Elevating Organizational Citizenship Behavior among Local Government Employees : The Mediating Role of Job Satisfaction

Local Government being the third-tier in Malaysian government has been identified as one of the government agencies mandated to drive the economic growth and development of a particular district and its residents. For years, local governments have been plagued with public complaints as the issues of inefficiency of service quality and service delivery has yet to meet the expectation of the public. Thus, the objectives of this study were to explore the extent to which the quality of leader-member exchange (LMX) affected the subordinates’ organizational citizenship behaviors’ (OCB) and job satisfaction as well as the role of job satisfaction in mediating this relationship. Data from a sample of 212 employees of local government through a structured questionnaire using disproportioned stratified sampling procedure were used to examine the hypothesized model. Result proved that, after controlling for selected demographic variables, the quality of LMX has a positive impact towards the subordinates exhibiting citizenship behavior and their level of job satisfaction. Furthermore, the finding confirmed that job satisfaction partially mediated the relationship between LMX and citizenship behavior. In conclusion, to fulfill the mandated objectives and to instill public confidence towards efficiency of the various services provided by the local government, fostering quality dyadic relationship and improving satisfaction among the workforce should be given priority as these elements have proved to be significant in elevating the citizenship behavior among the local governments’ employees.


Introduction
One of the biggest challenges facing service organizations is to cater the needs of its end-users in providing continuously high quality service on the back of limited resources like workforce, financial and infrastructure.Delivering consistently quality service is among the important factors that contribute to the reputation and strong standing of an organization in the eyes of the public.Public service organization like local government is no exception as this important agency plays a crucial role in providing public goods and services that are specific to its localities.Being close to the community, local government is burdened with multilayered tasks in their efforts to provide good service delivery and to maintain citizen satisfaction.Therefore, the ability to respond to the changing needs of the environment is a key factor in ensuring a nation's sustainable development, growth and global competitiveness (Kandan & Ibrahim, 2010).Hence, OCB is the possible solution in meeting citizen satisfaction as OCB is among the important factors in influencing organizational effectiveness at relatively low cost (Organ, 1988).
Leadership works best when there is a match between the identity level of followers and the focus of leaders, as people of similar behavior tend to be attracted to each other (Lord & Brown, 2001).Thus, in the public sector context, as mentioned by ex-General Secretary Tan Sri Ismail (2007), it invariably rests on the capability and capacity of the leaders to influence their subordinates towards accomplishing the vision, mission and objectives of the organization and, even more so, in meeting the expectation of the wider public interest.Basing on this point, the major concern of the public sector organization today, is to establish effective leadership management and the core element in developing this leadership management is through quality level of exchange relationship between leaders and subordinates.It is believed that understanding the process path by which LMX relationships impact important performance behaviors such as OCB are crucial in advancing leadership in an organization (Lo, Ramayah & Kueh, 2006).
Being the lowest in the government hierarchy, local government is identified as one of the most important agencies to drive the economic growth and social development of a particular district and its residents as they are managers of urban environments.Their performance has been the "topic of interest" by the public at large as their performance has yet to achieve the satisfactory level.The local government operation has been barraged with questions and complaints directly via electronic and written media and tougher higher ups at the state and federal levels, implying the inefficiency and ineffectiveness of the local government management (Zaherawati, Zuriawati & Hilme, 2010).They have often been criticized for their poor services, bureaucratic, inefficient, wasteful, unresponsive and not citizen-friendly and the allegations of bribery and misuse of power were frequently raised by the public (Hazman & Kalianan, 2008;Danilah & Nabiah, 2011).Based on statistics compiled by Public Complaints Bureau, the local government championed in the numbers of complaints by the public for three consecutive years from 2009 to 2011 (annual reports of Public Complaints Bureau).
In view of these alarming dissatisfaction by public towards local governments' management is the motivation factor in undertaking this study.Therefore, this study is a modest attempt to response to the calls made by Henderson, Wayne, Shore, Bommer & Tetrick (2008) and Liden, Wayne & Sparrowe (2000) that the relationship between LMX and employees work outcomes remains largely inconclusive and that a more detailed scrutiny of this relationship is clearly warranted.The study attempts to provide empirical evidence on these objectives: (1) to determine the effect of dyadic relationship (LMX) has on employees' citizenship behavior and on job satisfaction and (2) to investigate the role of job satisfaction in mediating the relationship between LMX and OCB.Surprisingly, little is known about whether and how job satisfaction mediates the relationship between LMX and OCB in Malaysian context and in the public sector focusing on local government in particular.
The remainder of this paper is organized as follows: literature reviews of the variables of interest are put forward along with hypotheses development and the conceptual framework.Research methodology, data collection and data analyses are then discussed.Discussion on empirical findings from the hypotheses being tested is presented and elaborated.Lastly, conclusion, limitations and future research directions wrap up this paper.

