This study aimed to examine the link between transactional leadership, supply chain quality (SCQ) and business performance in the FMCG industry in the Gauteng province. As stated by Suhong et al. (2006), effective SCM has consistently become one of the most valuable tools for acquiring a competitive advantage, leading to enhanced business performance. However, supply chains are managed by people, and these include leaders and managers who have diverse orientations in their practice. Because supply chain networks are becoming more demanding, leadership within them has to become less arduous and more methodical (Cruz & Wang 2018). Inconsistencies in leadership processes and approaches within the supply chain create complications in the execution of its supply chain responsibilities (Cruse & Wang 2018b). One area where organisational leadership matters most is in stimulating high quality within business processes, systems and outputs (Foster 2008; Teoman & Ulengin 2018). As mentioned by Wu, Yang and Chiang (2011), business leaders in particular have an essential role to play in ensuring an acceptable quality of their supply chains. This view inspires a continuous consideration of the relationship between leadership and SCQ.

Although there are numerous leadership styles, such as transformational, transactional, distributed, servant or other supporting leadership functions, this study places exclusive focus on the transactional leadership style. This preference is supported by Laohavichien, Fredendall and Cantrell’s (2009) suggestion that among the available array of leadership styles, transactional leadership is the most appropriate and natural model of quality improvement through leadership. Also, the ultimate barometer of the effectiveness of all organisational efforts is the level of business performance (Yildiz & Karakas 2012; Zulkiffli & Perera 2011). Accordingly, it is important to conduct this research to gain insight into the impact of transactional leadership on SCQ and business performance in the FMCG industry.

There is much literature on leadership, although only a handful focuses on both leadership and SCM. An analysis of the general trends in research showed that most of the existing studies (Kuei 2001; Laohavichien et al. 2009; Tiwari & Sharma 2017) are aligned to either the subcomponents of SCM practices or total quality management, transformational leadership and organisational performance. However, as indicated by Foster (2008), more research should be conducted in the areas of SCQ with reference to leadership and business performance and other factors such as customer-centred approaches, quality methodologies, mutually beneficial supplier relations, human resource practices and occupational safety. Within South Africa, studies that were conducted in the specific area of leadership and SCQ in the FMCG industry are uncommon. The closest evidence found is research by Rautenbach and Rothmann (2017), who included some aspects of leadership in their study of antecedents to flourishing at work in an FMCG firm. But this fact notwithstanding, that study was outside the domain of SCM. Hence, the present study contributes to the body of knowledge by being one of the first of its kind to be conducted on the linkage between transactional leadership, SCQ and business performance in the FMCG environment in South Africa.

This article is structured as follows: the next section presents a brief review of literature and the hypotheses to be tested followed by a presentation of the conceptual model. Thereafter, the research methodology is presented, succeeded by the research results and discussion. The closing sections of the article present the conclusion and managerial implications as well as the limitations and future research directions.

The literature review briefly discusses the FMCG industry in South Africa and the research constructs such as transactional leadership, SCQ and business performance.

In line with the above review of literature, the conceptual model presented in Figure 1 was developed.

As shown in Figure 1, the predictor variable is transactional leadership, represented by three subcomponents, namely contingent reward, management by expectation (active) and passive-avoidant. The mediating variable is SCQ that is connected to business performance, which is the outcome variable. This study included three of the four subcomponents of transactional leadership. The fourth subcomponent (management by exception – passive) was not considered in the study because its items were discarded during the data analysis since they were cross-loading with others in the factor analysis procedure (refer to Table 3 and its preceding paragraph).

For this study, a quantitative approach was followed, in which numeric data were applied in objectively measuring the variables of interest and determining the possible relationships between them (Leedy & Ormrod 2015). The study also used a cross-sectional survey strategy, in which data were collected from the selected respondents in a single snapshot of time (Blumberg, Cooper & Schindler 2014).

