Predictors of affective commitment at municipalities in the Nkangala district, Mpumalanga

Orientation: Less committed employees have a huge impact on organisational performance. The affective commitment of municipal employees is critical for a municipality to achieve its strategic objectives. Research purpose: To determine the predictors of affective commitment at the four municipalities in the Nkangala district, Mpumalanga. Motivation for the study: There is a shortage of studies conducted at South African municipalities that have simultaneously measured negative emotions, positive emotions and job satisfaction as the predictors of affective commitment. Research approach/design and method: The approach was quantitative and cross-sectional in nature. A closed-ended questionnaire was developed and administered to 808 respondents who were conveniently selected from four municipalities in the Nkangala district. The data were analysed using descriptive statistics. Inferential statistics were used to determine whether the predictors and affective commitment were unidimensional and to determine the relationships and the highest predictor. Main findings: The results showed that job satisfaction was the greatest predictor of affective commitment and negative emotions were the lowest predictor. Job satisfaction, positive and negative emotions explained 68% of the variance in affective commitment. Practical/managerial implications: This research will empower municipality managers in cultivating a conducive work environment. Improving the interpersonal skills of supervisors and implementing fair human resource practices that promote a conducive work environment have been highlighted as some contributors towards increased affective commitment. Contribution/value-add: This study has developed a two-path framework that could be used by human resource management practitioners to determine the affective commitment of employees in Mpumalanga municipalities.


Introduction
Research on affective commitment at municipalities is gaining traction in the human resource literature (Bright, 2007). Kim (2008) found that committed managers at Asian municipalities had the ability to encourage newly appointed employees to perform exceptionally well. Research conducted at European municipalities reported that managers who created conducive working conditions and gave employees constructive feedback enhanced employees' commitment levels (Steijn, 2008). In the same vein, it has been shown that satisfaction is a predictor of affective commitment at African municipalities (Chimwaso, 2000). However, employees from South African municipalities have shown less commitment (Bizana, Naude, & Ambe, 2015).
One of the indicators that showed the lack of commitment at one municipality was complaints from customers about the quality of service they received. Kanyane (2006) discovered that the majority of employees employed by one of the South African municipalities were less committed because of lack of growth and promotional opportunities. It was further established that only politically connected employees without appropriate qualifications were appointed in senior positions, even when they were not performing well. This had the unintended consequence of skilled workers leaving the municipality (Mailovich, 2019). Generally, irregular appointments adversely affect the affective commitment of employees at South African

Purpose of the study
Prior to this study, no similar study had been conducted at municipalities in the Nkangala district. It was against this background and to address this gap that the study was conducted. The main purpose of the study was to determine the predictors of affective commitment at four municipalities in the Nkangala district. The research objectives of the study were as follows: • to determine whether the predictors (refer to Figure 1) and affective commitment were unidimensional • to determine whether there was a significant relationship between the predictors • to determine whether positive and negative emotions mediated the relationship between the predictors and job satisfaction • to determine, between job satisfaction, negative emotions and positive emotions, which predicted affective commitment the most. Meyer and Allen (1991) defined affective commitment as an employee's emotional attachment to identification with, and involvement in an organisation. Similarly, Benjamin (2012) saw it as a customer's emotional attachment to a particular brand or strategic objectives. Affectively committed employees are less resistant to organisational change (Morin et al., 2016), have low rates of absenteeism, support human resource management strategic objectives, are team-oriented and are less likely to resign (Ditlev-Simonsen, 2015).

Theoretical framework and hypotheses
In this study, affective events theory (AET) was used as it was deemed appropriate for the study background. According to Wegge, Van Dick, Fisher, West and Dawson's (2006) study, which used AET as a theoretical model, it was established that affective commitment was predicted by different affective variables (e.g. autonomy, supervisory support, participation, employee welfare and work overload). In the same vein, affective variables and affective commitment are distinguishable (Wegge et al., 2006). Such hypotheses have been confirmed by two other studies conducted at one municipality in Mpumalanga (Maleka et al., 2016b). Several studies have tested the relationship between the study variables, and the hypotheses developed from these studies for the research reported on in this article are shown (Table 1).
Autonomy, participation, workload, well-being and employee performance were the independent variables. Positive emotions, negative emotions and job satisfaction were the mediating variables. Affective commitment was the dependent variable. The sections that follow briefly elucidate the relationships amongst the variables.

