish behavior, including not abusing power for one’s own good, and not taking personal revenge in the name of public interest.

Based on the above logic, we argue that even with the effects of transformational leadership behaviors with cultural generality controlled for, PL will still have significant and specific explanation power for subordinate responses. Accordingly, our first hypothesis is:

H1: With the effect of Western transformational leadership controlled for, PL still has

significant and additional explanative power for subordinate responses.

The three dimensions of PL and subordinate responses

What are subordinate responses? In the Western literature on leadership, the leader’s influence is effective only if the leaders are able to involve subordinates in three ways: Alienative involvement (induced by fear or respect of the leader), moral involvement (due to recognition of the leader), and calculative involvement (based on the subordinate’s self-interest) (Etzoni, 1961). Based on this perspective, what is the real relationship between PL’s three dimensions (benevolent, moral, and authoritarian leadership) and a subordinate’s psychological response? The answer to this question was partly addressed in Cheng’s (1995a, 1995c) research on PL’s dyadic model. Using a clinical approach, Cheng observed and interviewed a large enterprise owner, and conducted interviews with 18 Taiwanese family enterprise owners and 24 first-level managers. Cheng not only confirmed the prevalence of PL in Chinese business

organizations, but also noted that in response to the leader’s shi-en behavior, which is

analogous to benevolent leadership, a subordinate will show gratitude and repayment for the

leader’s care. In response to the leader’s li-wei behavior, which is analogous to authoritarian

leadership, a subordinate will demonstrate obedience, compliance, and fear in responding to the leader’s requests.

Using the employees of state-owned enterprises in China as informants, Ling (1991) found that Chinese leaders revealed their moral characters and behavior as well as emphasized team performance and maintenance. This revelation of moral characters and behavior is similar to the display of unselfish behavior of the participants in Cheng and Chuang’s (1981) study of Taiwanese soldiers. Both the informants and the participants in these two studies put emphasis on the leader’s behavior of setting an example and being selfless. This aspect of leadership behavior is similar to the moral leadership concept in the present study. In response

(7)

© Blackwell Publishing Ltd with the Asian Association of Social Psychology of the leader and internalize the leader’s values, and their job performance will be affected by the leader (Ling, 1991). Farh and Cheng (2000) proposed a preliminary model, which consisted of benevolent, moral, and authoritarian leadership styles and delineated the relationships between these three leadership styles and subordinate responses (Figure 1). According to Farh and Cheng’s (2000) model, authoritarian leadership evokes a stronger response of dependence and compliance; benevolent leadership induces more gratitude and repayment; and moral leadership results in respect and identification.

From the above research, the influence of leadership on a follower’s psychological reactions is referred to as subordinate responses. These psychological responses to the leader are examined: identification and imitation, compliance without dissent, and gratitude and repayment. Compliance without dissent, similar to alienative involvement, is built on the role-based norm of fear and conformity to the leader’s legal authority. Gratitude and repayment, with a superficial resemblance of calculative involvement, are imperative behaviors related to morality in the context of Confucianism and Chinese moral norms. Furthermore, the form and content of gratitude and repayment are not necessarily comparable, but repayment takes needs into consideration as well as the well-being of the favor giver (Wen, 1995). Identification and imitation are similar to moral involvement. Subordinates are expected to completely accept the leader’s values and to learn and imitate these values as if they were their own (Cheng & Jiang, 2000). From this discussion, we have derived the following hypotheses:

H2: PL has positive effects on subordinate responses.

H2a: PL has a positive effect on subordinates’ identification with and imitation of the leader,

with moral leadership having the most significant effect.

H2b: PL has a positive effect on subordinates’ compliance with the leader, with authoritarian

leadership having the most significant effect.

H2c: PL has a positive effect on subordinates’ gratitude and repayment to the leader, with

benevolent leadership having the most significant effect.

Figure 1 Paternalistic leadership and subordinate responses (adapted from Farh, J. L. & Cheng, B. S. (2000) Leader behaviors Morality Authoritarianism Respect and identification Gratitude and repay Dependence and compliance Subordinate responses Benevolence

(8)

What are the effects of the interactions among benevolent, moral, and authoritarian leadership behaviors of PL on subordinate responses? Leadership styles employing favor and fear simultaneously were found to improve subordinates’ work attitude (Cheng, 1997). Specifically, that leadership effects on subordinate responses were strongest when high authority was combined with high benevolence, which suggests that there may be a positive interaction effect of authoritarian and benevolent leadership. Regarding the interactions between benevolent and moral leadership, and between moral and authoritarian leadership, there has been no relevant empirical research. We expect that benevolent leadership and moral leadership are likely to have independent and significant effects on subordinate responses because these two types of behavior seem unrelated. That is, in Chinese organizations characterized by asymmetric superior–inferior relations, a subordinate’s reciprocation to a leader’s benevolence will not be affected by whether or not the leader demonstrates moral leadership behavior. Likewise, a subordinate’s identification with leader’s moral leadership behavior will not be affected by whether or not the leader grants favors. In contrast, moral and authoritarian leadership may have a negative interaction effect on subordinate responses. In traditional China, when righteous, incorruptible, and impartial government officials implemented their policies with an authoritarian approach, they would issue severe penalties to violators, no matter what the excuse was or how familiar the violator was with them. Consequently, civilians might form an inhumane and cruel impression of these government officials, and become alienated from them (Pye, 1985). Thus, if a leader exhibits moral behaviors as well as imposes strict discipline upon subordinates, the punitive behaviors will dominate the subordinate responses, resulting in disaffection and a reduction in identification, compliance, and gratitude.

