Bureaucratic culture, empowering leadership, affective commitment, and knowledge sharing behavior in Indonesian government public services

Abstract This paper aims to clarify the relationship among bureaucratic culture, empowering leadership, knowledge sharing behavior and affective commitment in the Indonesian public sector organizations. A total of 371 respondents from eleven local government public service institutions participated in this research. Structural Equation Modeling (SEM) was used for model testing. The results of this study indicate that bureaucratic culture has a negative and significant relationship with empowering leadership and affective commitment. Empowering leadership has a positive and significant impact on knowledge sharing behavior and affective commitment, and also mediated the relationship between bureaucratic culture with knowledge sharing behavior and affective commitment. The direct relationship between bureaucratic culture and knowledge sharing behavior proved insignificant, however, indirectly proved significant via empowering leadership. The tests of such relationship are expected to contribute to the theory and practice of the government sector.

Abstract: This paper aims to clarify the relationship among bureaucratic culture, empowering leadership, knowledge sharing behavior and affective commitment in the Indonesian public sector organizations. A total of 371 respondents from eleven local government public service institutions participated in this research. Structural Equation Modeling (SEM) was used for model testing. The results of this study indicate that bureaucratic culture has a negative and significant relationship with empowering leadership and affective commitment. Empowering leadership has a positive and significant impact on knowledge sharing behavior and affective commitment, and also mediated the relationship between bureaucratic culture with knowledge sharing behavior and affective commitment. The direct relationship between bureaucratic culture and knowledge sharing behavior proved insignificant, however, indirectly proved significant via empowering leadership. The tests of such ABOUT THE AUTHORS Hendryadi is a lecturer at Sekolah Tinggi Ilmu Ekonomi Indonesia Jakarta. His area of interest is human resource management, organizational behavior, Islamic work ethics, workplace incivility, and quantitative research methods.
Dr. Suratna is a professional psychologist and head of the human resources and organizational bureau, the agency for the assessment and application of technology (BPPT) Jakarta Indonesia. Her research interests include organizational behavior and human resources management.
Dr. Suryani is an Associate Professor at the Islamic State Institute of Lhokseumawe. She received the Doctorate in Islamic Economics from State Islamic University Syarif Hidayatullah Jakarta (2018). Her research interests include organizational culture, leadership, and ethics from the Islamic perspective.
Dr. Budi Purwanto received Doctorate for business management from Universitas Pancasila, Jakarta Indonesia. Working as a lecturer at Logistics and Transportation Institute of Trisakti and professionals in marine logistics. His research interests include supply chain management and marine logistics.

PUBLIC INTEREST STATEMENT
This study provides useful indications of how government organizations work to ensure knowledge sharing behavior in organizations works effectively. While the relationship among bureaucratic culture, empowering leadership, knowledge sharing behavior and affective commitment have been examined independently, few studies have investigated the association between the four concepts. This study extended prior research models and presents empirical evidence, which suggests that the relationship between bureaucratic culture and knowledge sharing behavior is mediated by the empowering leadership, especially in the Indonesian government sector relationship are expected to contribute to the theory and practice of the government sector.