Local Government
There are three levels in the Malaysian government with federal government being the upper or center government follows by the state government and lastly the local government which is the lowest level government.Forming the largest population in the Malaysian workforce, these three governments are the backbone of the social and economic growth in Malaysia.There are 149 local governments in Malaysia with 99 are located in Peninsular Malaysia while the balance 50 in East Malaysia.Local government is categorized into three levels of hierarchy with 12 city councils for city centers, 39 municipal councils for larger towns and 98 district councils for small urban town (Ministry of Housing & Local Government, 2012).The local government in Peninsular Malaysia follows the Local Government Act 1976 whereas Sabah and Sarawak (East Malaysia) follow their own respective Local Government Ordinances.
There are three categories of workforce in a local government: Professional and Administrative staff, Support I and Support II staff adopting a pyramid style organizational structure.The Professional and Administrative staff comprises of Mayor/President, Secretary and Head of Departments as they are involved directly in the administration, strategic town planning and decision-making purposes.Support I staff or the middle level managers are responsible for implementing strategies and policies and those who are responsible in overseeing the manner in which the operational staff discharging their duties.Support II staff or operational staff are those who are involved in day-to-day operations where their tasks are more towards executing the tasks assigned rather than giving out instructions.
Under the control of the State government, the power to make decisions and resolutions are transferred to local government in managing its respective locality.Among the power conferred to local government includes collection of taxes, creation of laws and rules as well as approval of licenses and permits for business activities.
Local government is also entrusted to safeguard public health and sanitation, environmental protection and building controls, social and economic development and general maintenance functions of urban infrastructure within its jurisdiction (Hazaman & Kalianan, 2008).