The three-way interaction effect of PL is very complicated and not easy to interpret. We will not attempt to propose a hypothesis in the absence of previous research. The proposed two-way interactions are summarized in the following hypotheses:

H3: Authoritarian, benevolent, and moral leadership have a significant interaction effect with

respect to subordinate responses.

H3a: Benevolent and authoritarian leadership have a significant, independent positive effect

on subordinate responses.

H3b: Authoritarian and moral leadership have a significant negative effect on subordinate

responses.

Moderating effect of authority orientation of subordinate’s traditionality

In traditional Chinese societies, rights were unequal between superiors and subordinates, and superiors demanded compliance, respect, and trust in authority from subordinates. This pattern from traditional Chinese culture has been influenced by Western culture and industrialization in the 20th century (Yang K. S., 1998). King (1996) indicated that the Chinese in Hong Kong do not accept Confucianism without criticism, but instead view it with a rational and instrumental attitude. They accept the traditional values that help them adapt to modern life, but discard traditional values that are unsuitable. Cheng and Farh’s (2001) research points to a similar conclusion. They surveyed employees of Taiwanese and Chinese business organizations and found that despite relatively lower scores in authoritarianism, employees both in Taiwan and China had high scores in granting favors and face, instrumental relationships, and familism. Results indicated that some cultural values were preserved, but an authoritarian orientation had faded, as it may be unsuitable to modern trends. K. S. Yang

(9)

© Blackwell Publishing Ltd with the Asian Association of Social Psychology (1996) also reached a similar conclusion about authoritarian orientation. In a study of Taiwanese, he found that compliance to authority is unable to coexist with modern values, such as egalitarianism or accepting new thoughts, and is becoming weaker gradually.

Chinese people’s attitudes towards filial piety have changed over the past century. Filial piety is fading away, and is no longer followed without question (Ho & Chiu, 1994; Ho, 1996). Although the main obligations, such as offering sacrifices to ancestors and reciprocating parents’ care from childhood are still common, absolute compliance and obedience to parents have largely disappeared (Ho, 1996). In addition, C. F. Yang (1988) found that in families from China, Taiwan, and Hong Kong, the family power structure has transformed from a focus on the father–son relationship to the husband–wife relationship. Although the parents are still respected, they have lost their full authority and absolute power over their children.

Modernization in contemporary Chinese societies has to some degree weakened the foundation of authoritarianism. Compliance to authority may no longer be a common value to all Chinese, especially the younger generation with higher education. In fact, recent studies of Hong Kong and Taiwan suggest that individual deviation in attitude towards authority compliance is a moderator variable for relationships between subordinates’ attitudes and

organizational behavior (Farh et al., 1997; Farh et al., 1998; Pillutla et al., 1998).

The above review suggests that different subordinates with different degrees of authority orientation may respond to PL differently. Specifically, the authority orientation of a subordinate’s traditionality may show a moderating effect on the relationship between PL and subordinate responses. For subordinates with a weaker authority orientation, the effects of authoritarian leadership may be weaker as well. In contrast, modernization has not altered the emphasis placed on leaders’ integrity and employee well-being, and moral and benevolent leadership should have the same effects as in a traditional society. That is, authority orientation should not impact on the effects of these leadership dimensions. Our reasoning is represented in the following hypothesis:

H4: The effect of authoritarian leadership on subordinate responses will differ according to

the level of the subordinate’s authority orientation. The effects of moral and benevolent leadership are not affected by authority orientation.

Methods

Sample

We tested our hypotheses by surveying employees who had worked together with their supervisors for more than half a year. Our sample consisted of 605 low- to mid-level managers

and staff from 60 Taiwanese enterprises (the return rate was 605/700 = 86.43%). After

excluding invalid questionnaires and employees who had worked with their leaders for less than half a year, we had complete data for 543 employees. Table 1 summarizes the demographic profile of the employee sample.

Procedure

Participants in this study were from two sources: business employees participating in an executive MBA program or participants sampled directly from business organizations. The former data were collected by group surveys. For the latter group, the personnel managers of

(10)

the businesses were entrusted to identify suitable employees as respondents. The completed questionnaires could either be given back to the personnel managers or be sent back to the researchers directly. Researchers delivered detailed instructions to the questionnaire administrators in either written or verbal forms in advance to ensure the quality of data collection. After collecting all the completed questionnaires, researchers excluded invalid questionnaires (those with too many missing responses or obvious response biases) before proceeding with statistical analyses.

Measures

The independent variables were paternalistic leadership and the corresponding Western leadership constructs.

Table 1 Demographic profiles of the sample

Variable Category n Percentage

Gender Male 295 54.30

Female 235 43.31

Missing data 13 2.39

Age Under 25 years old 59 10.90

26–30 years old 184 33.90

31–35 years old 176 32.33

36–40 years old 71 13.00

41–45 years old 36 6.67

46–50 years old 10 1.93

Over 51 years old 5 0.87

Missing data 2 0.40

Education Below junior high school 5 0.83

Senior high school 69 12.67

Bachelor degree 186 34.29

Graduate degree 243 44.70

Missing data 40 7.51

Tenure Less than 1 year 55 10.10

1–3 years 187 34.43 3–5 years 100 18.50 5–7 years 72 13.20 7–9 years 33 6.00 9–11 years 42 7.80 11–13 years 22 4.10

More than 13 years 29 5.30

Missing data 3 0.57

Position Staff 397 73.20

Low-level manager 110 20.20

Mid-level manager 31 5.70

(11)

Paternalistic leadership. Paternalistic leadership has three distinct elements: authoritarian

leadership, benevolent leadership and moral leadership. Each type was measured with 11 items using a six-point Likert scale ranging from 1 (strongly disagree) to 6 (strongly agree).