Introduction
The knowledge-based economy represents a trend in an economy that has a high dependence on knowledge, information, and high level of accessibility in human daily activities. Since knowledge is not widely distributed within organizations, knowledge sharing behavior (KSB) among individuals and groups becomes a tool to increase organizational effectiveness. Knowledge sharing behavior is a critical tool for knowledge creation and supports the effectiveness of organizational learning to achieve performance achievement (Bartol & Srivastava, 2002). Previous studies have identified that KSBs have a role to improve a firm's performance, take a crucial role to sustain the competitive advantage, enhance human and structural capital, increase the financial and operational performance, contribute to knowledge application and innovation, and can be encouraging the employee creative problem-solving capacity (Davenport & Prusak, 1998;Wang & Wang, 2012).
The researchers highlighted that fostering KSB in the organization is not an easy task and is a challenge for organizations to maximize knowledge dissemination through knowledge sharing behavior among employees (Cleveland & Ellis, 2015). Organizational culture has gained wide recognition as a key to success as well as the barrier of KS (Riege, 2005;Lin, 2008;Zhang, De Pablos, & Xu, 2014;Cavaliere, Lombardi, & Giustiniano, 2015;Islam, Jasimuddin, & Hasan, 2015;Rafique & Anwar, 2019). The similarity between leadership and top management supports and rewards (Lin & Lo, 2015) empowering leadership (Srivastava, Bartol, & Locke, 2006;Xue, Bradley, & Liang, 2011). From organizational context, culture and leadership are two dominant factors in explaining the KSB within an organization. Leadership and organizational culture also play an important role in enhancing competitiveness and gaining sustainable advantage (Kargas & Varoutas, 2015).
Although many studies indicate that organizational cultures and leadership are important components of KSB, there are still some potential issues. First, the relationship between BC and KSB has been contradictory in recent literature. For example, Cavaliere et al. (2015) proved a positive effect of BC (stability, order, and efficiency) on KSB within organizations. Similarly, Islam et al. (2015) have found that centralization is positively related to knowledge sharing. Wang and Chen (2017) found that bureaucratic culture is the dominant factor that influences knowledge management. Other empirical evidence mentions competitive culture is negatively related to knowledge-seeking when mixed with bureaucratic culture (Baker & Ellis, 2018). Since the national culture has an impact on management styles, organizational culture, and employee behaviors, thus the incongruencies may have existed due to the limitations regarding regional differences (Hosftede, 1980;1993;House, Javidan, Hanges, & Dorfman, 2002). Furthermore, most studies of the relationship between leadership and knowledge sharing have focused on transformational leadership, thus potentially overlooking other important leadership concepts (Wu & Lee, 2017). Other studies have attempted to explain the relationship between empowering leadership and knowledge-sharing (Srivastava et al., 2006;Xue et al., 2011) in the business sector may differ when implemented in the government sector. In other words, it is still unclear how BC and empowering leadership can explain KSB in the government sector.
The present study contributes to explain the knowledge-sharing behaviors and affective commitment by focusing on a specific leadership style (empowering leadership) and specific organizational culture (bureaucratic culture) within the government service organizations. In so doing, the contribution of this study is threefold. First, different results have been found with regard to the influence of bureaucratic culture on knowledge-sharing behaviors (Cavaliere et al., 2015;Islam et al., 2015;Wang & Chen, 2017;Baker & Ellis, 2018), this study aims to resolve such shortcomings. Second, study of the relationship among empowering leadership and knowledgesharing behavior (Srivastava et al., 2006;Xue et al., 2011), and empowering leadership-affective commitment relationships (Freire & Azevedo, 2015;Kim & Beehr, 2018) was conducted in the business sector, while this study expands the topic area in the government sector. Third, by testing three structural models, this study developed a detailed understanding of how bureaucratic culture may directly or indirectly affect knowledge-sharing behavior and affective commitment.
The tests of such relationship are expected to contribute to the theory and the practice of knowledge-sharing behavior in the government sector.

Literature review
2.1. Bureaucratic culture, empowering leadership, knowledge sharing, and affective commitment Schein (2004) reveals that organizational culture is a basic assumption discovered or developed by a group of people while they learn to solve problems, adapt to the external environment, and integrate well to solve the problem in the organization. In this study, bureaucratic culture is seen as a unique feature of government organizations. The generic features of the bureaucratic culture are: management style is relatively authoritarian, a high degree of control, top-down communication, individuals search for stability, limited initiatives, and centralized decision making (Claver et al., 1999). Contrary to empowering leadership that facilitates in empowered work environments by enabling and encouraging workers in their work roles, including supporting, coaching, informing, and develop the participative decision-making to enhances the meaning and significance of work (Raub & Robert, 2010). Jogulu (2010) confirmed the organizational culture related to leadership style, and there are significant differences between leadership style and culture of the group. In a specific organizational culture dimension, Taormina (2008) has found that leader behaviors to be more control-oriented in bureaucratic culture. This indicates that the bureaucratic culture is in line with the behavior of controloriented leaders. Based on the empirical evidence, the hypothesis is proposed: H1. Bureaucratic culture is negatively related to empowering leadership Knowledge sharing behavior is defined as individuals sharing information and work-relevance experiences within the organization (Davenport & Prusak, 1998;Zhang et al., 2014). Organizational culture is regarded as knowledge sharing barriers as well as drivers, individual factors such as lack of interpersonal skills, motivation, self-efficacy; or organizations such as structures, processes, and systems in the organization, or technology are the primary components of knowledge sharing support (Lin, 2008;Riege, 2005). Some empirical studies show that organizational culture has a relationship with knowledge-sharing behavior (Chen & Cheng, 2012;Islam et al., 2015). Based on previous studies, supportive behaviors through actual support from top management were expressed as an important factor for developing employee knowledge-sharing behavior (Connelly & Kelloway, 2003;Peltokorpi, 2006;Li, p. 2010), and develop a culture that encourages employees to attempt innovation. Thus the bureaucratic culture is more appropriate to be a barrier than the drivers of knowledge sharing, so the hypothesis proposed: H2. Bureaucratic culture is negatively related to knowledge sharing behavior Meyer, Stanley, Herscovitch, and Topolnytsky (2002) define organizational commitment as the degree that employees identify with the goals and values of the organization and is willing to exert the effort to help the success of the organization. Allen and Meyer (1996) proposed a threedimensional construct of organizational commitment basis, namely affective, continuance and normative commitment. Affective commitment refers to the employee's emotional attachment, identification, and their involvement and emotional connection with the organization. Continuance commitment refers to rational consideration of employees regarding costs when staying or leaving the organization. Normative commitment refers to a sense of responsibility from employees to serve based on internalizing the values and goals of the organization that are inherent in themselves (Allen & Meyer, 1996). In this study, affective commitment is used as an independent construct of the other dimensions of organizational commitment.
Many studies examine the relationship between organizational culture as general and employees commitment. For example, Lok and Dan Crawford (2001) concluded that organizational culture is an important role to improve employee commitment. In particular, researches in various industries and countries show that an innovative and supportive culture have a strong positive effect on commitment, while the bureaucratic culture has a negative impact, and most authors agree that affective commitment depends on organizational culture (Lok & Crawford, 2004;Lok & Dan Crawford, 2001;Silverthorne, 2004;Yiing & Ahmad, 2009). The present study aims to investigate the relationship between bureaucratic culture and affective commitment in the government service unit. Hence, the following hypothesis in this study is proposed: H3. Bureaucratic culture is negatively related to affective commitment