Organizational Citizenship Behaviour
OCB has attracted many scholars to continuously investigate and understand this important work outcome as it has been found to affect organizational effectiveness (Organ, 1988).The social exchange theory is often used to examine various aspects of employee reciprocity including OCB (Hopkins, 2002).OCB may contribute to organizational effectiveness through enhancing coworker and managerial productivity, enabling the organization to adapt to environmental changes and strengthening the coordination within and across work groups (Podaskoff, MacKenzie, Paine & Bachrach, 2000).OCB can enhance organizational performance because they lubricate the social machinery of the organization, provides the flexibility needed to work through many unexpected contingencies, reduce friction and leads to increase efficiency (Smith, Organ & Near, 1983;Podsakoff et al., 2009).As noted by Organ (1988), it is the aggregate summation of OCB performed across time and across persons in the group, department and organizational levels, that contributes to the overall performance of the organization.It is the behavior that are clearly noticeable by peers, supervisors or the public at large (Noormala & Alam, 2009).Organ (1988) argued that OCB is held vital to the survival of an organization and organizations that foster good citizenship behaviors are more interesting places to work and are able to employ and retain the best people (George & Bettenhausen, 1990).In other words, OCB is an employee's beneficial behaviors that leaders want but cannot require (Motowildo, 2000).
In examining the literature, scholars cannot agree on the particular dimensions of OCB.Podsakoff et al. (2000) defined 30 different citizenship behaviors in their review study; while Williams & Anderson (1991) divided OCB into two types: OCB towards individuals and OCB towards the organization.Despite the many types of OCB proposed in literature , the most frequently cited taxonomy of OCB is the five dimensional model developed by Organ (1988) and the scales developed by Podsakoff, MacKenzie, Moorman & Fetter (1990).The five dimensions of OCB are (1) altruism -employees help co-worker with job relevant duties and tasks, (2) courtesy -being polite, considerate of others and treat them with respect, (3) conscientiousness -when employee goes beyond normal requirements or expectations, (4) sportsmanship -a positive attitude and willing to tolerate less than ideal circumstances without complaining and (5) civic virtue -participating in the governance of the organization.
In meta-analysis review of attitudinal and dispositional predictors of OCB, Organ & Ryan (1995) concluded that job attitudes such as job satisfaction, perceived fairness, organizational commitment and leader supportiveness are robust predictors of citizenship behavior.According to Aquino & Bommer (2003), OCB can enhance the social attractiveness in a work unit and thus, this may create a favorable and conducive working environment in a work unit.OCB is able to foster the effectiveness of organizations in terms of achieving a higher level of work group performance from the perspective of quantity and quality of work (Podsakoff, Ahearne & MacKenzie, 1997).In general, organization with a higher degree of OCB leads to reduce absenteeism and turnover, increase employees' satisfaction and retention which subsequently lead to improve organizational performance and success and elevate customers' satisfaction and loyalty (LePine, Erez & Johnson, 2002;Chughtai & Zafar, 2006;Chahal & Mehta, 2011).Thus, when considering the importance of OCB, it will be useful to examine the possible interactions between LMX, job satisfaction and OCB.

Relation between Leader-Member Exchange and Job Satisfaction
Job satisfaction is defined as having positive feelings about a job based on their perceptions and evaluation of the job's characteristics (Robbins & Judge, 2007).It refers to employees' satisfaction with the general aspects of work situation such as pay, supervision, the firm as a whole, the job itself, fellow employees and the prospects of advancement.Job satisfaction, in general, can be categorized into intrinsic and extrinsic satisfaction (Weiss, Dawis, England & Lofquist, 1967).Intrinsic satisfaction originates from within the individual and has psychological value while extrinsic satisfaction originates from the environment which is outside the control of the employees.
Empirical evidence proved that LMX is a positive predictor of job satisfaction (Bhal & Ansari, 2007;Lapierre & Hackett, 2007;Gerstner & Day, 1997;Cheung & Wu, 2012).In a meta-analysis of LMX literature, Gerstner & Day (1997) asserted that there is a significant correlation between LMX and overall job satisfaction.According to them, the affective bond employees' form with their organizations is proportional to the affective bond they have with their immediate supervisor.For that reason, their supervisor generally has a significant and direct influence over their daily work experiences (Hackett & LaPierre, 2004).Thus, this can be inferred that this affective bond with the supervisor will impact the employees' overall job satisfaction.In the same vein, a research study conducted by Mardanov, Heischmidt & Henson (2008) in a banking industry found that good quality LMX has a significant impact on job satisfaction where 80.7% of variance in job satisfaction was explained by LMX.In contrast, Pillai, Scandura & Williams (1999) found that LMX failed to influence employees' job satisfaction.Hence, the following hypothesis is put forward: H1: Leader-member exchange is positively related to employees' job satisfaction.

