Effects of Supervisor-Subordinates Relations on Subordinates’ Commitment Behaviour in Ghana

This study examined the effects of subordinates’ perception of supervisor favouritism on subordinates’ commitment to their organizations in the Ghanaian context. A cross-sectional survey was employed to study 296 employees comprising supervisors and subordinates in two organizations (public and private). Results from t-test, analysis of variance, Pearson product-moment correlation coefficient and hierarchical multiple regression showed: (a) no significant difference in the amount of perceived supervisory support; (b) that perceived favouritism based on ethnicity had a greater tendency to influence supervisor-subordinate exchanges and also the amount of perceived supervisory support; and (c) that subordinates who received more supervisory support were highly committed. There was an inverse relationship between perceived favouritism and organizational commitment of subordinates. Surprisingly, subordinates who perceived favouritism did not show a higher tendency to quit their organizations but rather continued to stay. The organizations could benefit greatly if the lower turnover intents are translated into practical work outputs but that, of course, will involve reducing perceived favouritism to the barest minimum. Suggestions for future research are made.


Introduction
Research findings suggest that workforce diversity could be both beneficial for higher productivity and low turnover if diversity is well-managed but challenging for the implementation of policies fairly and managing employees' perception of fairness. It is more exigent when employees have varying perceptions of fairness and favouritism which affect their work behaviour, attitude and the organization's task (Ensher et.al, 2001). Graves and Powell (1995) show how perceived similarity between recruiters and applicants influence recruiters' decision to employ an applicant. Such a situation may affect the subsequent work behaviours and attitudes of workers in the organization. For organizations in developing countries, the challenge of diversity is enormous because the workforce is largely multi-ethnic.
Given the high rates of unemployment in these countries, the interplay of nepotism and favouritism tend to undermine the use of the merit system in organizational practices leading to poor work behaviour and productivity. The case of Ghana is illustrative. Post-independence governments focused on industrial sector development by establishing state-owned enterprises (SOEs) to compete with private firms, then solely owned by expatriates (Esseks, 1971a;1971b;Appiah-Kubi, 2001;Arthur, 2005). However, most of these SOEs performed so badly and also accumulated huge debts that a programme was launched between 1987 and 1999 to divest and reform these SOEs in order to improve management and performance (Appiah-Kubi, 2001).
Interestingly, research is inconclusive on whether private organizations perform better than public organizations (Cakmak and Zaim, 1992;Mohammed, 1992) and leaves puzzling the tendency towards privatization as the panacea to the SOEs problems. One plausible way to explain the performance of the SOEs is to explore supervisor-subordinates relations (i.e. the quality of their interactions on the job). It has been suggested that the industrial drive attracted most people (some with no requisite skills for formal sector work) to move to urban centers in search of jobs (Peil, 1979). For some of those people, massaging ethnic identities for reciprocal favours created a leeway for overcoming the challenge of finding employment (Asamoah, 1990). Shellunkindo and Baguma (1993) have pointed out how recruitment practices at the workplace in Ghana undermine the merit system but promote favouritism and tribalism. The situation in Ghana, as elsewhere in Africa, affects work attitudes including commitment behaviour (i.e. tendency to stay) of members and might, to a large extent, account for the nonproductivity of the organizations involved (Abdulai, 2000). Therefore any attempt at improving productivity in Ghanaian public organizations should situate the interaction between the members' relational demography and the quality of their work relations at its core.