Empowering leadership, knowledge sharing behavior and affective commitment
There are numerous works are the antecedents of knowledge sharing behavior from organizational dimensions, including organizational structure, management support and rewards (Lin, 2007;Islam et al., 2015), empowering leadership (Srivastava et al., 2006;Xue et al., 2011), and culture (Riege, 2005;Lin, 2008;Zhang et al., 2014;Cavaliere et al., 2015;Islam et al., 2015;Rafique & Anwar, 2019). Specifically, in empowering leadership, Xue et al. (2011), Srivastava et al. (2006, and Kim and Beehr (2018) found that knowledge sharing behavior is positively significant related to empowering leadership. This study argues that empowering leadership can increase the sense of identification (Liu & Shieh, 2015), emotional closeness and the sense of responsibility for the decisions that have been made. This condition then encourages employees to have a concern and intention to share workrelated information with colleagues to achieve effective organizational, thus the hypothesis proposed: H4: Empowering leadership is positively related to knowledge-sharing behavior Empowering leadership is a condition created by leaders to share power, giving decision-making autonomy to subordinates, expressing employee confidence's capabilities, and elimination of performance barriers (Zhang & Bartol, 2010). The empowering leadership has a direct effect on empowerment (Kwak & Jackson, 2015), and also may have indirect effects on affective commitment and turnover intentions (Albrecht & Andreetta, 2011). A study by Freire and Azevedo (2015), demonstrated that the empowering work context was a significant effect on affective commitment and trustworthiness of the supervisor. For the affective commitment, most studies show that affective commitment depends on leadership (Lok & Crawford, 2004). Hence, the following hypothesis is proposed: H5: Empowering leadership is positively related to affective commitment Based on the social exchange theory (Blau, 1967), this study proposes that when employees have access to a set of empowering structures, such as information, resources, support, and opportunities, it can make the employees act by responding favorably when relating to the organization, through a greater affective commitment. In this sense, this study proposes that empowering leadership will stimulate knowledge sharing behavior and affective commitment, and empowering leadership can act as mediation in the relationship between organizational culture and affective commitment. Hence, proposes the following additional hypotheses: H6: Empowering leadership mediates the effects of organizational culture on knowledge sharing behavior and affective commitment 3. Methodology