Relation between Leader-Member Exchange and Organizational Citizenship Behavior
The LMX theory posits that the relationship between leaders and subordinates develop as a result of work-related exchanges.In the dyadic relationship, leaders treat their subordinates differently over a period of time, leading to the development of relatively stable dyads that range from lower (out-group) to higher (in-group) quality exchanges (Sherony & Green, 2002).In-group subordinates are those employees who have created high-quality relationships with their leaders, characterized by mutual trust, respect, liking and reciprocal influence (Liden & Maslyn, 1998).On the other hand, the out-group subordinates are those employees who have low-quality relationships with their leaders, characterized as "limited reciprocal trust and support and few rewards" (Liden & Maslyn, 1998).The linkages between LMX and OCB have become the foundation of a new era of managing a diversified workforce in the advent of a globalized world (Lo et al., 2006).
Grounded on social exchange theory, in-group subordinates with whom the leaders share valued resources such as time, information and personal support tend to consistently perform OCB with passion as compared to out-group subordinates whose leaders adhere to contractually establish job roles (Hui, Law & Chen, 1999).A meta-analysis by Hackett, Farh, Song & Lapierre (2003) reported a mean correlation of 0.32 between LMX and overall OCB, thus, conclude that OCB plays a key role in the reciprocal social exchange process of LMX.Hence, the effectiveness and the willingness of subordinates in performing OCB are very much depended on the leadership effectiveness and the relationship being established as noted by Podsakoff et al., (2000).Indeed, leaders can encourage the employees' engagement in OCB even in the case of thwarting personality characteristics (Emmerik, Hetty & Euwema, 2007).Study by Bo-Kai (2013) in 44 elementary schools in Taiwan found loyalty and contribution dimensions of LMX have direct influence on OCB towards individuals while affect and respect dimensions of LMX have direct influence on OCB towards organization.This is supported by LePine et al. (2002) in their meta-analysis on OCB that leader support emerged as the strongest predictor.In contrast, Aryee & Chen (2006) found an insignificant link between LMX and task performance and psychological withdrawal.
H2: Leader-member exchange is positively related to employees' citizenship behavior.

Job Satisfaction as the Mediator
The inconclusive findings as noted in the relationships between LMX and job satisfaction and LMX and OCB indicate that the process through which LMX quality influences the work outcomes is still not fully understood and there is a strong possibility of the existence of mediators (Cheung & Wu, 2012).The fact that satisfaction functions as a mediator within the organizational behaviors has been explicitly articulated from both theoretical and empirical aspects (Crede, Chernyshenko, Strak, Dalal & Bashshur, 2007).However, literature does not provide much evidence whether and how job satisfaction mediates the relationships between LMX and OCB in Malaysian context focusing specifically on one of the important government agencies, the local government.
Several studies have provided evidence the importance of job satisfaction in mediating the LMX and work outcome relationship.Among them were Songbo, Xiao-Shung & Wei (2013) found that leaders can reduce unethical behaviors by nurturing social exchange with their subordinates as well as their job satisfaction and Cheung and Wu (2012) found influence of job satisfaction fully mediated the relationship between LMX and two work outcomes in Chinese firms, namely, organizational commitment and intention to leave.Another study by Tsai & Su (2011) indicated a mediating effect of job satisfaction on transactional leadership and service-oriented OCB was higher than transformational leadership among flight attendants employed in international airlines operating in Taiwan.This was because flight attendants working in a tight scheduled need to perform their level best during the short period of flight time to satisfy its passengers, thus, in this situation, transactional leadership was perhaps the more appropriate behavior in achieving the mission within a strict time period.Another recent study by Bo-Kai (2013) on the mediating effect of job satisfaction in the relationship between LMX and OCB in 44 elementary schools in Taiwan found that job satisfaction has a partial mediating effect between respect dimension of LMX and OCB towards the organization.Kasemsap (2013) found that job satisfaction positively mediated the influence of LMX on OCB in textiles factories operating in Thailand.Ballinger, Lehman & Schoorman (2010) reported that affective and cognitive satisfaction weakened the link between LMX quality and intention to leave.
By reviewing these mediating literature, job satisfaction has always been an important attitude outcome in organizational behavior researches and it is not difficult to infer that part of the mechanism through which LMX influences OCB level is through job satisfaction.Thus, based on this argument, the following hypothesis is proposed: H3: Job satisfaction mediates the relationship between LMX and OCB.
Based on the discussion of the literature, the theoretical model that guides the present study is presented in Figure 1.
Figure 1.Conceptual framework of the relationship between leader-member exchange and organizational citizenship behavior mediated by job satisfaction