The scale was developed by Cheng et al. (2000).1_{Participants rated their supervisors’}_shi-en_,

li-wei and shuh-der behaviors. The internal reliability coefficient was 0.94 for the benevolent

leadership scale, 0.90 for the moral leadership scale, and 0.89 for the authoritarian leadership scale.

Corresponding Western leadership scales. Parallel with paternalistic leadership, the

corresponding Western leadership scales included individualized consideration, modeling,

and performance-emphasizing leadership. The six-factor scale2_{was translated and modified}

from Podsakoff et al., 1990), and this scale also used a six-point Likert scale (1 = strongly

disagree and 6 = strongly agree). The internal reliability coefficient was 0.87 for the four

items of the individualized consideration scale, 0.97 for the five items of the modeling leadership scale, and 0.75 for the four items of the performance-emphasizing leadership scale.

The dependent variables were three types of subordinate responses: response-respect and identification; indebtedness and obligation to repay; and dependence and compliance; which reflected the degree of the psychological influence of leadership on subordinates. Based on the definition provided by Farh and Cheng (2000), we modified Cheng’s (1995b) subordinate

responses items and Cheng et al.’s (1999) loyalty to supervisor scale to obtain these three

types of subordinate responses.3

Identification and imitation refers to respecting and identifying with the supervisor’s behaviors and values, and being willing to imitate them. Seven items were used, and the internal reliability coefficient was 0.93. Compliance without dissent is concerned with whether or not subordinates obey and follow the supervisor’s command completely, and the internal reliability coefficient for the five items used was 0.81. Gratitude and repayment is concerned with whether or not the subordinate is influenced by the supervisor’s benevolence and is willing to strive to repay it, even if it requires self-sacrifice. Six items were used, and the internal reliability coefficient was 0.88. All the items in these scales were based on six-point, strongly agree–strongly disagree Likert scales.

The moderating variable was authority orientation, which was adopted from the respect

authority subscale of the traditionality scale4_{of K. S. Yang}_{et al.}_{(1989). The internal}

reliability coefficient for the 11 items used was 0.84, and the items were also based on six-point, strongly agree–strongly disagree Likert scales.

We included a number of control variables: gender, age, education, tenure and position. They were included because they are proxies of human capital and power and may influence

a person’s reactions to a leader. Gender was coded with 0 = males, and 1 = females. Age,

ranging from under 25 to over 50, was classified into seven categories. Education level was

measured by four categories (1 = below junior high school, 2 = Senior high school,

3 = bachelor degree, and 4 = graduate school). Tenure was coded with eight categories

(1 = less than 1 year, 2 = 1–3 years, 3 = 3–5 years, 4 = 5–7 years, 5 = 7–9 years, 6 = 9–

11 years, 7 = 11–13 years, and 8 = more than 13 years). Position was measured by three

categories (1 = staff, 2 = low-level manager, 3 = mid-level manager).

We used an even number of points (i.e. six points, on the Likert-like scale rather than an odd number, because Chinese people tend to answer in the middle (Chiu & Yang, 1987). We hoped to prevent this response bias by not including a mid-point on the scale.

(12)

Analyses

Several statistical procedures were adopted to examine the hypotheses. First, correlation analysis was conducted to understand the relationships among the variables, and to provide explanations for the results of the regression analysis. Second, for Hypothesis 1 and

Hypothesis 2 (M1~M3), block regression analysis was used to compare the difference in

subordinate responses to Western leadership as compared to paternalistic leadership. Partial

R2_and

b were used to explore the additional explanatory power of paternalistic leadership

after controlling for the impact of Western leadership.

Third, hierarchical regression analysis was used to test Hypothesis 3 (M4, M5), which are concerned with the main effects and interactions on subordinate responses to leadership by benevolence, morality, and authoritarianism. If Hypothesis 3 was supported, researchers would trisect subordinates into high, middle, and low groups according to the leader’s benevolence, morality, and authoritarian leadership score, respectively. Based on the categorization, the high and low groups were selected and formed four groups by crossing two dimensions: high-high, high-low, low-high, and low-low. Group means of the three responses were calculated and diagramed to clarify the tendency of the interaction of paternalistic leadership in affecting subordinate responses.

Finally, hierarchical regression analysis was used to test Hypothesis 4 (M6, M7), which is about the mediating effect of authority orientation on the relationship between authoritarian leadership and subordinate responses. Then, high and low groups were formed according to their authority orientation for block regression analyses to understand the nature of the moderating effect of authority orientation (M8, M9). Steps for the analysis of each hypothesis are shown in Table 2.

Table 2 Plan for regression analysis

Hypothesis Method of analysis Model

Block or sequence of variables

1 2 3 H1 & H2 BRM M1 CV WL M2 CV PL M3 CV WL PL H3 HRM M4 CV WL PL 2-way interactions of PL M5 CV WL PL 2-way interactions of PL 3-way interaction of PL H4 HRM M6 CV WL PL Authority orientation M7 CV WL PL Authority orientation AO * PL BRM

for low and high authority orientation groups

M8 CV PL

M9 CV WL PL

AO, authority orientation; BRM, block regression analysis; CV, control variables; PL, paternalistic leadership; HRM, hierarchical regression analysis; WL, corresponding Western leadership.

(13)

Results

Correlations between all the variables

The means, standard deviations, reliabilities, and intercorrelations of all of the variables used in the study are reported in Table 3. At the zero-order correlation level, subordinates’

gender and education were negatively correlated with authority orientation (r = - 0.11,

p < 0.05; r = - 0.18, p < 0.001), and organizational tenure was positively correlated with

authority orientation (r = 0.11, p < 0.05). In addition, there was a negative correlation

between subordinates’ gender and gratitude to the leader (r = - 0.14, p < 0.01), and a

positive correlation between age and compliance (r = 0.11, p < 0.05), and position and

gratitude (r = 0.10, p < 0.05). All the subordinate responses to the leader (identification,

compliance, and gratitude) were significantly and positively correlated with organizational tenure.