Sample and procedure
The respondents of this survey are employees from eleven different government service units in a province of eastern Indonesia. This study was conducted in three stages to ensure the aims and objectives are clear within a formal framework for methodological rigor. The first stage involved eleven heads of local government public service units and experts in human resources management as the early stages of scale development. The main purpose of this stage is to check for the relevance and validity of the questions, and any ambiguities.
Based on the feedback given, the changes are incorporated into the final questionnaire. These activities are carried out through focus group discussion as the process of item-generation and content validity. Each expert judgment provides ratings based on their opinions and Aiken's V is used to calculate the content-validity coefficient based on the assessment results of the expert panel. The items valued as highly relevant will be included in the final test for organizational culture scale. The pilot testing is to identify those items that most clearly represent the content domain of the underlying construct using exploratory factor analysis (EFA). A total of 150 participants were used in this stepas suggested by Hair, Anderson, Tatham, and Black (1998) to apply factor analysis. The final stage is to analyze the relationship between variables through structural equation modeling (SEM). A total of 371 participants were used for the final survey. Fifty-five percent of the respondents were male, and most had at least a diploma or bachelor's degree (53%). The average of respondents was 36 years (s.d = 9.03).

Measurement
The development of bureaucratic culture scale refers to the stage of scale development by Hinkin (1997) including item generation, content adequate assessment, questionnaire administration, factor analysis, internal consistency analysis, and construct validity. An initial pool of approximately seven items gathered from the literature. Content adequate assessment using experts in a content domain and leaving 5 items are considered most relevant. Items comprising the bureaucratic culture scale are: "centralized in decision-making", "high degree of control", "top-down communication", "individuals search for stability", "change-resistant". The participants were to rate the items on a five-point Likert-type item where 1 represents ''strongly disagree" and 5 represents ''strongly agree".Cronbach's alpha coefficient for the present total sample was 0.79, thus acceptable internal consistency (Nunnally & Bernstein, 1994). From these 5 items, the factor analysis extracted one factor (unidimensional), and loading is greater than 0.50 (See Table 1).
Empowering leadership adapted and modified from Arnold et al., (2000) with the subscales of participative decision making (three items) and informing-coaching (three items). All of the items were evaluated by a five-point Likert scale where 1 represents "'strongly disagree'" and 5 represents "'strongly agree.'" As a first step in the analyses, the dimensionality of the measurement model was tested with exploratory factor analysis with varimax rotation. Table 1 shows that all items extracted one factor (Cronbach alpha = 0.78). The Cronbach alpha values of were acceptance above the cut-off point 0.70.
Knowledge sharing behavior was adapted 3-item from knowledge collecting subscale by Lin (2007). Each of the 3 items is rated on a 5-point scale from 1 "strongly disagree" to 5 "strongly agree". Lin et al., (2007) reported the composite reliability is 0.80 in her study, and that is not different from this study (Cronbach alpha = 0.83). Affective commitment (AFC) was measured with the 6-item affective commitment scale from Rhoades, Eisenberger, and Armeli (2001). Scale anchors ranged from 1 (strongly disagree) to 5 (strongly agree). Examples of the items asked are: "I feel a strong sense of belonging to my organization", and "I feel personally attached to my organization". Cronbach Alpha Coefficient for the present total sample was 0.81.

Method of analysis
Data analysis in this study was performed using structural equation modeling (SEM) to validate the research model. SEM is a multivariate statistical analysis tool that provides researchers with a thorough method for the examination and theories testing (Joreskog & Sorbom, 1996). This approach was chosen because of its ability to estimate direct and indirect relationships between variables and relationships between latent variables. The measurement model (confirmatory factor analysis) was examined for instrument validation, followed by an analysis of the structural model for testing associations hypothesized in the research model. Table 2 shows descriptive statistics and correlations for all the latent variables. As expected, bureaucratic culture (BC) was negatively correlated to empowering leadership (EMP), knowledge sharing behavior (KSB), and affective commitment (AFC). Also, empowering leadership was positively related to knowledge sharing behavior and affective commitment. Compared to bureaucratic culture, empowering leadership had the stronger correlation between knowledge sharing behavior and affective commitment.