Study Setting and Sampling
East Coast Malaysia consists of three states namely, Pahang, Terengganu and Kelantan.There are 29 local governments in these states comprising of one city council, six municipal councils and twenty-two district councils.Stratified random sampling was used in selecting sample population.This technique was appropriate where various departments with different job functions existed in a local government while functions between local governments were almost similar in nature.
The local governments were initially categorized into city council, municipal councils and district councils.
Since there was only one city council, the said city council was excluded from the sampling process, leaving 6 municipal councils and 22 district councils.Two local governments from each state were considered sufficient to represent the population.Thus, two municipal councils and four district councils (totaling six local governments) were randomly selected where Pahang and Terengganu were represented with one municipal council and one district council each while Kelantan with two district councils.
With a population size of 1,025 employees, the sample size required for this study was 285 respondents (Krejcie & Morgan, 1970).As the number of professional and administrative group was much smaller compared to Support I and Support II group (low to middle level management), a disproportionate stratified random sampling procedure was employed.Data was collected through survey questionnaires as this method provided certain advantages such as convenience, accessibility and minimal costs (Malhotra, 2007).The questionnaires were personally delivered to chief clerk of each local government who acted as the 'personal contact' person.Brief explanation as to the variables and objectives of the study was given to these chief clerks and these questionnaires were self-administered.Out of 300 questionnaires distributed, 220 questionnaires or 73.3% were returned and only 212 questionnaires were found useable, thus marked a return rate of 70.6%.

Pilot Testing
Before undertaking the full scale study, a pilot test was conducted to test the reliability and validity of the questionnaires with 32 respondents were randomly selected from one local government in Terengganu.They were selected to represent the target group since the job description and structure are similar to the actual research being conducted.Data was analyzed using SPSS version 20 for windows and results indicated that the questionnaire was reliable and valid.Cronbach's alpha values indicated good measurement of reliability with all variables of interest exceeded the 0.7 benchmark (Sekaran & Bougie, 2010).Normality tests revealed that the data was normally distributed with skewness and kurtosis analyses further indicated that the data was normal.
The assumptions of linearity and homocedasticity in multivariate analyses have been fulfilled and the data was also found to be free from multicollinearity.

Leader-Member Exchange
Job Satisfaction Organizational Citizenship Behaviour

Research Instruments
The scales used in this study were adapted from established instruments with proven reliability and validity.Since they were originally worded in English, translation to Malay language was carried out as the respondents were predominantly Malay and majority from low to middle management.In order to guarantee cross-linguistic quality, these translated items were then translated back into English following the method suggested by Brislin (1970).
The research instrument was divided into four parts: demographic information and three scales to measure OCB, LMX and job satisfaction using a five-point Likert scales ranging from (1) strongly disagree to (5) strongly agree.Each construct measurement is discussed as follows:

Organizational Citizenship Behavior
This study adopted a 22-item scale developed by Podsakoff et al. (1990) based on the five dimensions proposed by Organ (1988) namely altruism, conscientiousness, sportsmanship, courtesy and civic virtue.As recommended by Hoffman, Blair, Meriac & Woehr (2007), the operationalizations of OCB are best viewed as indicators of a general OCB factor.Le Pine et al. ( 2002) in their meta-analysis also supported this recommendation that predictive relationships with the broader OCB criterion are as good as, or superior to, those with narrower dimensional criteria.Thus, based on these recommendations, a latent construct was used in measuring the OCB level.

Leader-Member Exchange
This variable was measured using the 12 item questionnaire adapted from Liden & Maslyn (1998) based on four dimensions of namely affect (mutual affection members of the dyad have for each other based on interpersonal attraction), contribution (perception of the current level of work-oriented activity each member of the dyad puts forth), loyalty (expression of social support among each member of the dyad) and professional respect (perception of each member of the dyad has built a reputation of work-related activity).An arithmetic mean was used as the measurement for this scale.