Autonomy
Autonomy refers to the freedom employees require to execute their tasks and to make decisions without or in limited consultation with their supervisors and/or line managers (Robbins & Judge, 2019). Previous research at a municipality in Mpumalanga showed that employees who were given autonomy to execute their job functions were happy and were not disgruntled . Other scholars have emphasised that municipal management should empower employees to make independent decisions (Brunetto, Farr-Wharton, & Shacklock. 2011).

Participation
Participation involves consulting and engaging employees in the decision-making surrounding no matter what directly or indirectly affects them (Lee & Kim, 2011). Previous research at a municipality in Mpumalanga found that employees who were able to control the scheduling of their daily tasks were the most likely to be satisfied and happy (Maleka et al., 2016b).

Supervisor support
According to Kim, Hur, Moon and Jun (2017), supervisory support refers to employees' perception about support from their superiors, their provision of good advice and the ability to support employees in work-related challenges. Supportive supervisors who provide guidance and treat subordinates fairly and ethically are most likely to enhance the job satisfaction and affective commitment of their employees (Mohsin et al., 2015).

Employee performance
In this study, the focus of employee performance is feedback. This is defined in the literature as performance appraisal (Campbell, McHenry, & Wise, 1990). Even though this variable was not included in Wegge et al.'s (2006), in this study, it was included because earlier it was discussed that employees in the municipalities are disgruntled about the way they are given feedback. Previously, it was established that employees who were given positive and objective feedback were more likely to display positive emotions and were satisfied (Stazyk, Pandey, & Wright, 2011).

Job satisfaction
Job satisfaction is one of the affective feelings that employees have about their job, the working environment and work procedures (Locke, 1976). Some of the work procedures that enhance employees' job satisfaction include, but are not limited to, the chances of promotion (Maleka, 2016;Riley, 2006), opportunities to do different tasks, job rotation, as well as commitment, effectiveness and efficiency at work (Akpinar et al., 2013).

Positive emotions
Positive emotions can be described as an emotional response towards certain events or objects. This response reveals one's positive feelings towards certain events or objects. For instance, evidence suggests that organisations that promote autonomy are likely to increase employees' positive emotions (Hospel & Galand, 2016). Furthermore, research has revealed  Adil and Kamal (2016) are of the view that negative emotions can be expressed through anger, fear, guilt and nervousness. Adil et al. (2013), Fida et al. (2015) and Wegge et al. (2006) found that negative emotions mediated the relationships amongst workload, well-being and job satisfaction. This means that, when negative emotions are included as a third variable, they enhance the impact of or the relationships amongst workload, well-being and job satisfaction.

Work overload
Excessive workload is likely to contribute towards employees' negative emotions. In this sense, employees with excessive workloads experience considerable difficulty in completing their tasks. Consequently, they experience higher stress levels and other health-related challenges because of unbearable work demands (Williams, 2010).

Well-being
The concept of well-being can be viewed through the concepts of physiological, social, psychological and emotional well-being (Lee & Peccei, 2007). One of the indicators of well-being is when employees come to work and are physically fit to do their task (Brunetto, Shacklock, Teo, & Farr-Wharton, 2014). Other indicators that enhance employees' well-being are social interaction and support from co-workers, managers and family (Dávila & García, 2012).
The methods followed to determine the predictors of affective commitment at municipalities in the Nkangala district, Mpumalanga are discussed next.

Research participants
Altogether 1000 questionnaires were printed and 808 were completed. Therefore, the response rate was 80.80%. The respondents were conveniently selected from four municipalities in the Ngankala district. The questionnaires were distributed and collected from December 2018 to February 2019. A total of 53.30% of the participants were male and 42.70% were female. A total of 70.30% were employed on a permanent basis and 29.70% were employed on a part-time basis. The majority (52.48%) of the participants had grade 12 and the other 47.52% had post-matric qualifications. The majority (60%) of the participants had ages ranging from 45 to 65 years and the other 40% had ages ranging from 18 to 44 years. In terms of positions, 87.38% were in non-management positions and 12.62% were in middle and top management positions.