Measurement model
Model fit evaluation determines the degree to which the sample variance-covariance data fit the structural equation model. The criteria commonly used are chi-square (χ2), the goodness-of-fit index (GFI), the adjusted goodness-of-fit index (AGFI), and the root-mean-square residual index (RMR) (Joreskog & Sorbom, 1996). The Chi-Square = 204.70 (df = 159, p-value = 0.0085), this model-fit index indicates that the model is unacceptable (Schumacker & Lomax, 2010). However, since the chi-square model-fit criterion is sensitive to sample size (when the sample size increases, generally above 200), the χ2 statistic tends to indicate a significant probability level (Schumacker & Lomax, 2010), so evaluation of model fit can be considered with other criteria. The Goodness of Fit Index (GFI) = 0.95; Adjusted Goodness of Fit Index (AGFI) = 0.93; and Comparative Fit Index (CFI) = 0.99, they exceeded the recommended cut-off level of 0.9 (Bagozzi & Yi, 1988). The root-mean-square error of approximation (RMSEA) is equal to 0.028, which is below the typical acceptable level of model fit (criterion RMSEA < 0.08 or 0.05), and Standardized RMR = 0.046 (< 0.05). The combination of these results suggested that the measurement model exhibited a good level of model fit.
Construct validity involves providing psychometric evidence of convergent validity and discriminant validity. Convergent validity is the degree to which items in the constructs truly represent the intended latent construct and indeed have the correlation with other measures of the same latent construct (Hair et al., 1998). Convergent validity was assessed by using the factor loadings, average variance extracted (AVE) and composite (Bagozzi & Yi, 1988). Discriminant validity is demonstrated when the respective average variance extracted is larger than the squared correlation between the two constructs (Fornell & Larcker, 1981).
The measurement model with all four constructs was assessed using confirmatory factor analysis (CFA) show that all factor loadings exceed 0.5 and each indicator was significant at 0.05 levels. The results show that the composite reliability of all constructs exceeded the benchmark of 0.7 recommended by Nunnally and Bernstein (1994). In terms of average variance extracted, all constructors exceed the suggested value of 0.5 (Bagozzi & Yi, 1988), indicating the measure has adequately convergent validity. The comparison between the average variance extracted for each construct (diagonal elements) and squared correlations of two constructs (off-diagonal elements) show that all constructs have high discriminant validity. In summary, the measurement model demonstrated adequate reliability, convergent validity, and discriminant validity. Figure 1 presents the estimates of both measurement and structural parts of the hypothesized model. Path coefficients and the overall model fit the supposed structural model regarding the strength and relationship among the variables. Turning to empower leadership as the endogenous variable, the squared correlations (R2) was 0.04, indicating that nearly four percent of the empowering leadership variations were explained by bureaucratic culture. The second substructural, the knowledge sharing behavior as the endogenous variable show that squared correlations (R2) was 0.038, indicating that 3.8 percent variations of knowledge sharing behavior were explained by empowering leadership and bureaucratic culture. The last sub- Notes: Correlations without any asterisks are not significant; ** p < 0.01 structural, affective commitment as the endogenous variable show that squared correlations (R2) was 0.30. This result indicates that nearly 30 percent of the variations of the construct were explained by the proposed model. Table 3 shows the direct, indirect, and total effects and their level of significance between predictor and dependent variables in the final model. The standardized path coefficients (SPC) revealed in the analysis suggested that each of the hypothesized path was in the expected direction. Specifically, the bureaucratic culture had a significantly negative effect on empowering leadership (SPC = −0.20, t-value = −3.18), confirming the result revealed in the previous studies (H1 supported). As hypothesized, empowering leadership exerted a direct effect on knowledge sharing behavior (SPC = 0.19, t-value = 2.95), but the results show that bureaucratic culture had no significant relationship with knowledge sharing behavior (SPC = −0.008, t-value = −0.14). Thus, hypotheses H4 was supported, and H2 was not supported.

Structural model
Both bureaucratic culture and empowering leadership were found to be significantly associated with affective commitment, but magnitudes of the impacts varied. The direct impact of bureaucratic culture on affective commitment was a significantly negative effect on knowledge sharing behavior (SPC = −0.35, t-value = −5.23), and the effect of empowering leadership on affective commitment was significantly positive (SPC = 0.35, t-value = 2.95), supporting hypotheses H3 and H5.
Finally, the indirect effects in the structural model were further examined to gain in-depth insights into the knowledge sharing behavior and affective commitment. The indirect effect of bureaucratic culture on knowledge sharing behavior via empowering leadership was found to be significant and negative (β = −0.04, t-value = −2.19). Also, the result showed that the indirect effect of bureaucratic culture on affective commitment via empowering leadership was also significant with β = −0.07 (t-value = −2.86). Thus, the impact of bureaucratic culture on affective commitment was linked indirectly through empowering leadership. (H6a and H6b supported).