Context and the problem
Ghana is a multi-ethnic nation-state (Agyeman, 1998) and strong ethnic loyalties tend to adversely affect work relations especially in the application of the merit system in recruitment and promotion processes (Woode, 1997). Although some authors (e.g. Chazan, 1982) suggest a declining recourse to, and the importance of ethnicity, the ethnic factor continues to influence work processes in most organizations. Abdulai (2000) noted complaints about discriminatory practices in public organizations which breed suspicion, hatred and tension among workers. Ametewee (2007) observed that while ethnic group membership may not be important in specific social settings, elites quickly resort to massaging ethnic sentiments even from the down-trodden and illiterate employee once their interest and position in that social class is threatened. Collier (2000) cited a situation in Ghana during the 1980s, where workers from the locally-dominant ethnic groups were paid 25 percent higher than other workers (after controlling attributes such as skills, age, and education) to demonstrate how those from locally dominant ethnic groups were favoured in terms of promotion. This situation questions the rationale for massaging ethnicity in work settings and how it affects work interactions among organizational members.
Ethnic group membership imposes loyalty demands on group members (Agyeman, 1986;Tonah, 2007) and it is plausible that leaders will give more favours to their ethnic group (in-group) subordinates than other ethnic group (out-group) subordinates. This is because the leader and member will appear to grapple with the question of whether they can work and trust each other.
Therefore in-group subordinates will receive more resources than out-group members which may translate into differential work attitudes and outcomes. In-group members favoured by the leader may produce desirable outcomes such as commitment, citizenship behaviour and higher job performance. In contrast, out-group members may develop negative work attitudes which find expression in lateness, absenteeism, tardiness, slowdowns and unionism (Hodson and Sullivan, 1994).
These behaviors may adversely affect productivity in organizations, public or private. Since productivity involves efforts of all participants with different ethnicities, the objective of this paper is to systematize the extent to which ethnic favouritism shapes the quality of supervisor-subordinate interactions and the organizational commitment behaviour of subordinates. The remainder of the paper discusses theory and research on leadersubordinate interactions. This is followed by a brief note on data and methods. Afterwards, an analysis of results and findings is provided. It then concludes with some recommendations.

Leader-member exchange (LMX) theory
The leader-member exchange (LMX) theory, one of numerous approaches on supervisor-subordinate interaction, argues that leaders do not adopt a consistent leadership style towards all subordinates (Graen et.al, 1982). Rather, the theory suggests that leaders do discriminate and, by so doing, can construct negative situations i.e. leadership does not necessarily bring about positive results (Dansereau, 1995). Central to the LMX theory is the working relationship between a leader and the various members of a work unit, team, department or organization. One core assumption is that time and resources required of leaders make them develop trust for subordinates to help them manage the work unit (Graen and Uhl-Bien, 1995).
Most leaders develop high-quality exchange relationship (high-quality LMX) with only a few subordinates (Dansereau et al., 1975;Graen and Uhl-Bien, 1995). Lower quality LMX is informed by limited levels of reciprocity and support and render subordinates mere followers of formal organizational rules (Graen et.al, 1973). In contrast, higher quality LMX involves mutually supportive behaviours such as interpersonal attraction, mutual trust (e.g. Liden and Graen, 1980), loyalty and communication channels (Dienesch and Liden, 1986). Such subordinates may enjoy favours and opportunities in job assignments, promotions, and performance assessments. Consequently, work units split into an out-group consisting of subordinates with whom the leader has only superficial contact and an in-group consisting of highly-valued employees with whom there is intensive cooperation and communication (Van Breukelen et al., 2006). The fundamental point of departure between LMX and average leadership style (ALS) perspectives is that the LMX theory considers the relationship domain as essentially different from the leader domain (Graen and Uhl-Bien, 1995) and the quality of the LMX is more predictive of organizational outcomes than leader traits (House and Aditya, 1997).

Leader-member exchange research
Much evidence from correlational studies on LMX consistently supports the idea that high-quality exchange between a leader and subordinates predicts a number of relevant work outcomes (Vecchio, 1985;Lunenberg, 2010;Sun, Chow, Chiu and Pan, 2013). A study of 60 leader-subordinate dyads in a US university by Dansereau et al. (1975) found that ingroup leader-member exchanges produced greater leader attention and support in personal job problems. Consistently, Liden and Graen's (1980) study of 41 leader-member dyads from three semiautonomous service departments observed that highquality LMX yielded greater job responsibility, higher subordinate contributions to the work unit and high job performance.
Further, in a study of 109 supervisor-subordinate dyads, Graen and Schiemann (1978) found that leaders and subordinates highly agreed on job issues within high-and intermediate-quality dyads. Graen et.al's (1982) study of turnover intentions of 48 systems analysts and computer programmers in a large public utility found that the quality of LMX was an effective predictor of employee turnover. Recent studies show how high-quality LMX develops between leaders and subordinates in higher education (Power, 2013) and open and distance learning (Mapolisa and Kurasha, 2013) and in health management systems (Hays and Luo, 2013;Hunt, 2014).
The LMX theory remains the only leadership theory that makes the concept of dyadic relationship the centerpiece of the leadership process (Dansereau et al., 1975;Lunenberg, 2010;Barbuto and Gifford, 2012;Power, 2013). However, it downplays the impact of inter-group comparison which forms the basis of inter-group behaviour such as inter-group favouritism, ethnocentrism, and out-group denigration (Hogg et al., 2005). Given that the quality of LMX is influenced by a relational demography (Duchon et al., 198;Greenberg and Baron, 2000), it is argued that ethnicity may predict the quality of LMX in multi-ethnic Ghana where ethnicity and favouritism in work relations are well known (Woode, 1997). Thus leaders' differential treatment of their in-group subordinates vis-à-vis out-group subordinates may create a situation in which in-group members most likely receive more attention and support from the leader than out-group members. Allport's (1958) frustration-aggression approach, however observed that out-group subordinates who perceive being discriminated against by their leaders have the tendency to produce negative work attitudes and behaviours including tardiness, turnover, withdrawal, and aggression.