Procedure
One of the researchers distributed and collected the questionnaires at the four municipalities. Prior to the distribution of the questionnaires, the researchers were given ethical clearance (FCRE:2018/FR/07/009-MS02). An informed consent form was attached to each questionnaire, and the participants completed the form before they  participated in the study. The study followed ethical principles such as voluntarism, confidentiality and anonymity. In terms of the latter, the participants were not coerced to complete information that could be linked to them. They were also requested not to include their names on the questionnaires.

Research instrument
The questionnaire used in this study had two sections. The first measured biographical information discussed in the participants sections. The second part of the questionnaire measured affective commitment, as developed by Meyer and Allen (1991) and its antecedents. The latter comprised the following: • autonomy, supervisory support and participation scales developed by Patterson et al. (2005) • job satisfaction scale developed by Spector (1985) • employee performance scale developed by Babin and Boles (1996) • workload scale developed by Riggs, Warka, Babasa, Betancourt and Hooker (1994) • well-being scale developed by Ryff (1989) • positive and negative emotions scales developed by Burke, Brief, George, Roberson and Webster (1989).
Each scale consisted of five items that were measured on a five-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree).

Statistical analysis
The descriptive statistics involved frequencies reported in the participation section and the associated mean. The analysis entailed calculating the measurement model. The average variance extracted (AVE) was calculated, and a cut-off of 0.5 was used to determine the construct validity (Hair, Hult, Ringle, & Sarstedt, 2017;Ravand & Baghei, 2016) of affective commitment and its antecedents. To achieve discriminant validity, the Fornell-Larker criterion was used. This criterion states that the diagonal correlation scores must be higher than the horizontal correlation scores (Hair et al., 2017). To test the direct and indirect effects, structural equation modelling (SEM) was utilised. The level of significance to accept the hypotheses was set at 0.05 or 5% in the SMART-PLS.

Results
As it can be observed, the kurtosis and skewness scores were within the range of -2 to 2 (Table 2). Tabachnick and Fidel (2014) claimed that such a distribution shows that the data were normally distributed. Only negative emotions had a skewness value of 2.47 and had a highest mean score (M = 4.34, SD = 0.74). This suggested that participants agreed that they were frustrated with the workload, felt angry when they were criticised, feared losing their jobs and were nervous when they submitted their reports. The variable that was rated the second highest was workload (M = 3.72, SD = 1.06). The rating can be interpreted to indicate that participants did not have enough time to execute their duties, perceived their jobs to be difficult and agreed that their workload increased because the municipal managers were not replacing employees who had resigned. The third variable that had a mean score above 3 was well-being (M = 3.53, SD = 1.05). This result showed that the participants agreed that they did not feel optimistic about their future in the municipalities had very high stress levels and did not feel confident about their jobs. The other seven variables had mean scores under 3, meaning that the participants did not agree with the items listed (Table 2).

Measurement model
The measurement model is presented (Table 3). All the loadings were above 0.7. As the AVE scores were above 0.5, it can be claimed that convergent validity was achieved (Hair et al., 2017) and that the study variables were unidimensional. The Cronbach's alphas (α) were above the 0.6 lower limit for acceptance as suggested by Maree (ed. 2016).
The data presented show that discriminant validity was achieved (Table 4). Fornell-Lacker was used to determine discriminant validity. This statistical test is about comparing 'the square root of each construct's average extracted with its correlations with all other constructs in the model' (Hair et al., 2017, p. 317). The average square roots are bold diagonal scores and since they are higher than the horizontally and vertically correlations it can be argued that there were no discrimination validity issues.

Hypothesis testing
As  values for the variables were less than 5, and according to Hair et al. (2017), this meant that there were no issues with multi-collinearity. The SMRR was 0.11, which was less than the 0.12 criterion to determine model good fit (Hair et al., 2017). The SEM model of the study is displayed (Figure 1).
Twenty-six hypotheses were tested, and the results are displayed (Table 5 and Table 6). The results for direct effects are shown (Table 5).
The results for indirect effects are displayed (Table 6).