Intercorrelations between each pair of leadership behaviors were positive except for the correlations with authoritarian leadership. Furthermore, the subordinate’s authority orientation of traditionality was positively correlated with benevolent and authoritarian PL leadership traits, as well as modeling and performance-emphasizing leadership traits of corresponding Western leadership. The correlation between authority orientation and moral leadership was negative. All the intercorrelations between the variables of the subordinate responses were positive. The subordinate’s authority orientation was positively correlated with all of the subordinates’ responses to the leader (identification, compliance, and gratitude). Finally, the correlations between leadership behavior and subordinate responses were significantly positive. These correlations permit examination of the unique effects of each leadership variable on subordinate responses while controlling for other variables.

Comparison of PL and corresponding Western leadership

Table 4 lists the block regression results for testing PL and corresponding Western leadership’s relative effects on subordinate responses. Demographic variables had little relationship to subordinate responses. After controlling for demographic variables, M1 and M2 show that both corresponding Western leadership and PL produced a significant

and sizable correlation with the three subordinate responses (DR2

= 0.62, 0.27, and 0.40,

p < 0.001; DR2

= 0.56, 0.27, and 0.45, p < 0.001). Benevolent, moral, and authoritarian leadership styles of PL, and the corresponding individualized consideration, modeling, and performance-emphasizing leadership styles all related to identification, gratitude, and compliance.

Next we entered PL and corresponding Western leadership together into the regression equation (M3). As hypothesized, when controlling for the effects of corresponding Western leadership, M3 indicated that PL still had significant partial R2 with all subordinate responses

(DR2

= 0.03, 0.04, and 0.08, p < 0.001). Thus, the findings support H1: With the effect of

Western transformational leadership excluded, PL still has significant and additional explanative power for subordinate responses.

There were different patterns of impact on the three types of subordinate responses from (i) PL and corresponding Western leadership; and (ii) benevolent, moral, and authoritarian

leaderships. First, for identification, corresponding Western leadership (DR2

= 0.09,

p < 0.001) explained more variance than PL (DR2

= 0.03, p < 0.001). For compliance, both

were the same. However, for gratitude, PL (DR2

(14)

102 Bor-Shiuan Cheng et al.

T

able 3

Means, standard deviations, and correlations (

N = 514–525) V ariable Mean SD 1234 5 6 789 1 0 11 12 13 14 15 Control v ariable 1. Gender 0.45 0.50 2. Age 3.79 1.21 - 0.32*** 3. Education 3.45 0.83 0.03 - 0.14*** 4. Position 1.24 0.69 0.00 0.40*** 0.00 5. T enure 3.37 1.92 - 0.14** 0.57*** - 0.26*** 0.28*** Moderator v ariable 6. A O T 3.20 0.69 - 0.11* 0.09 - 0.18*** 0.04 0.11* (0.84) P aternalistic leadership 7. Bene v olent 3.84 0.86 - 0.09* 0.02 - 0.04 0.01 0.07 0.11* (0.94) 8. Moral 4.48 0.85 0.05 0.00 - 0.03 0.00 0.04 - 0.13** 0.61*** (0.90) 9. Authoritarian 3.18 0.83 - 0.18*** 0.09* - 0.06 0.07 0.03 0.27*** - 0.35*** - 0.54*** (0.89) Corresponding W estern leadership 10. IC 4.00 0.94 0.00 0.00 0.01 0.04 0.07 - 0.03 0.72*** 0.67*** - 0.48*** (0.87) 11. Modeling 4.23 1.12 - 0.09 0.01 - 0.07 - 0.01 0.07 0.09* 0.67*** 0.68*** - 0.27*** 0.63*** (0.97) 12. PE 4.12 0.82 - 0.15*** 0.04 - 0.05 0.02 0.10* 0.09* 0.58*** 0.52*** - 0.17*** 0.50*** 0.64*** (0.75) 13. II 3.98 0.86 - 0.05 0.06 - 0.07 0.02 0.09* 0.21*** 0.70*** 0.64*** - 0.28*** 0.66*** 0.73*** 0.61*** (0.93) 14. CWD 4.12 0.66 - 0.08 0.11* 0.08 0.04 0.14** 0.30*** 0.42*** 0.41*** - 0.06 0.43*** 0.44*** 0.43*** 0.71*** (0.81) 15. GR 3.77 0.81 - 0.14** 0.08 0.00 0.10* 0.09* 0.25*** 0.65*** 0.48*** - 0.17*** 0.55*** 0.57*** 0.51*** 0.79*** 0.68*** (0.88) * p < 0.05, ** p < 0.01, *** p < 0.001.