Hypotheses
For the purpose of the present paper, the following hypotheses were stated: H1: Similarity in ethnicity of leader and subordinate will be positively associated with higher supervisory support.

H2:
Perceived supervisory support will be positively related to level of subordinates' organizational commitment.

H3:
Perceived favouritism and organizational commitment of subordinates will be inversely related after controlling for the effects of age, education and tenure in organization.

Study organizations and respondents
Cross-sectional study design was used to collect data from respondents in two alcoholic beverage distilling organizations herein referred to as public and private organizations in Accra and Kumasi respectively. The former is a quasi-public organization situated in a relatively heterogeneous environment while the latter is a privately-owned company located in a relatively homogenous sociocultural environment. These characteristics allowed for comparisons of relative demography and work behaviours of employees. A census of all 384 active employees of the two organizations comprising supervisors and 363 subordinates was targeted (see Sabbi, 2012). However, 312 questionnaires were retrieved and 16 were discarded for incomplete responses. Thus the total number of respondents came to 296 giving a return rate of 81.3% and a response rate of 77.1%.

Survey procedure
The survey questionnaires were administered by the researcher and two assistants who spent four weeks in each organization to administer and retrieve completed questionnaires. Usable questionnaires were then processed and analyzed electronically using SPSS version 16. The dimensions of work behaviour and the measures used in the study are provided in Appendix I.

Socio-demographic data
The respondents' socio-demographic profile included the following: Age: The mean age of the respondents was 33.3 yrs (SD=7.9). The mean age for males was 34.6 yrs (SD=8.5) while the mean age for females was 30.7 (SD=5.9). Age is predictive of turnover and the generally youthful age may predict respondents' mobility between organizations.
Sex: Out of the 296 respondents, there were 198 (66.9%) males and 98 (33.1%) females. The public organization had 138 (69.7%) males compared with 60 (30.3%) from the private organization. Again, 57 (58.2%) of the females were from the public organization with 41 (41.8) from the private organization.  Only a handful (0.3%) had no formal education. This high educational achievement may also predict their turnover behaviour. The next section of the paper explores how subordinates' perception of favouritism affects their work-related attitudes and behaviour.

Subordinates' perception of organizational support
Subordinates' perception of support from their organization was measured using a four-item perceived organizational support scale.

The Quality of supervisor-subordinate interaction
For subordinates to continue to stay in their organization, they must receive support from their supervisors and the work group i.e. local (proximal) hypothesis (Becker, 1992). But this perceived support depends on the quality of LMX between the leader and the subordinate. Subordinates' perception of supervisory support was assessed with a 4-item perceived supervisory support scale and they were observed to hold above average notions of support from their supervisors ( revealed the subordinates had slightly higher perception of organizational support than perceived supervisory support. Thus they were more globally than locally committed (Becker, 1992) and more likely to produce organizational relevant outcomes such as job satisfaction and lower turnover (Cheng et al., 2003). When perceived supervisory support was matched with ethnicity, interesting results were observed for the three dominant ethnic groups: Akan, Ewe, and Ga-Dangme in the two organizations (the Guan, Grusi, Mole-Dagbani and 'other' groups had statistically insignificant scores). The Ga-Dangme had the overall supervisory support (over 75.0%) compared with the Ewe (just over 60%) and the Akan (a little over 50%). Thus ethnicity related to the amount of supervisory support subordinates received.

Demographic similarity and supervisor-subordinate exchange relationships
There is evidence on how relational demography (age, sex, education, ethnicity, etc) shapes the quality of LMX between supervisor-subordinate (Duchon et al. 1986;Tsui and O'Reilly, 1989). In this study, subordinates on average did not perceive their superiors as discriminatory in allocating resources and responsibilities (M=2.04; SD=0.68). Over 50% of the 275 subordinates saw no favouritism when their supervisors allocate resources and more than 70% of this group disagreed that their supervisors favour same ethnic group subordinates more than subordinates from other ethnic groups. The impression one gets from the perspective of the subordinates is that favouritism may not play any significant role in the distribution of organizational rewards.

