Organizational commitment (OC) means the desire to achieve the goals of the organization and to establish a long-term relationship with the organization (Hussain, Ali, Khalid, Shafique, & Ahmad, 2016). It predicts workers’ behaviour and attitude, which than directly impacts firms’ competitiveness and financial performance (Bashir, & Ramay, 2010; Khan, & Zafar, 2013). It is found that those employees whose commitment level is high are more inclined towards organization (Danish, Ramzan, & Ahmad, 2013) and their state of absenteeism and turnover decreases (Vangel, 2011).

In addition, the three-component OC model (Meyer, & Allen, 1991) is used to predict high labour productivity such through increased OCB, high job performance, reduced Turnover rate etc. (Hussain et al., 2016). These three components of OC are continuous, normative and affective commitment.

Continuous commitment is defined as “workers’ affiliation with the organization due to the greater cost of disassociation” (Mengenci, 2015, p.147). Normative commitment is about “employees’ willingness to sustain association with the firm due to perceived obligation towards the firm” (Fard, & Karimi, 2015, p.147).

In a Chinese context, AC is considered the most important element in the three-component OC model, just as in the Western context (Chen & Francesco, 2003). The findings of their study suggest that AC is the most relevant element as compared to other components. Therefore, this study emphasises the AC,) three-component model based on Meyer and Allen (1991 and aims at examining the antecedents of AC in the service sector such as banks of Khyber Pakhtunkhwa, Pakistan.

In the banking sector of Pakistan, the literature shows that different climatic factors such as authority, trust, leadership style and organizational support are responsible for high commitment (Hassan, Bano, Shaukat, & Nawaz, 2013). Additionally, Abdullah and Ramay (2012) highlighted the significance of AC in Pakistan’s banking sector, whereas Khan and Zafar (2013) reported demographics and personal factors like age, tenure and management level as the antecedents of

AC in the banking context. Similarly, working environment, compensation plan, employees’participation in decision making and job safety are considered the most significant predictors of AC (Milliken, Morrison, & Hewlin, 2003).

Therefore, the aim of this study is to enhance the theoretical base and contribute to the literature on AC and its antecedents. Furthermore, this research contributes empirically by examining the variables in new context of service sector such as different banks situated in different area of Khyber Pakhtunkhwa, Pakistan.

Lack of Leader Openness to Voice and Affective Commitment

Leader openness to voice is described as leader’s ability to encourage employees to exchange information, ideas and work-related issues (Milliken et al., 2003). Leader with the positive attitude towards employees’ voice listens, motives, assist, supports and allow employees to provide suggestions during decision making process while making them feel safe to talk in front of the top level (Detert & Edmondson, 2011; Lu & Xie, 2013).

Lack of Open Communication Opportunity and Affective Commitment

Open communication opportunity is described as the open communication channels available for the free flow of information exchange with the managers and colleagues (Lu & Xie, 2013). In the presence of open communication channels employees feels empowered to provide useful input during the decision-making process, which than increases their level of motivation, trust and commitment towards the organization (Dedahanov, Kim, & Rhee, 2015; Deniz et al., 2013).

Defensive Norms of Organizational Culture and Affective Commitment

Organizational culture states to a team's emotional process of matching their interests, problems and results to their work (Bogosian & Stefanchin, 2013). Similarly, Mengenci (2015) findings shows that conductive norms in organization promote positive climate and workers will feel safe and will express their ideas. If the climate is not good, then it will decrease commitment.

Defensive Silence and Affective Commitment

(DS) is defined as an intentional withholding of relevant information, issues, suggestions and ideas due to the risk associated with speaking up (Beheshtifar, Borhani, & Moghadam, 2012; Brinsfield, 2013; Morrison, 2014). It is found that employees adopt defensive silence behaviour due the fear of facing negative outcomes.

However, many studies in past has been conducted on aforementioned area. Though, the relationships between these elements are under investigated in banking sectors of Pakistan (Milliken et al., 2003; Panahi et al., 2012; Ponnu & Chuah, 2010; Vakola & Bouradas, 2005).

Conceptual Model

H1:Lack of leader openness to voice adversely associates with affective commitment.

H2: There is negative effect of Lack of open communication opportunity on affective commitment.