Numbers in parentheses are the Cronbach

a of the scales. Gender , 0 = male and 1 = female. AO T,

authority orientation of traditionality; CWD, compliance without dissent; GR, gratitude and repayment; IC, indi

vidualized

(15)

T

able 4

Regression analysis for unique and common effects of paternalistic leadership and corresponding W

estern leadership scales

on subordinate response ( N = 503) V ariables Identification Compliance Gratitude M1 M2 M3 M4 M5 M1 M2 M3 M4 M5 M1 M2 M3 M4 M5 Control v ariables Gender – – – – – – – – – – - 0.09* - 0.09* – – – Age – – – – –– – – – – – ––– – Education – – – – – – – – – – – – – – – Position – – – – – – – – – – - 0.09* 0.09* 0.09* 0.08* 0.08* T enure – – – – – – – – – – – – – – – D R 2 –– –– –– 0.02* 0.02* 0.01* Corresponding W estern leadership Consideration 0.31*** 0.18*** 0.17*** 0.17** 0.22*** 0.20*** 0.19** 0.19** 0.29*** 0.11* 0.10* 0.10* Modeling 0.40*** 0.29*** 0.26*** 0.26*** 0.18** – – – 0.28*** 0.14** 0.11* 0.11* PE 0.21*** 0.15** 0.17*** 0.17*** 0.22*** 0.17*** 0.17*** 0.18*** 0.18*** 0.10* 0.11* 0.11* D R 2 0.62*** 0.09*** 0.27*** 0.04*** 0.40*** 0.03** P

aternalistic leadership Bene

v olent 0.49*** 0.23*** 0.21** 0.21*** 0.29*** – – – 0.57*** 0.42*** 0.40*** 0.40*** Moral 0.39*** 0.14** 0.17*** 0.17*** 0.37*** 0.23*** 0.28*** 0.28*** 0.21*** – 0.12* 0.12* Authoritarian 0.09** – – – 0.21*** 0.21*** 0.23*** 0.21*** 0.10** 0.09* 0.10* 0.10* D R 2 0.56*** 0.03*** 0.27*** 0.04*** 0.45*** 0.08*** Tw o-w ay interaction BE*MO – – 0.10* – – – BE*A U 0.16*** 0.16** 0.16** 0.16** 0.19*** 0.19*** MO*A U - 0.13** - 0.13** - 0.19** - 0.18** - 0.17*** - 0.17** D R 2 0.02*** 0.03*** 0.02*** Three-w ay interaction D R 2 –– – Ov erall R 2 0.63 0.57 0.66 0.68 0.68 0.29 0.28 0.33 0.36 0.36 43 48 51 53 53 F -v alue 104.95 83.09 86.04 71.30 66.42 26.17 25.69 22.48 19.89 15.58 47.41 57.79 46.08 38.62 35.98 –, Not significant. * p < 0.05, ** p < 0.01, *** p < 0.001. Gender , 0 = male and 1 = female. All F -v

alues are significant (

p < 0.001). D R 2 is the unique v ariance attrib utable to a set of v

ariables independent of all other sets in a gi

v

en equation.

A

U, authoritarian; BE, bene

v

(16)

than corresponding Western leadership (DR2

= 0.03, p < 0.001). Moreover, with

corresponding Western leadership controlled for, benevolent and moral leadership had a

positive impact on subordinate’s identification (b = 0.23, p < 0.001; b = 0.14, p < 0.01). Moral

and authoritarian leadership had a significant positive effect on compliance (b = 0.23 and

0.21, p < 0.001). Benevolent and authoritarian leadership had a positive influence on gratitude

(b = 0.42, p < 0.001; b = 0.09, p < 0.05).

Main effects of benevolent, moral, authoritarian leadership on subordinate responses

According to the regression coefficients of M2 in Table 4, all the benevolent, moral, and authoritarian leadership aspects of PL had a notably positive effect on subordinate responses.

H2 that PL has positive effects on subordinate responses is supported. In other words, the

leadership behaviors leaders show to their subordinates, including shi-en, shuh-der, and

li-wei, are conducive to subordinate identification, compliance, and gratitude.

Yet, surprisingly, no matter whether we controlled for corresponding Western leadership or not, M2 and M3 of Table 4 show that the three leadership aspects of PL had different significance for subordinate responses. Benevolent leadership had the strongest effect on

identification (b = 0.49 and 0.23, p < 0.001). Next was moral leadership. Regarding

compliance with a leader, the biggest effect came from moral leadership (b = 0.37 and 0.23,

p < 0.001). The type of leadership that is most conducive to subordinate gratitude, as

hypothesized, was benevolent leadership (b = 0.57 and 0.42, p < 0.001). These findings

support H2c in that PL has a positive effect on subordinate gratitude and repayment to the

leader, with benevolent leadership showing the most significant effect. However, H2a and H2b

were not completely supported.

Interaction of benevolent, moral, authoritarian leadership on subordinate responses

The interactive effects of benevolent, moral, and authoritarian leadership are showed in M4 and M5 of Table 4. By examining the change in R2 attributable to the interaction of the three

elements of PL, two-way interactions were found (DR2

= 0.02, 0.03, 0.02, p < 0.001), but not the three-way interaction. Except for the interaction of benevolent and moral leadership on subordinate responses, the other two-way interactions (benevolent-authoritarian and moral-authoritarian leadership) had a clear pattern. As shown, the interaction between benevolent

and authoritarian leadership had a positive beta coefficient (b = 0.16, p < 0.001; b = 0.16,

p < 0.01; b = 0.19, p < 0.001), and the interaction between moral and authoritarian leadership

had a negative beta coefficient of (b = - 0.13, p < 0.01; b = - 0.19, p < 0.01; b = - 17,

p < 0.001).

In sum, the results for interactions are as follows: H3, that authoritarian, benevolent, and

moral leadership have a significant interaction effect with respect to subordinate responses,

was supported. In addition, the hierarchical regression analysis supported H3a, that benevolent

and authoritarian leadership have a significant positive effect on subordinate responses, and

H3b, that authoritarian and moral leadership have a significant negative effect on subordinate

responses. Finally, there is no significant three-way interaction effect between benevolent, moral, and authoritarian leadership.