The target population for this study consisted of managers and employees of two selected firms operating within the FMCG industry in the Gauteng province. From this population, respondents were selected using a non-probability purposive sampling technique. Purposive sampling is the deliberate choice of a specific respondent because of the qualities she or he possesses (Battaglia 2011; Etikan, Musa & Alkassim 2016). It was the sampling technique of choice in this study because respondents had to have some elementary knowledge of the subject under investigation, as demonstrated by their level of education, as well as sufficient knowledge of the FMCG industry which is indexed on their work experience.

A total of 150 questionnaires were distributed to managers and professional employees of the selected firms. A sum of 82 were returned, of which 73 were used in the final data analysis, which produced a response rate of 48.7%. A total of nine questionnaires were discarded because they had errors. The sample size of 73 respondents was considered to be acceptable because, as recommended by Forcino et al. (2015), sample sizes for multivariate analysis in quantitative studies should have at least 50 respondents. Other authors (DeVellis 2017; De Winter, Dodou & Wieringa 2009) suggest that the rule of thumb for factor analysis (the procedure also applied in this study) should be at least n = 50. Thus, the sample size of n = 73 was adequate to deliver sufficient data to minimise statistical errors, and increase validity as well as reliability and statistically significant results.

This study used a four-section self-administered survey questionnaire to collect data. Section A of the questionnaire provided information on the demographic profiles of the individual respondents. Section B consisted of 14 questions focusing on leadership styles as adapted from the Management Leadership Questionnaire (MLQ) (Bass & Avolio 2000). The questions in this section were presented in five-point Likert-type scales anchored by 1 – not at all, to 5 – frequently, if not always. Section C sought responses on SCQ using a six-item scale adapted from previous studies by Koh et al. (2007) and Li et al. (2009). In section C, questions were presented in five-point Likert-type scales anchored by 1 – strongly disagree, to 5 – strongly agree. Section D sought responses to business performance and used seven questions adapted from studies by Koh et al. (2007) and Li et al. (2009). The questions in section D were presented in five-point Likert-type scales anchored by 1 – much worse than industry average, to 5 – much better than the industry average. Measurement scale items are indicated in Tables 3–5.

A survey questionnaire was distributed using a combination of the drop and collect and email methods. The targeted respondents were given 1 week in August 2017 to complete the questionnaire. A confidentiality agreement and an informed consent form accompanied each questionnaire to facilitate the data collection. All respondents in the survey completed the same questionnaire and were not given incentives for participating in the study.

The collected data were analysed using the Statistical Packages for the Social Sciences (SPSS version 25.0). The statistical analysis techniques used in the study included descriptive statistics, exploratory factor analysis (EFA), correlations and regression analysis. Descriptive statistics such as frequencies and percentages are used to summarise the collected data and to show the patterns in that data (Nick 2007). In this study, descriptive statistics were used to analyse the demographic profile of respondents. Exploratory factor analysis is a procedure used to determine the underlying structure of a set of variables (Norris & Lecavalier 2009). In the present study, the EFA procedure was applied to determine the factor structure of the collected data regarding the constructs of the study, namely transformational leadership, SCQ and business performance. Correlation analysis is a statistical procedure used to test the strength and direction of linear associations between research constructs (Székely, Rizzo & Bakirov 2007). Regression analysis is a statistical procedure for testing for causality or prediction between a set of research constructs (Freedman 2005). Both correlation and regression analyses were used to test the relationships between the constructs under consideration in the current study.

The study was conducted after an ethical clearance had been granted by the Ethics Review Committee of the School of Business Leadership Research of the University of South Africa (Reference number: 2018_SBL_MBL_161_FA).

The profile of the respondents is presented in Table 1.