H3: There is negative effect of defensive norms of organizational culture on affective commitment.

H4: Defensive silence negatively influences affective commitment.

Research Setting, Participants and Procedures

This research was carried out in banking sector of KP province of Pakistan, and data was collected in two steps. Stratified proportionate random sampling technique was used and sample of 1236 workers of 258 different branches of 8 conventional banks in 12 different districts of KP was drawn. Whereas, in the second qualitative phase, convenience sampling was used to draw sample of 24 informants, 2 from each 12 districts.

Instrument and Measurement

The questionnaire was developed using five-point Likert scale in the first quantitative phase. Five items of LLV and five items of LOC were adopted from the instrument developed by Vakola and Bouradas (2005) whereas, DNO was measured through four items extracted from the Organizational Culture Inventory (OCI) of Cooke and Rousseau (1988). Five items of DS and eight items of AC were adopted from (Dyne et al., 2003) and Meyer and Allen (1991) scale respectively.

Result and Analysis

Results for the Quantitative Phase I

Reliability and Validity

Reliability analysis depicts that the coefficient values for each scale (LLV = .876, LOC = .766, DNO = .716, DS = .846, AC = .785) are greater than 0.70 and AVE as well CR value also greater than 0.5 and 0.7 correspondingly. (LLV: AVE = .715, CR = .873; LOC: AVE = .822, CR = .924; DNO: AVE = .632, CR = .791; DS: AVE = .759, CR = .866; AC: AVE = .865, CR = .829).

Descriptive Statistics

Descriptive statistics of the study variables reports the highest mean value of 3.77

(SD = 0.74) for LLV followed by LOC 3.39 (SD = 0.71). Moreover, DS presents the mean value of 3.13 (SD = 0.82) whereas the mean value for AF is reported as 2.70 (SD = 0.64) while DNO presents the lowest mean value of 2.40 (SD = 0.60).

Inter-correlation investigation was accompanied to inspect the association between the affective commitment and its antecedents and shown in Table 1.

**p<.01

Result of Table 1 illustrates the highest negative correlation between LOC and AC (r = -0.48) as compared to LLV (r = -0.44), DS (r=-0.39) and DNO (r = -0.23). Hence, hypothesis 1, 2, 3 and 4 were supported by the results of the study.

Structural Equation Modelling

To validate that LLV, LOC, DNO and DS impacts AC, structure equation modelling was utilized. The results of standardized estimates obtained through SEM are depicted in table 2 which revealed that LLV has negative and significant effect on AC such as when LLV increase one unit on the other side AC decrease 0.14 units. Furthermore, LOC (-0.56), DNO (-0.38) and DS (-0.31) shows a negative association with AC, whereas p-value (0.00) indicated the significance of these relationship.

: Regression Standardized Estimates between Climatic Components and Affective Commitment

Moreover, the result of model fit indices shown in Table 3 below, depicts that model does not meet the minimum condition for the model fit due to the chi-square (4937.296), degree of freedom (270) and its associated P value (.000). Root mean square residual (RMR) is 0.073 which is less than 0.08, so model shows good fit. GFI and AGFI value shows not good model fit as the range of both are less than 0.90. The value of CFI is .930 which is greater than the threshold and considered as good model fit. And at last, the value of RMSEA is 0.068 which is less than 0.80 and considered that data fit the model (McDonald & Ho, 2002).

Table 3. Model Fit Indices

Data Analysis and Result of Qualitative Phase II

Result of Qualitative Phase II

The result of the quantitative phase revealed that there were no insignificant or outlier results however all hypothesis had significant findings. Thus, it was decided to explore the reasoning for the strongest and weakest predictor of affective commitment in the banking sector of KP, Pakistan. Thus, thematic coding and causal networking was used to derive themes for explaining the reasons for such relationship to exist among the study variables.

The second part of the qualitative study, semi-structure interview guide was established to get knowledge of how and why these particular associations arise in KP banking sector.

Through the triangulation of data collected in these two phases two models for antecedents of AC in the banking sector of KP, Pakistan were developed.

: Relationship between LCO and AC: Banking Sector of KP Pakistan

The result of quantitative phase also depicts that DNO negatively relates with AC but act as the weakest predictor as compare to other climatic components. The following themes were drawn in the interview as the responsible factors explaining this association in the banking environment.