Discussion
This study examines the interrelationships between bureaucratic culture, empowering leadership, knowledge sharing behavior and affective commitment in the context of government organizations in Indonesia. This study indicates that bureaucratic culture was a negative significant relationship with empowering leadership and affective commitment. This result implies that when employees feel that a stronger bureaucratic culture is applied, employees feelings for empowering leadership behavior are reduced. Additionally, bureaucratic culture also reduces affective commitment. The higher the employees' feelings on bureaucratic culture, the lower the affective commitment of the employees.
This study supports previous studies such as Sigler and Pearson (2000) and Taormina (2008) on the relationship of organizational culture to empowering, and in line with Lok and Dan Crawford (2001); Lok & Crawford (2004)), Silverthorne (2004), Taormina (2008); Yiing and Ahmad (2009) on the relationships between organizational culture and affective commitment. In contrast to expectations, bureaucratic culture empirically has no significant effect on knowledge sharing behavior. Although analytical results show that most respondents provide low ratings on bureaucratic culture, the inclination of knowledge sharing behavior is inconsistent at high and low ratings. These results do not support the findings of previous studies (Cavaliere et al., 2015;Islam et al., 2015;Wang and Chen (2017) and emphasize that national culture bias has an impact on employee behaviors (Hosftede, 1980;1993;House et al., 2002).
As expected, empowering leadership plays a role to improve knowledge sharing behavior and affective commitment. Consistent with previous studies (Srivastava et al., 2006;Xue et al., 2011), these findings found that empowering leadership play a unique role in influencing employees' desire to share work-related knowledge with colleagues within an organization. This result implies that employees are pleased in leadership behaviors that empower them and thus help others tend to be more motivated to knowledge sharing with colleagues. This condition can be a strategic finding for government organizations that want to increase the effectiveness of knowledge management through knowledge sharing between employees. Additionally, a sense of the empowered by leadership can be a requirement for employees to have a higher affective commitment. These results are consistent with previous studies which establish a relationship between empowerment and affective commitment (Albrecht & Andreetta, 2011;Freire & Azevedo, 2015;Kwak & Jackson, 2015), and support the empirical study of relationship between empowering leadership with knowledge sharing behavior (Srivastava et al., 2006;Xue et al., 2011).
Finally, the results indicate that empowering leadership can serve as a mediator of bureaucratic cultural relationships with knowledge sharing behaviors and affective commitment. In contrast to the direct relationship model, the indirect effect test shows that bureaucratic culture has significant influences on knowledge sharing through empowering leadership. The findings suggest that the model of bureaucratic cultural relationships is indirect via empowering leadership.

Conclusion and future recommendations
Based on the investigation, it can be stated that bureaucratic culture is directly related to empowering leadership and affective commitment, but not to knowledge sharing behavior. And also, empowering leadership plays a role to improve knowledge sharing behavior and affective commitment. A management should recognize that the application of bureaucratic culture with various attributes such as low initiative in decision-making, lack of new ideas, top-down communication models, individuals search for stability, and limited scope for the initiative is not an effective way to manage positive attitudes and behavior, specifically knowledge sharing behavior and affective commitment. Knowledge sharing and affective commitment can be enhanced through empowerment by providing broader opportunities in decision making, expanding autonomy, giving employees the flexibility to convey new ideas, implementing two-way communication, and encouraging individuals to work and expand employee scope for initiative. Bureaucratic culture has a negative effect on empowering leadership and affective commitment. The present study provides an important contribution as the latest empirical evidence regarding the relationship of bureaucratic culture, empowering leadership, and affective commitment in the context of government organizations. The study also provides unique evidence which supports that the bureaucratic cultural relations are fully mediated by empowering leadership. Therefore, this research model can be used as a useful reference for future research in the topic area of bureaucratic culture, empowering leadership, knowledge-sharing behavior, and affective commitment in the context of government organizations, especially in Indonesia.
Future studies should focus on three areas to overcome the limitations of the present study. First, the current study is focused on relationships of variable studies among government public service agencies in Indonesia. The selection of government organization settings in a single country study may raise concerns about generalizability. Therefore, replication studies in the different location and various private and public sector in Indonesia are needed for crossvalidating the findings reported here and for further generalizations. Second, this study only uses organizational factors (bureaucratic culture and empowering leadership) as antecedent knowledge sharing behavior. Future research can examine the individual factors, including motivation, work experience, training, job autonomy and self-efficacy (Lin, 2007), goal orientations (Kim & Lee, 2013), and technology/ICT (Lin, 2007). Finally, this study adopts the cross-sectional design that examines staff level at one point in time. This design has a weakness in explaining the causality relationship between variables. Thus, future studies can gather longitudinal design to examine the causality and interrelationships between variables that are important to knowledge sharing behavior and affective commitment.