As shown in Table 1, males constituted 60.3% (n = 44) and females constituted 39.7% (n = 29) of the sample. The age groups most represented in the sample were 26–30 years at 21.9% (n = 16), 31–35 years at 23.3% (n = 17) and 41–50 years at 20.5% (n = 15). In the highest qualification category, diploma holders constituted 41% (n = 30), while matriculation and degree holders made up 21.9% (n = 16) apiece. In terms of the employment period, 31.5% (n = 23) of the respondents had 2–5 years’ work experience, followed by those with 11–15 years at 23.3% (n = 17). The race category composed of Africans with 71.2% (n = 52), Indians with 6.8% (n = 5) and white people with 21.9% (n = 16). With reference to the positions occupied, semi-skilled respondents were the highest in terms of representation in the sample at 35.6% (n = 26), followed by junior management with 31.5% (n = 23). Occupations represented in the sample constituted the following: senior officials and managers with 12.3% (n = 9); professionals with 13.7% (n = 10); technicians and associate professionals, clerks, and service and sales workers all with 16.4% (n = 12), respectively; and plant and machine operators and assemblers followed with 24.7% (n = 18).

Because the measurement scales were adapted from other studies, an EFA procedure was performed to determine the factor structure of the data collected in the study. A Kaiser–Meye–Olkin (KMO) test of sampling adequacy and the Bartlett’s test of sphericity were performed to assess whether the data captured were suitable for EFA. The results of the Bartlett’s and the KMO tests are presented in Table 2.

As indicated in Table 2, the KMO was 0.664 for the transactional leadership scale, 0.744 for the SCQ scale and 0.804 for the business performance scale. These values were above the 0.5 minimum threshold suggested by Kaiser (1974). The Bartlett’s test yielded significant chi-squares of 273.146 (df = 120) for the transactional leadership scale, 222.929 (df = 15) for the SCQ scale and 161.826 (df = 21) for the business performance scale. These test results were all significant at the 5% level. Because the results of the Bartlett’s and the KMO tests were within the recommended thresholds, it was determined that data were factorable, and hence EFA could be performed.

Correlation analysis was applied to measure the linear associations between transactional leadership subcomponents, SCQ and business performance. Pearson correlation analysis, which assesses the degree that quantitative variables are linearly related in a sample (Maxwell & Moores 2007), was used. The results of this analysis are presented in Table 6.

As shown in Table 6, positive inter-factor correlations ranging from r = 0.141 to r = 0.521 were observed between contingent reward, management by exception, SCQ and the two subcomponents of business performance. This result depicts that an increase in one factor would attract an increase in the other factors as well and vice versa. However, correlations between passive-avoidant and other constructs were either almost non-existent (e.g. with management by exception) or negative (e.g. with the rest of the constructs). These negative correlations depict an inverse association in which an increase in the passive-avoidant leadership would result in decreases in SCQ and both financial and non-financial performance.

To determine whether there were predictive relationships between the different variables, regression analysis was used. Regression analysis is a statistical technique that facilitates the summarising of the relationship between a variable of interest and one or more variables foreseen to influence that variable (Schroeder, Sjoquist & Stephan 2017). In this study, regression analysis was applied using the enter method, in which all variables are entered in a single step. Because there were three sets of relationships, three regression equations and models were computed. In the first regression model, contingent reward, passive-avoidant and management by expectation (active) were entered as the independent variables, while SCQ was entered as the dependent variable. The linear regression equation for model 1 was formulated as follows:

SCQ=β0+β1(contingent reward)+β2(passive-avoidant)  +β3(management by expectation; active)[Eqn 1]

where SCQ is supply chain quality, β₀ is the constant or intercept and β_1–3 are the coefficients of the independent variables.

The results of the first regression analysis are presented in Table 7.