Discussion
The study intended to determine the predictors of affective commitment at municipalities in the Nkangala district. The study made multiple contributions to the body of knowledge. Firstly, a methodological contribution was made in terms of the sample size. Because the sample for Skosana's (2016) was drawn from one municipality, the sample for this study was drawn from four municipalities in the Nkangala district.
Secondly, a theoretical contribution was made as employee performance, which had not been included in the study conducted by Wegge et al. (2006), was added as independent variable. This variable is salient in the municipality context because previous research has revealed that, when employees are given proper feedback, coached and given training opportunities, they are most likely to offer customers excellent service (Maleka et al., 2016b). The data also showed that this variable was the greatest predictor of positive emotions. The beta (β) score between employee performance and positive emotions was 0.56 and was significant at the 5% level. In addition, the data showed that employee performance was the greatest predictor of job satisfaction (β = 0.39, p < 0.05). These results are consistent with findings by Stazyk et al. (2011). The results can be interpreted as indicating that, when employees are given constructive feedback, their satisfaction level increases. Similar to findings by Wegge et al. (2006), the study results showed a positive and significant relationship between autonomy, supervisor support and positive emotion and job satisfaction. Unlike in Skosana's (2016) study, which was conducted in one of the municipalities in Mpumalanga, it was expectedly found that participation negatively predicted positive emotions. However, the results were not significant (β = 0.39, p > 0.05). When positive emotions were used as a mediator variable, the results were also not significant. This was in contrast to what Wegge et al. (2006) found.
The third contribution of this study was testing the relationship between workload and well-being and negative emotions, as this relationship had never been tested in Mpumalanga municipalities prior to this study. The results of this study were consistent with what Wegge et al. (2006) found. A positive and significant relationship was found between workload and negative emotions (β = 0.34, p < 0.05).
In addition, the relationship between well-being and negative emotions was positive and significant (β = 0.18, p < 0.05). Based on the study results, it can be deduced that workload was the greatest predictor of negative emotions. It can be interpreted as indicating that employees were uncertain about their futures and were not given enough time to execute their duties, which caused negative emotions.
When negative emotions were measured as a mediator between well-being and workload, the results were also consistent with findings by Wegge et al. (2006). Similar to Wegge et al. (2006), the researchers found that job satisfaction mediated the relationship between negative emotions and affective commitment. As expected, the relationship between negative emotions and job satisfaction was negative and significant (β = -0.14, p < 0.05). It is noteworthy that negative emotions were the lowest predictor of affective commitment. The data showed that the relationship between negative emotions and affective commitment was negative and significant (β = -0.07, p < 0.05).
On the other hand, even though job satisfaction was the greatest predictor of affective commitment, the difference between it and positive emotions was not large: β (0.468)β (0.446) = 0.022. This can be interpreted as indicating that employees are most likely to be committed when (1) they are happy and joyous, (2) they work in a conducive environment, (3) they experience fair promotion practices, (4) they are rotated and (5) they are given meaningful tasks. Consistent with Wegge et al. (2006), this study found that job satisfaction mediated the relationship between positive and negative emotions and affective commitment. As the results between negative and positive emotions were significant, this shows that partial mediation took place. Furthermore, the beta (β = -0.06, p < 0.05) was higher than the direct relationship between negative emotions and affective commitment (β = -0.07, p < 0.05). This can be interpreted as indicating that, when job satisfaction was included as a mediator, the relationship between negative emotions and affective commitment was stronger. The same interpretation is applicable when including job satisfaction as a mediator between positive emotions and affective commitment.
The data showed that the combination of positive and negative emotions plus job satisfaction explained 68% of the variance in affective commitment. This suggests that the  other variables contributed 32% of the variance in affective commitment. One South African research found that workplace spirituality (Kokt & Palmer, 2019), work-life balance and friendship (Coetzee, Ferreira, & Potgieter, 2019) and remuneration (Bussin & Brigman, 2019) were predictors of affective commitment. Another South African study established that perceived organisational support was a predictor of affective commitment (Satardien, Jano, & Mahembe, 2019). In Tshwane municipality, it was established that employee engagement was a variable that predicted employees' affective commitment (Maleka, Mpofu, Hlatywayo, Meyer, Carr, & Parker, 2019).