Job Satisfaction
This variable was measured using the short version of Minnesota Satisfaction Questionnaire (MSQ) developed by Weiss, et al. (1967).A latent construct was used to measure job satisfaction based on the suggestions made by Hancer and George (2003).

Control Variable
Type of local government, service category and LMX tenure were controlled in this study to reduce error variance and to rule out their possible impact on the variables under study (Aiken & West, 1991).Sapie (2012) in his study of local governments in East Coast Malaysia found that type of authority has significant impact on competency level of the employees being studied while Suliman (2007) found that service category was positively related to job performance.Furthermore, the length of the dyadic relationship was controlled for its potential effects on the quality of leader-subordinate relationships where Van Dyne & LePine (1998) found that working relationship with the present supervisor has negative impact on voice behavior (one of the components under OCB stable).Thus, these demographic variables were included as covariates in this study.

Characteristics of Respondents and Procedure
Out of the 3 States in East Coast Malaysia, Pahang recorded the highest respondents (35%) followed by Kelantan (33%) and lastly Terengganu (32%).In terms of local government, the municipal councils from Pahang and Terengganu recorded the highest number of respondents since municipal council has larger jurisdiction area compared to district council, hence, bigger workforce compared to district council.Majority of the respondents were Malays (99%) with female (53.3%) outnumbered the male counterpart (46.7%).54.7% of the respondents aged 35 years and below while 42.5% of them were in the range of 36 to 55 years and 2.8% were approaching retirement age.60.8% of the respondents attained below certificate level of education while the balance 39.2% has achieved higher educational qualifications.Corresponding to the level of education achieved, 85.8% of respondents fell under Sokongan I and Sokongan II category (low & middle level management).31.1% of the respondents have been in service for less than 5 years and the balance 68.9% have been in service for a period more than 5 years with 27% of them have surpassed 15 years in service threshold.
In terms of the respondent's present supervisor, the majority of them was reported to male supervisors (72.2%) compared to female supervisors (27.8%).Furthermore, 67% of these respondents have worked with the current supervisor less than 5 years while 25% of them have been reporting to their supervisors in the range of 5 to 15 years.8% of the respondents have not changed their supervisors where they have been reporting to the same supervisor for the past 15 years.This indicated that job rotations were being practiced by these local governments.

Factor Analysis
Confirmatory factor analysis was carried out independently on these variables to ensure that all the items were loaded on their hypothesized factors.Factor analysis conducted on OCB yielded one factor with an eigenvalue greater than 1 (eigenvalues = 6.96), explaining a total variance of 31.64%.The Kaiser-Meyer-Olkin measures of sampling adequacy stood at 0.85 and the Bartlett's test has indicated significance of overall correlations within the correlation matrix with a value of 1760.14.All 22 OCB items were retained as they have factor loadings greater than 0.30 (Hair, Black, Babin & Anderson, 2010).
Factor analysis conducted on LMX also gave rise to one factor with an eigenvalue exceeding 1.0 (eigenvalues = 5.98), explaining a total of 49.80% of the variance.The Kaiser-Meyer-Olkin measures of sampling adequacy stood at 0.88 and the Bartlett's test has indicated significance of overall correlations within the correlation matrix with a value of 1498.24.As these items have a factor loading of above 0.30 (Hair et al., 2010), all 12 items were maintained.Lastly, factor analysis on job satisfaction was carried out where one factor emerged with an eigenvalue greater than 1 (eigenvalues = 8.07), explaining a total variance of 40.37%.The Kaiser-Meyer-Olkin measures of sampling adequacy stood at 0.90 and the Bartlett's test has indicated significance of overall correlations within the correlation matrix with a value of 2098.03.19 items were retained and one item was omitted due to low factor loading of 0.30.The item deleted was "The chance to be somebody in the community".