As shown in Table 7, the regression analysis in model 1 revealed that the three transactional leadership predictors (adjusted R² = 0.252) explained approximately 25% of the variance in SCQ. This further denotes that the remaining 75% of the variance in SCQ is explained by other factors that were not considered in this study. The study considered the effect of multicollinearity, which suggests that several of the independent variables are closely linked in some way, which causes unusual results when attempting to study how well individual independent variables contribute to an understanding of the dependent variable (O’Brien 2007). Murray et al. (2012) suggest that the variance inflation factor (VIF) value should preferably be below 10. Concerning tolerance, larger tolerance values of more than 0.5 are more desirable as they are more indicative of lesser problems with multicollinearity (Denis 2011). Table 7 further shows that VIF values were lower than the recommended maximum of 10 and their tolerance values were higher than 0.5 minimum cut-off value. Thus, multicollinearity problems were almost negligible in the model.

In the second regression model, SCQ was entered as the independent variable, while non-financial business performance was entered as the dependent variable. The linear regression equation for model 2 was formulated as follows:

NFBP=β0+β4(SCQ)[Eqn 2]

where NFBP is non-financial business performance, β₀ is the constant or intercept and β₄ is the coefficient for SCQ.

The results of this regression analysis are presented in Table 8.

Table 8 reveals that in the regression analysis in model 2, the single predictor variable (SCQ: adjusted R² = 0.180) explained approximately 18% of the variance in non-financial business performance in this study. Concerning evaluating the assumptions of multicollinearity, the VIF for the predictor (1.000) was acceptable because it was less than 10 and the tolerance value (1.000) was higher than 0.5. Hence, multicollinearity problems were minimum in this regression model.

In the third regression model, SCQ was entered as the independent variable and financial business performance as the dependent variable. The linear regression equation for this model was formulated as follows:

FBP=β0+β4(SCQ)[Eqn 3]

where FBP is financial business performance, β₀ is the constant or intercept and β₄ is the coefficient for SCQ.

The results of this regression analysis are presented in Table 9.

Table 9 shows that in the regression analysis in model 3, the single predictor (SCQ: adjusted R² = 0.210) explained approximately 21% of the variance in financial business performance. In further assessing the assumptions of multicollinearity, the VIF for the predictor (1.000) was acceptable because it was less than 10, while the tolerance value (1.000) was also higher than 0.5. This also signals that multicollinearity problems were minimum in this model.

The five relationships that were hypothesised in the conceptual framework (Figure 1) of this study were tested using both correlation and regression analyses. The outcomes of the hypotheses tests that show the correlation coefficients, beta coefficients, the corresponding t and p-values and the outcome of each relationship are provided in Table 10.

The results of the hypotheses tests show that except for H₂, which initially suggested a positive relationship between passive-avoidant and SCQ, the other four hypotheses (H₁, H₃, H₄ and H₅) were supported. Further discussions on each hypothesised relationship are provided in the next section.

The first hypothesised relationship (H₁), which suggested that contingent reward has an influence on SCQ, was supported. Correlation analyses indicated a moderate positive association between contingent reward and SCQ (r = 0.496; p < 0.01). This result illustrates that when contingent reward increases, SCQ will also increase. Regression analysis indicates that contingent reward was statistically significant in predicting SCQ (β = 0.479; t = 4.487; p = 0.000), which demonstrates that the practice of contingent reward leads to SCQ. As such, there is a high likelihood of greater SCQ in FMCG environments where contingent reward is practised by leaders and managers. These results are parallel to those found in a study by Laohavichien et al. (2009), where the quality of the supply chain was higher in organisations where contingent reward was practised. Consistently, Fröber and Dreisbach (2014) found that having a strong contingent reward system in place can strengthen the competitive advantage through SCQ. A contingent reward system that is motivationally based is used to reward the achievement of goals and provides for positive reinforcement in meeting the set objectives. For instance, employees that meet their predetermined goals can be awarded some performance-based bonuses. In turn, such employees become motivated to exert additional effort in various areas of their work, leading to the improved quality of the supply chain. Thus, the awards encourage optimal performance in various areas of the organisation (Walumbwa, Wu & Orwa 2008), which explains the positive influence on SCQ in the FMCG industry as observed in this study.