Figure 2 depicts the positive interaction effects on subordinate responses. The relationship between benevolent leadership and subordinate responses was stronger for high than for low

(17)

© Blackwell Publishing Ltd with the Asian Association of Social Psychology authoritarian leadership. Figure 3 presents the negative interaction effects for moral and authoritarian leadership. The relationship between moral leadership and subordinate responses was stronger for low than for high authoritarian leadership.

The moderating effects of the traditional authority orientation

The regression results for H4 are summarized in M3, M6, and M7 of Table 5. Based on M6,

subordinate’s authority orientation had a significant positive effect on subordinate responses

(b = 0.18, 0.29, and 0.21, p < 0.001; DR2_{= 0.03, 0.07, and 0.04, p < 0.001). As expected,}

Figure 2 Benevolent leadership and subordinate response by authoritarian leadership. , low-authoritarian; , high-authoritarian 5.0 4.5 4.0 3.5 3.0 2.5 Identification Compliance Gratitude 5.0 4.5 4.0 3.5 3.0 2.5 5.0 4.5 4.0 3.5 3.0 2.5 3.64 3.11 4.61 4.63 4.00 3.83 4.65 4.43 3.30 3.00 4.47 4.32 low high Benevolent low high Benevolent low high Benevolent

(18)

subordinates exhibited more identification, compliance, and gratitude to leaders with higher authority orientation. M7 reveals further that a subordinate’s authority orientation had a

moderating effect on paternalistic leadership on subordinate responses (DR2_{= 0.01, p < 0.01}

and DR2_{= 0.01, p < 0.05). Furthermore, authority orientation did not moderate the effects of}

benevolent and moral leadership, but showed a positive moderating effect on the effects of authoritarian leadership (b = 0.09, 0.12, and 0.11, p < 0.01).

To clarify the moderating effects of authority orientation, we divided the sample into high and low authority orientation groups, based on a median split, and performed separate block

Figure 3 Moral leadership and subordinate response by authoritarian leadership. , low-authoritarian; , high-authoritarian 5.0 4.5 4.0 3.5 3.0 2.5 Identification 3.41 2.94 4.38 4.59 low high Moral 5.0 4.5 4.0 3.5 3.0 Compliance Gratirude 2.5 3.95 3.69 4.34 4.47 low high Moral 5.0 4.5 4.0 3.5 3.0 2.5 3.43 2.83 4.20 4.08 low high Moral

(19)

T

able 5

Hierarchical regression analysis for moderating effects of authority orientation (

N = 504) V ariables Identification Compliance Gratitude M3 M6 M7 M3 M6 M7 M3 M6 M7 Control v ariables Gender –––––––– – Age –––––––– – Education –––––––– – Position –––––– 0.9* 0.8* 0.8* T enure –––––––– – Corresponding W estern leadership Consideration 0.18*** 0.18** 0.18** 0.20*** 0.20*** 0.21*** 0.11* 0.11* 0.11* Modeling 0.29*** 0.26*** 0.26*** – – – 0.14** 0.11* 0.11* PE 0.15** 0.15** 0.15** 0.17*** 0.16** 0.17** 0.10* 0.9* 0.10* P

aternalistic leadership Bene

v olent 0.23*** 0.19** 0.16** – – – 0.42*** 0.38*** 0.34*** Moral 0.14** 0.18** 0.20** 0.23*** 0.30*** 0.31*** – 0.13* 0.14** Authoritarian – – – 0.21*** 0.15** 0.16** 0.09* 0.09* 0.09** Authority orientation 0.18*** 0.17** 0.29*** 0.28*** 0.21*** 0.20*** D R 2 0.03*** 0.07*** 0.04*** Moderating ef fects of authority BE* authority – – – MO * authority – – – AU * authority 0.09** 0.12** 0.11** D R 2 0.01** 0.01* 0.01* Ov erall R 2 0.65 0.69 0.70 0.34 0.41 0.42 0.51 0.54 0.55 F -v alue 85.81 88.83 73.11 22.41 27.82 23.08 45.74 48.41 40.03 –, Not significant. * p < 0.05, ** p < 0.01, *** p < 0.001. Gender , 0 = male and 1 = female. All F -v

v

en equation.

A

U, authoritarian; BE, bene

v

(20)

regression for each group (median = 3.18 and SD = 0.69 in the overall sample. The results of

the F-test for the differences in the variance of three dependent variables are: identification,

F = 2.01 and p = 0.16; compliance, F = 0.01 and p = 0.93; gratitude, F = 3.20 and p = 0.07.

The variance of two groups was homogeneous. The results of subgroup regressions are shown in Table 6 (M8, M9).

As shown in Table 6, the mechanism of how PL induces subordinate responses (identification, compliance, and gratitude) also differed according to the degree of authority orientation. First of all, authoritarian leadership shows no effect on subordinates with low

authority orientation, but shows a strong positive effect (P < 0.1) on subordinates with high

authority orientation for all three subordinate responses. Second, regardless of authority orientation, benevolent and moral leadership styles both show a notable influence on subordinate responses, especially moral leadership. We used Clogg et al.’s (1995) formula to test the difference between the beta coefficients:

as shown in Table 7. Only the regression coefficients of benevolent leadership for identification under M9 were significantly different. But, all such tests for authoritarian

leadership show a significant difference for the two groups. Consequently, H4 (the effect of

authoritarian leadership on subordinate responses will differ according to the level of the subordinate’s authority orientation. The effects of moral and benevolent leadership are not affected by authority orientation) was confirmed.

The variance accounted for by PL on subordinate’s identification and gratitude is larger for low than for high authority orientation groups. These findings suggest that PL in modern business organizations continues to play a major role, but its primary effects come from benevolent and moral leadership.