Descriptive Statistics
Means, standard deviations and zero-order correlations were computed and displayed in Table1.The mean scores for job satisfaction, OCB and LMX were above their respective midpoints ranged from 3.76, 3.77 and 3.80 respectively on the five-point scale, indicating that the respondents perceived moderate amount of agreements on these variables.Correlation analysis was performed to determine the directionality and the magnitude of the relationship among these variables.Correlation matrix indicated that statistical independence existed between these variables and they have demonstrated accepted levels of correlation.As expected, LMX was found to have a moderate positive significant influence of r = 0.54, p < 0.01 on OCB while job satisfaction was also found to have a strong positive significant influence on OCB at r = 0.61, p < 0.01.Furthermore, a strong positive significant relationship of r = 0.66, p < 0.01 was established between LMX and job satisfaction.The Cronbach's alphas for these variables were well above the minimum acceptable reliability of 0.7 (Sekaran & Bougie, 2010) which indicated good internal consistency.

Mediation Analysis
In testing the mediating effect of job satisfaction in LMX and OCB relationship, Baron and Kenny (1986) three-step regression procedure was applied.Three conditions must be satisfied to support the mediating effect: (1) the independent variable (LMX) must effect the mediating variable (job satisfaction); (2) the independent variable (LMX) must effect the dependent variable (OCB) and (3) after the inclusion of the mediating variable (job satisfaction) in the second regression, the regression coefficient of the independent variable (LMX) decreases for partial meditation or becomes insignificant for full meditation and the mediating variable (job satisfaction) must have a significant relationship with dependent variable (OCB).Nevertheless, it is important to note that psychological concepts often have a number of interrelationships.Thus, according to Baron and Kenny (1986), it is more realistic to say that meditation significantly decreases rather than completely eliminate the relationship between the individual variable (LMX) and dependent variable (OCB).The results of mediation regression analysis were shown in Table 2.
As evidenced, in step 1, after restraining the co-variates, LMX has a significant relationship with job satisfaction (Beta = 0.655, p < 0.001), thus step I has been met and Hypothesis 1 was accordingly fully supported.In Step 2, after controlling the effect of control variables, LMX has a significant positive relationship with OCB (Beta = 0.520, p < 0.001), thus fulfilling step 2 and accordingly Hypothesis 2 was confirmed.30.8% of the variance in OCB is explained by LMX.
The last step 3, when job satisfaction (mediating variable) was included into the regression model, job satisfaction has a significant positive relationship with OCB (Beta = 0.459, p < 0.001) and the regression coefficient of LMX and OCB was reduced from Beta 0.520 to Beta 0.219.This decrease showed that job satisfaction partially mediate the relationship between LMX and OCB.Thus, hypothesis 3 was supported.Furthermore, with the inclusion of job satisfaction in the regression model, the variance in OCB being explained by LMX and job satisfaction has increased to 42.8%.Notes: significant at**p < 0.01, ***p < 0.001; N = 212