The second hypothesised relationship (H₂), which suggested that management by expectation (active) has an influence on SCQ, was supported. There was a moderate positive association between management by expectation (active) and SCQ (r = 0.117; p < 0.01) in the correlation analysis. This result presents the picture that when management by expectation (active) increases, SCQ will also increase. The regression analysis indicates that management by expectation (active) was statistically significant in predicting SCQ (β = 0.500; t = 0.479; p = 0.000). This result implies that the practice of management by expectation (active) leads to SCQ. For instance, when managers are attentive in monitoring the activities of their subordinates, they may easily identify areas where standards are being compromised. Such managers may then immediately apply some remedial action to address any adverse situation, which catapults both organisational inputs and outputs back to the expected standards. In this manner, there is a high likelihood of greater SCQ in FMCG environments where managers and leaders practise management by expectation (active). Willis, Clark and O’Connor (2017) found similar results which showed that a positive relationship exists between management by expectation and contextual performance. In the setting of this study, this contextual performance is represented by SCQ.

The third hypothesised relationship (H₃), which suggested that passive-avoidant leadership has an influence on SCQ, was not supported. There was a weak negative association between the passive-avoidant and SCQ (r = -0.139; p < 0.01) in the correlation analysis, which implies that SCQ will decrease whenever the prevalence of passive-avoidant leadership increases. The regression analysis indicates that passive-avoidant leadership was statistically insignificant in predicting SCQ (β = -0.060; t = -0.569; p = 0.571). This result demonstrates that passive-avoidant leadership practices do not promote SCQ. Put in other words, there is a high likelihood of inadequate SCQ in FMCG environments where passive-avoidant leadership practices are common. Consistent with these results, Skogstad et al. (2014) found that passive-avoidant leadership practices may have negative consequences in critical situations where strong leadership is critical but is absent. This applies to scenarios where superior SCQ is a requirement for the overall performance of the supply chain. For instance, when supervision is lacking, which is what happens during passive-avoidant leadership, some employees generally tend to lower the expected standards of performance and apply minimum efforts in their work (Kabede & Demeke 2017). As a result, organisational systems and procedures suffer, leading to a decrease in quality standards. This perhaps explains why no significant relationship was found between passive-avoidant and SCQ in the present study.

The fourth hypothesised relationship (H₄), which suggested that SCQ has an influence on non-financial performance, was supported. In the correlation analysis, there was a moderate positive association between SCQ and non-financial business performance (r = 0.424; p < 0.01). This result shows that when there is an improvement in SCQ, non-financial business performance will likewise increase. The regression analysis indicates that SCQ was statistically significant in predicting non-financial business performance (β = 0.424; t = 3.947; p = 0.000). This result suggests that there is a high prospect of significant non-financial business performance in FMCG environments where there is greater SCQ. Consistent with this result, Zizlavsky (2014) suggests that there is a positive relationship between aspects of SCQ and non-financial business performance, which is indicated in areas such as customer and employee satisfaction, innovation, systems and procedures, and social and environmental performance. Accordingly, it is important to ensure that aspects of non-financial business form part of supply chain performance measurement tools in the FMCG industry.

The fifth hypothesised relationship (H₅), which suggested that SCQ has a positive influence on financial business performance, was supported. In the correlation analysis, there was a moderate positive association between SCQ and financial business performance (r = 0.470; p < 0.01). This result demonstrates that when there is an increase in SCQ, financial business performance will, in the same way, increase. The regression analysis indicates that SCQ was statistically significant in predicting financial business performance (β = 0.424; t = 3.947; p = 0.000). This result exhibits the finding that there is a high probability of higher financial business performance in FMCG environments where there is greater SCQ. In sync with this result, Lin et al. (2005) found a causal link between quality management practices that incorporate SCQ subcomponents and economic performance. Hence, the ability of a business in the FMCG industry to achieve better financial results is also linked to the quality of the supply chain in which the business operates.