Discussion

This study probed the relationships between PL and subordinate responses to establish a preliminary model of effective PL behavior. First, the present study showed that PL has extra power in explaining subordinate reactions after controlling for corresponding Western concepts of transformational leadership. This finding suggests that PL is, indeed, different from Western leadership and manifests some indigenous characteristics. Furthermore, as was illustrated in Cheng’s (1996) study of subordinate loyalty in the East and West, role obligations are emphasized in the loyalty responses of Chinese subordinates, whereas personal affection is stressed by the loyalty responses of Western subordinates. Following this logic, when comparing subordinate responses to PL versus Western transformational leadership, affective commitment may be more influenced by Western leadership, whereas normative commitment may be more influenced by PL.

Second, we found that all three elements of paternalistic leadership have positive effects on subordinate responses. Specifically, benevolent leadership has the most significant effect on subordinate gratitude and repayment to the leader. Benevolent leadership has the most powerful effect on identification and imitation. Surprisingly, for subordinate compliance, it is moral leadership that has the largest effect, not authoritarian leadership as might be expected. The results seem to imply that in modern Chinese organizations, because the relational orientation is still salient (Cheng & Farh, 2001), the superior’s shi-en behaviors, such as personal care, forgiving and consideration towards subordinates, not only evoke

z=

(

bˆg1-bˆg2

)

[

s2

( )

bˆg1 +s2

( )

bˆg2

]

1 2

(21)

T

able 6

Block regression analysis for high and low authority orientation groups

V ariables Identification Compliance Gratitude High ( N = 250) Lo w (N = 252) High (N = 250) Lo w (N = 252) High (N = 250) Lo w (N = 252) M8 M9 M8 M9 M8 M9 M8 M9 M8 M9 M8 M9 Control v ariables Gender ––––––– – – – - 0.10* – Age ––––––– – – – – – Education ––––––– – – – – – Position ––––––– – – – 0.11* 0.11* T enure –––––– 0.16* – – – – – D R 2 ––––––– – – – 0.03* 0.02* Corresponding W estern leadership Consideration 0.25*** 0.15* 0.30*** – 0.16* – Modeling 0.26*** 0.29*** – – – – PE 0.23*** – 0.24*** – 0.15* – D R 2 0.12*** 0.06*** 0.09*** – 0.04** – P

aternalistic Leadership Bene

v olent 0.40*** – 0.49*** 0.25*** 0.21** – 0.25*** – 0.50*** 0.32*** 0.52*** 0.40*** Moral 0.44*** 0.16** 0.41*** 0.20*** 0.47*** 0.30*** 0.36*** 0.25** 0.24*** – 0.26*** 0.15* Authoritarian 0.14** 0.13** – – 0.23*** 0.25*** – – 0.15** 0.15** – – D R 2 0.50*** 0.02** 0.65*** 0.05*** 0.31*** 0.06*** 0.25*** 0.03** 0.40*** 0.07*** 0.52*** 0.08*** Adjusted R 2 0.49 0.62 0.65 0.71 0.30 0.38 0.27 0.28 0.41 0.45 0.54 0.55 F -v alue 31.15 37.17 59.81 56.68 14.40 14.99 12.66 9.84 22.90 19.44 37.32 28.43 –, Not significant. * p < 0.05, ** p < 0.01, *** p < 0.001. Gender , 0 = male and 1 = female. All F -v

v

en equation.

High, high authority orientation group; Lo

w

, lo

(22)

subordinate gratitude and repayment, but also lead to the perception of benevolent behavior as a virtue and to the display of affection, dependence and identification toward the superior. Along with the evolution of modern concepts, such as equalitarianism, moral leadership seems to have replaced some dated elements of authoritarian leadership. It is now more probable for moral rather than authoritarian leadership to induce subordinate compliance through acts of unselfishness and exemplary behaviors.

The analysis also suggests that authoritarian leadership influences subordinate responses only for authority-oriented subordinates, and that its influence is not as strong as that of moral and benevolent leadership. Apparently, modernization has weakened the influence of authoritarian leadership, supporting the view of leadership functionalism (Bons, 1984): the leadership style that conforms to the demands of the era is more effective than one that does not.

The present study also confirmed that leadership styles with both en (favor) and wei (might) have a positive and strong effect on subordinate responses. In fact, the favorability of subordinate responses show the following ranking: ‘equivalent en and wei are used’ > ‘more en is used than wei’ > ‘neither of en nor wei is used’ > ‘more wei is used than

en’. This result is similar to the classic study of the two-element theory of effective leadership

conducted by researchers at Ohio State University. They showed that consideration and initiative structure as two important dimensions of leadership have a significant, positive interaction that impacts subordinate responses (Fleishman, 1953; Blake & Mouton, 1982). Indeed, our result is also consistent with a previous study (Cheng, 1996) showing that, for Chinese subordinates, leadership with both en (favor) and wei (might) is the most effective way to induce positive subordinate responses to the superior. Interestingly, we also found that the leadership with both der (morality) and wei (might) showed a negative effect on subordinate responses. The effects on subordinate responses can be ranked as follows: ‘high

der (morality) and low wei (might)’ > ‘equivalently high der (moral) and high wei

Table 7 Tests of regression coefficient for high and low authority orientation groups

Variable

Identification Compliance Gratitude

M8 M9 M6 M7 M6 M7 Benevolent D 0.118 0.226 0.031 0.095 0.078 0.131 s(D) 0.066 0.074 0.065 0.079 0.069 0.084 z-value 1.79 3.05** 0.48 1.20 1.13 1.56 Moral D 0.027 0.005 0.058 0.021 0.043 0.058 s(D) 0.071 0.074 0.069 0.079 0.074 0.085 z-value 0.38 0.07 0.84 0.27 0.58 0.68 Authoritarian D 0.114 0.133 0.137 0.153 0.122 0.134 s(D) 0.060 0.057 0.060 0.061 0.062 0.065 z-value 1.90† _2.33* _2.28* _2.51* _1.97* _2.06* *p < 0.05 (Z > 1.96), **p < 0.01 (Z > 2.58), two-tailed. †_z-value_{= 1.90, p 0.056.}

D = |bHigh authority orientation - blow authority orientation|.

s(D) = [s2_(b

High authority orientation) + s2(blow authority orientation)]1/2.

z-value = D/s(D).