Discussion and Conclusion
The objective of a significant relationship between LMX and OCB was reaffirmed in this study which supported the notion that LMX and OCB are two important components related to the establishment of favorable relationships between leaders and subordinates, and, therefore motivates subordinates to perform one of the most important work outcomes, citizenship behaviors.As put forward by Lo et al. (2006), the central premise of OCB theory was that LMX was partially an expression of subordinates gratitude and reciprocity for positive work-related behaviors originating from a high-quality relationship with one's supervisors.In conformity with previous studies, empirical evidence obtained from this study provided support of the significant influence of LMX and OCB in Malaysian context.In other words, in-group employees tend to exhibit better citizenship behavior.This result was congruent with several findings based on Malaysian settings such as Lo et al. (2006) in manufacturing sector and Kandan and Ibrahim (2010) in local government sector where they found that LMX and OCB have positive relationships in a work unit.As noted by Ali, Abu Daud, Aminah & Bahaman (2008), OCB helped to fulfill the reciprocity obligations of followers and represented an exchange currency that was diffuse, unspecified and weakly time-bound.Although 30.8% variance in OCB was being explained by LMX, this provided enough evidence that maintaining and fostering good-quality relationship between leaders and subordinates are important in the context of local government.As hypothesized, LMX was positively related to job satisfaction and this finding is congruent with that of the past researches by Mardanov et al. (2008), Hackett & LaPierre (2004) and Bhal & Ansari (2007) that establishing quality LMX will elicit the citizenship behaviors among the subordinates and employees' satisfaction in the job.Thus, this will help in reducing the numbers of complaints by the public which in turn will made local government an important engine of growth economically and socially for a particular district.
Responding to researchers' calls from Henderson et al. (2008) and Liden et al. (2000) for more detailed scrutiny arising from inconclusive relationship between LMX and work outcomes, this study provided empirical evidence that job satisfaction partially mediated the relationship between LMX and OCB in which influence of job satisfaction translated into positive effect of LMX in elevating the citizenship behavior among local governments' employees.This mediation result extended the understanding of social exchange process and the norms of reciprocity between leaders and subordinates.The results suggested that when subordinates with a high degree of LMX were contended with their jobs, the positive effects of high-quality LMX influenced their attitudes and behaviors by being cooperative with the organization and performing beyond their call of duty.The increment in the percentage of variance being explained by LMX and job satisfaction to 42.8% further reaffirmed that when employees with high satisfaction were always thankful to the organization and were likely to produce positive behaviors such as OCB as a gesture of "pay back" the organization.This finding was also congruent with that in non-Western context based on the findings by Kasemsap (2013) in textile factories operating in Thailand, Songbo et al. (2013) and Cheung and Wu (2013) in Chinese firms which these researches found the positive mediating effect of job satisfaction in improving employees work outcomes.These findings supported the propositions that LMX and job satisfaction can exert significant influence on OCB and empirical evidence also provided support that part of mechanism through which LMX influence work outcome; specifically OCB is through job satisfaction.Being the government's lowest tier and has the closest relationship with the communities, local governments have a high responsibility to provide quality services needed and expected by the communities, namely the stakeholders.Hence, holding the title as the managers of urban environments, it is important that quality dyadic relationship being established and it is also equally important to foster the aspects of employees' job satisfaction which in turn will promote OCB among local governments' employees.As empirical evidence suggested that the combination of these elements will help organizations like local governments to achieve its mandated objectives and to instill public confidence towards efficiency of the various services provided at a relatively low cost.

Limitations and Directions for Future Studies
The present study makes useful additions to the current body of management literature and to the field of organizational behavior by examining the effects of job satisfaction in mediating the relationship between LMX and OCB among local governments' employees in East Coast Malaysia.There were, however, some limitations noted in this study.The first was that the sample size was limited to three states in East Coast Malaysia, thus, it may raise the issue of generalization of the findings.Thus, it is recommended that future studies to expand to other geographical zones to include all local governments throughout Malaysia in determining whether these findings can be generalized.The second limitations noted was the cross-sectional data were used which limited the findings to the time of research.The results might be different in a longitudinal approach which would give deeper insight into the subject matter and the corresponding relationships.Moreover, all research variables were reported by employees themselves, this might lead to a problem of common source bias (Podsakoff & Organ, 1986).As pointed out by Spector (1997), common source bias is unlikely to cause a problem for well-developed instruments.This research used well established questionnaires with proven reliability and validity, hence, the effect of common source bias is minimized.
Following the calls by Henderson et al. (2008) and Liden et al. (2000) for more detailed scrutiny arising from inconclusive relationship between LMX and work outcomes, future studies should aim to theoretically and empirically explore the relationships between LMX and other key organizational outcomes such as turnover intention and perceived organizational support as well as counterproductive work behavior or unethical behavior.Since the understanding of OCB is improving with the exploration of job satisfaction as mediator in relation to LMX, future researches may examine other potential mediating variable such as organizational commitment or work engagement on LMX and OCB relationship.Furthermore, comparative studies across other government sectors especially those agencies that have close contact with the public like armed forces, hospitals or to private sector that has direct contact with the customers like the hospitality sector or banking sector are needed so as to improve and increase the ability to generalize the research findings.

Table 2 .
Results of hierarchical regressions analysis of job satisfaction mediating LMX and OCB