Practical implications
Approaches such as cultivating a conducive work environment, improving supervisors' interpersonal skills and implementing fair human resource practices that promote organisational justice have been highlighted as some contributors towards increased affective commitment (Liou, 2008). Managers should be aware that employees who are satisfied and happy have enhanced affective commitment. This can be achieved by ensuring that promotion practices and working conditions are improved. Municipal managers can create conducive working conditions by listening to employees, consulting employees before making decisions and giving employees carte blanche to make decisions and schedule their work. This would mean that municipal managers are not micro-managing employees. Furthermore, municipal managers should give employees reasons and justifications for their workload. In addition, municipal managers should be aware that employees who are positively and constructively given feedback are most likely to be committed (Maleka et al., 2016b).
When municipal managers engage employees, they are most likely to support strategic objectives and not abuse leave (Ditlev-Simonsen, 2015). Moreover, in this study, job satisfaction emerged as the greatest predictor of affective commitment. Therefore, senior managers should improve the job satisfaction level of employees in order to achieve a high level of affective commitment.
In addition, even when municipal managers experience that their subordinates are not performing optimally, they should impugn such subordinates in private rather than in public. It has been found that managers who criticise non-performing employees privately and give them an opportunity to develop, enhance these employees' affective commitment. Such managers also do not dent employees' self-confidence and sometimes refer employees to the employee assistance programme if employees have emotional challenges like depression and stress. Such employees are not able to complete their tasks timeously or do work that is not up to the expected standard. Furthermore, municipal managers who engage employees and share information about the municipality's financial status are able to deal with incorrect information that can create fears about the future of the municipality amongst employees (Maleka et al., 2016b).

Recommendations
Similar research should be conducted in the future. Other studies can be conducted in other municipalities in Mpumalanga and in South Africa, and if possible, can use probability sampling techniques and mixed methods designs so that qualitative data can also be solicited. The research can include remuneration as a variable because it has been found to be a point of dispute in the South African municipal context (Maleka, 2016). The following is recommended for managers to enhance affective commitment amongst employees: • Establish management forums to allow employees an opportunity to participate in all decision-making in the municipality. Alternately, managers can solicit employees' views during meetings and incorporate them when decisions are made that affect employees' duties and responsibilities. This will increase employees' level of attachment to the organisation whilst striving for positive outcomes and a successful organisation. • Implement a culture of two-way communication where managers listen to employees' concerns and provide them with guidance by supporting them equally and fairly in work-related and personal issues. Managers should encourage two-way communication and provide constructive feedback to enable employees to improve where necessary and to identify interventions to assist employees to exert more effort in their jobs. Most importantly, employees' concerns should be listened too. • Implement new strategies that will allow all employees to feel optimistic and keep their stress levels at an acceptable level for emotional well-being. This can include employees attending coaching sessions and being referred to the employee assistance programme. Where necessary, guidance and support should be provided to address professional and personal issues that may hinder employees' performance. • Encourage the use of a transparent management system and policies in terms of employee development and promotion within the organisation. This should be coupled with creating a conducive work environment and reasonable working conditions where employees will feel affectively committed. • Train employees on emotional intelligence and explain to employees how workload is determined. Furthermore, emotional intelligence interventions can be implemented. This will empower employees to not easily give up under difficult conditions in the organisation and not to display anger or act in an angry manner when they are under pressure or criticised.

Limitations
Firstly, this study was conducted at four of the six municipalities based in the Nkangala district. This makes it difficult for one to generalise the results of the study across all the municipalities within the district. Secondly, the study used a cross-sectional research design, which gives a once-off snapshot. In addition, because the researchers did not have a sampling frame, the results can be generalised to the population. Finally, the study could not measure the relationship (path) amongst the antecedents of positive affect (autonomy, participation, supervisory support, employee performance, etc.) and those of negative affect (well-being and work overload) to affective commitment.

Conclusion
The results from this study can assist municipal managers who intend to improve employees' affective commitment. Consistent with previous research, this study revealed that the predictors and affective commitment were unidimensional. Using AET as a theoretical framework, this study revealed that job satisfaction and negative and positive emotions explained 68% of the variance in affective commitment. This shows that these variables play a significant role in enhancing affective commitment and should be taken into consideration. The study showed that some mediations were not significant. For example, it was established that job satisfaction did not mediate the relationship between positive emotions and affective commitment; however, positive emotions were the second greatest predictor of affective commitment. The difference between the effect of job satisfaction and positive emotions as greatest predictors of affective commitment was small. Municipal managers should ensure that when they implement human resource management process such as promotions, they are transparent and fair. In addition, managers should ensure that when they give employees feedback on their performance, they should not evoke negative emotions such as anger. In conclusion, when employees are satisfied and happy, they will be affectively committed and offer excellent service to customers. This should lead to municipalities being perceived positively and generate revenue that will make them sustainable.