._. =

(23)

(might)’ > ‘low der (morality) and high wei (might)’ > ‘low der (morality) and low wei

(might)’. This ranking suggests that for a leader with high morality, authority has a negative effect on subordinate responses. For a leader with low morality, authoritarian leadership still has some positive effect.

Finally, for subordinates with high authority orientation and who value compliance and obligations, authoritarian leadership shows a larger effect. For subordinates who stress equal rights, the influence of authoritarian leadership is low or even non-existent. Nevertheless, the effects of both moral and benevolent leadership on subordinate responses are not moderated by authority orientation. This result shows, again, that, along with modernization, industrialization, and globalization, the influence of authoritarian leadership of PL is disappearing, as more and more Chinese are giving up authority orientation. In contrast, moral and benevolent leadership is alive and well, and may even become more important in modern organizations.

Limitations and future research

One limitation of the present study is that subordinate responses may not be equivalent to subordinate work attitudes or efficiency. We have not examined whether subordinate responses affect work attitudes (such as job satisfaction) and efficiency (such as organizational citizenship behavior, job performance, absence/leave behavior). It is necessary to examine how PL affects work attitude and efficiency in future research. Another issue is that the correlations among the three subordinate responses are high. Based on our conceptual analysis, ‘compliance’, ‘gratitude and repayment’ and ‘internalization of identification should be three distinct responses. In future research, it is necessary to refine these three scales to enhance their distinctiveness.

As to the three dimensions of transformational leadership adopted in this study, more items may be needed to capture individualized consideration, modelling, or performance, which should be undertaken in future research. Furthermore, the PL data and the subordinate responses are based on the respondents, which might lead to the same-source bias or common method variance. However, some researchers have claimed that the same-source bias does not necessarily have a serious impact (Spector, 1987; Kline et al., 2000). The explanative power of PL on subordinate responses is quite large and there are significant interaction effects between them. Thus, the explanative power of PL should still be significant, even if the effects of common method variance are removed. Nonetheless, future research should use different variables for different respondents, and longitudinal studies should also be attempted to shed light on the causality involved and sidestep the same-source bias.

Last, as the present study does not focus on mid- and high-level managers from family businesses, the PL effect may have been weakened. Generally, PL is common in higher-level managers in family businesses, yet the participants in this study were mostly only mid- and low-level managers. Thus, future research should include high-level managers from family businesses as participants.

Although research about PL is very important to Chinese organizations, few studies have probed deeply into this topic. This study is only a beginning, and many issues remain to be explored. These issues include: (1) The concepts of authoritarian leadership discussed in this study consist mainly of negative behaviors such as autocracy and severity. Positive behaviors such as awe, dignity, and prestige should be added to tap the totality of authoritarian leadership. (2) What is the nature of PL? Is it dispositional or situational? (3) When is PL

(24)

effective? What are the effects of PL on group processes and group performance as opposed to individual processes and performance? (4) What are the elements comprising PL? In addition to cultural elements, what are the influences of industrial and organization types? Can PL also describe organization types other than business organizations? Besides the Taiwanese society, is PL also applicable to other Chinese societies such as China, Singapore, and Hong Kong? (5) In addition to subordinates’ authority orientation as a part of traditionality, what other personal or managerial factors may have moderating effects on PL? For instance, a subordinate’s need for dependence or the manager’s competence may be probable factors.

Finally, some practical implications are obtained in this study. We found that when compared with the Western leadership style, PL based on Chinese culture has a unique effect on subordinate responses. While authoritarian leadership of PL is losing its influence, benevolent and moral leadership styles are influential, and may even become more important. Leadership with equivalent en (favor) and wei (might) has a positive and significant effect, yet leadership with equivalent der (moral) and wei (might) has a negative effect. Consequently, we believe that enterprise leaders in Chinese societies should employ a unified leadership style of shi-en (granting favors) or shuh-der (setting an example of morality), or an integrated leadership style with equivalent en (favor) and wei (might) to achieve the best results. In addition, whether or not a leader should display authoritarian leadership depends on the subordinate’s authority orientation. Authoritarian leadership can lead to positive effects only for highly authority-oriented subordinates.

Acknowledgment

We thank the contract/grant sponsor, Ministry of Education, Republic of China - Contract/

grant number: 89-H-FA01-2-4-4.

End notes

1. See Appendix 1 for the PL scale items. 2. See Appendix 2.

3. For the subordinate responses items, see Appendix 3. 4. See Appendix 4.

References

Bass, B. M. (1985). Leadership and Performance Beyond Expectations. New York: Free Press. Bass, B. M. (1996). A New Paradigm of Leadership: An Inquiry Into Transformational Leadership.

Alexandria, VA: U.S. Army Research Institute for the Behavioural Social Sciences.

Bass, B. M. (1998). Transformational Leadership: Industrial, Military, and Educational Impact. Mahwah, NJ: Lawrence Erlbaum.

Bass, B. M. & Avolio (1990). The implication of transactional and transformational leadership for individual, team, and organizational development. Research in Organizational Change and