Incompetency of Colleagues

One of the most common complaint made by the informants was incompetency of co-workers. They mentioned that the team work is adversely affected by the ineffective and incompetent team members who do not participate in the teamwork or else produce inefficient quality of work. Informants reported that they are forced not to raise voice against this issue due to the prevalence of defensive norms and fear of social rejection by the other colleagues. Hence, it decreases their level of commitment.

Negative Experience of the Colleagues

Informants mentioned that the negative experience of colleagues/ seniors associated with voice behaviour repels them from speaking up. When co-workers express their frustration about the negative attitude of top management towards voice and threaten others about the punishment and negative consequences, ultimately defensive norms originate and thus level of commitment decreases.

Lack of Leader Consultation Attitude

The other issue highlighted by the informants was lack of consultation by the supervisor. They reported that managers consider our silent as consent and do not encourage us to participate in decision making and we are unable to suggest any innovation idea or solutions to our problems. This lack of consultation promotes defensive norms to remain silent and reduces AC.

Social Boycott

Informants mentioned that if we get to know about any fraud case by the other colleagues or the managers, or any other kind of illegal activity we mostly remain silent and do not speak up the hierarchy due to the fear of social boycott and isolation. We also fear that top management might not trust our information and we may get social rejection by the co-workers and managers. Thus, remain silent for the sake of self-defence which adversely affects AC.

Avoiding Conflicts

Moreover, informants shared that for the sake of avoiding conflicts they mostly remain silent and do not talk about controversial issues. Instead of developing conflicts with others they prefer to remain silent and this become a common set of norms followed by everyone, but AC reduces.

Groupthink

Groupthink is another very important factor highlighted by the informants which effects the work environment, and everybody is forced to follow common norms to remain silent. This pressure created by groupthink negatively effects AC.

These themes that are derived from the informant’s feedback, to predict the relationship among DNO and AC, are presented in fig. 3.

Relationship between DNO and AC: Banking Sector of KP Pakistan

The aim of this study was to determine the antecedents of AC in banks situated in different area of KP. The first quantitative phase strongly supports all hypotheses and the results of the study are aligned with the prior studies (Nikmaram et al., 2012; Vakola & Bouradas, 2005).

Hypothesis 1 states that LLV negatively impacts AC. Previous research work also supports this argument by stating that negative attitude of supervisor towards voice behavior leads to low level of AC (Nikmaram et al., 2012; Vakola & Bouradas, 2005). Moreover, hypothesis 2 of the study reveals that LCO decreases AC. While supporting this statement, previous studies mentioned that in the absence of open communication channel employees do not get an opportunity to exchange information with the top management and the co-workers, hence their AC reduces (Panahi et al., 2012). Hypothesis 3 of the study found that there is a negative correlation between DNO and AC. In support of this argument Morrison (2014) stated that common set of defensive norms creates a collective sense of not to raise voice to avoid negative consequences, which ultimately results in low level of AC. In the same way, the results also support H4 which shows that there is negative association between defensive silence and affective commitment. Previous research results also aligned to this finding (Deniz et al., 2013; Laeeque & Bakhtawari, 2014).

The second qualitative phase of the study concentrate on further exploring the findings of quantitative phase to better understand how and why LOC considered the strongest predictor while DNO to be the puniest predictor of AC in banks operating in KP.

Limitations of the Study

Although there are significant contributions of the study, it also comprised of some limitations. Only one dimension of organizational commitment i.e., affective commitment was the main focus of the study whereas other dimensions were ignored. Also, the study highlighted few of the antecedents of affective commitment in the banking context. Cross-sectional data was collected during both quantitative and qualitative phase of the study. Furthermore, the study focusses on banks in general while ignoring the sector discrimination as public and private banks.

Future Research Guidelines

A comprehensive study could be conducted while incorporating all three dimensions of organizational commitment to better understand the phenomenon. Additionally, some other important antecedents of organizational commitment such as compensation, motivation, job satisfaction etc. could be added to study the variable in the banking context. Similarly, longitudinal data and/or experimental design could be incorporated to further explore the concept. Moreover, a comparative study between public and private sector banks could be investigated to further generalize the results.