Factors influencing supervisors' allocation of resources and responsibilities
Even though subordinates seemed to think otherwise, supervisors were found, on average, to use relational demography when allocating organizational resources (M=3.02, SD=0.68). Although over 90% said they were comfortable assigning resources and responsibilities to trustworthy and loyal subordinates irrespective of ethnic background, more than 60% of the 21 supervisors were willing to give more resources and responsibilities to subordinates from their ethnic groups. This they claimed would yield support and loyalty from such subordinates. Compellingly, more than 50% of the 21 supervisors said they sometimes give more resources to subordinates from their own ethnic group members in return for their loyalty and support. Results of analysis of variance (see Appendix II) found a significant relationship between one's ethnic group and the amount of perceived supervisory support. Therefore, subordinates in ethnic group with more support most likely get high-quality LMX than those from other ethnic groups even if they do not perceive such favours.

The Perceived supervisory support-organizational commitment nexus
Subordinates on average held high level of commitment to their organization (M=3.37, SD=0.88). Indeed, more than 70% out of the 275 subordinates said they were willing to help make their organization successful and more than 60% were loyal to their organizations. Male commitment to making the organization successful was relatively higher (75.3%) than female (67.7). And proportionately more males than female subordinates who received supervisory support were committed: 97.5% of 120 males and 88.7% of 53 females. This is contrary to previous studies (e.g. Mellor et al., 1994;Bauer and Green, 1994) which reported that females more often than males show strong feelings of commitment. Elizur (1994) and Furnham (1984) even claim that females are more committed than males because men tend seek money and cognitive outcomes e.g. competitiveness, and long-term career goals while women seek affiliative needs e.g. social approval and with short-term career goals. The present finding might arise from the fewer number of female supervisors in the organizations; thus female subordinates perceived less support and were not more committed.  (100) An observed trend between ethnicity and commitment levels becomes clearer if perceived supervisory support was added. Subordinates from ethnic groups with more support tend to show more commitment to their organization. Indeed, over 77% of 22 Ewes who perceived supervisory support in personal crises were proud of their organization compared with over 76% of 46 Ga-Dangme, and over 71% of 94 Akans who held similar views. Again, subordinates' perception of favouritism by supervisors influenced their sense of commitment to their organizations. Over 95% of 162 subordinates who thought that their supervisors do not show ethnic favouritism said they received on-the-job support and were more willing to expend extra effort for the organization to succeed. In contrast, 83.9% of 56 subordinates who thought that their supervisors play ethnic favouritism said they do not receive on-the-job support and were less willing to expend extra effort for the organization to succeed. It can thus be inferred that much as majority of the subordinates did not perceive favouritism from their supervisors, the few who perceived favouritism from their supervisors, especially on ethnic grounds, and therefore received very little support, showed a lower degree of commitment to their organizations. Perceived favouritism statistically predicted subordinates' organizational commitment. Step 2 (ps.005); *** p.001.

Conclusions
The findings of the study generally give some support to the LMX model (Graen and Uhl-Bien, 1995), which places emphasis on the assertion that leaders do not treat all subordinates equally but rather, leaders engage in preferential treatment of subordinates in view of the limited resources and time in exchange for subordinates' support and loyalty. It was observed that subordinates' perception of differential treatment by supervisors is influenced by ethnicity but those who perceive lower quality of LMX (less support) may continue to be committed to their organizations. Thus their perception of little supervisory support does not necessarily lead to negative work behaviours and attitudes.
The data presented give modest support to the three hypotheses and the study concludes as follows: First, where supervisor and subordinates come from the same ethnic group, they tend to have positive relationship between themselves; the amount of supervisory support received by subordinates is influenced by the relationship between supervisors and subordinates; and the more supervisors and subordinates feel belong to the same ethnic group, the more subordinates are likely to receive support from their supervisors. Second, there is a positive relationship between perceived fairness and the amount of perceived supervisory support. Thus, subordinates who think that their supervisors are fair also perceive greater amount of support from them. Third, there is a positive relationship between the degree of supervisory support and organizational commitment. Thus, subordinates who perceive supervisory support also express higher levels of commitment.