Abusive customer behaviour and frontline employee turnover intentions in the banking industry: The mediating role of employee satisfaction

Abstract

Purpose—The study aims to examine the mediating role of frontline employee job satisfaction in the relationship between abusive customer behaviour and employee turnover intentions in the banking industry.

Design/Methodology/Approach—The study sampled 186 frontline employees from 10 banks in the Ashanti Region of Ghana. The data analysis software used was STATA. The main method of data analysis was structural equation modelling; however, a confirmatory factor analysis was first conducted to check for the validity and reliability of the research instrument. A mediation analysis was finally conducted and confirmed using Sobel test.

Findings—The study revealed that abusive customer behaviour has a negative and significant effect on frontline employee satisfaction in the banking sector. Frontline employee satisfaction also has a negative and significant effect on employee turnover intention in the banking industry. Abusive customer behaviour has a positive and significant effect on frontline employee turnover intentions in the banking sector. Frontline employee satisfaction also has a partial mediating effect on the relationship between abusive customer behaviour and employee turnover intentions in the banking industry.

Originality/Value—In the banking industry, frontline employees such as tellers play a critical role as they serve as intermediaries between the banks and the public. However, empirical studies on the influence of customer behaviour on employees have largely concentrated on other sectors. Although literature supports the individual relationships among the variables studied, there existed a gap on the mediating role of employee satisfaction, especially in the banking sector.

Keywords:

• Abusive
• customer
• misbehaviour
• employee
• satisfaction
• turnover
• intentions

View correction statement:

Correction

Public Interest Statement

It is a widely held notion in marketing that customers are reason for the existence of a firm. Implying a firm with no customer will run out of business. This notion is further strengthened by the competitive nature of the business environment in recent times, making customers very valuable to business entities. It is of no surprise that customers are regarded as Kings and Queens. This has led to the development of a marketing mantra dubbed “The Customer is Always Right”. These practices and ideologies place much burden on the service employees, especially frontline staff, who have direct contact with customers. In the banking sector, these personnel are mostly the tellers and the customer care representatives. But how could these frontline staff work with satisfaction in the face of abusive customers? And how could abusive customer behaviours influence employee’s turnover intention, despite the competition for jobs in the Ghanaian banking sector?

1. Introduction

In the field of marketing, “service with a smile” has remained a very important mantra. This has led to much marketing research to focus on the relationship between customer service encounter and customer satisfaction and related concepts (Bamfo, Dogbe, & Osei-Wusu, 2017; Hennig-Thurau, Groth, Paul, & Gremler, 2006; Ma & Dubé, 2011; Oliver, 1997). These studies, more often than not, looked at service employees’ behaviour as the determinant of customer behaviour (Hennig-Thurau et al., 2006; Tsai, 2001). That is, customer satisfaction and customer loyalty have been widely attributed to service employees’ behaviour (Bamfo & Dogbe, 2017; Oliver, 1999; Oliver & DeSarbo, 1988). Some studies have also pointed out that service encounter (or interaction) is a bidirectional process, and the success of the encounter depends on both the customer and service employee (Madupalli & Poddar, 2014; Rafaeli & Sutton, 1989). It therefore stands to reason that the behaviour of one group (customers or service employees) could influence the other.

In a study conducted in America by Patterson, McColl-Kennedy, Smith and Lu (2009), about 25% of workers encountered some form of physical or psychological aggression from customers. It has also been identified that, in the call centre of telecommunication companies, about 20% of all customers who called were hostile (Grandey, Dickter, & Sin, 2004). Customers usually take advantage of the mantra “The Customer is Always Right”, by making unreasonable demands from service employees, or treating them unjustly (Berry & Seiders, 2008). In a service firm, as customer expectations are not met (either through company policies or service employee inabilities), they sometimes react with anger towards the service employees (Grove, Fisk, & John, 2004; McColl-Kennedy, Patterson, Smith, & Brady, 2009). These reactions or behaviours from customers, either duly or unduly, have some effect on employee satisfaction and turnover intentions (Li & Zhou, 2013; Madupalli & Poddar, 2014; Poddar & Madupalli, 2012; Sawyerr, Srinivas, & Wang, 2009; Yoo, 2017). For example, Sawyerr et al. (2009) found that high turnover rates among frontline service employees are often due to problematic customers making jobs stressful for employees.

In the banking industry, frontline service employees such as tellers play a critical role as they serve as intermediaries between banks and the public (LeBlanc & Kelloway, 2002). However, previous studies on the influence of customer behaviour on employees have largely concentrated on industries other than the banking sector. For example, Han, Bonn and Cho (2016) looked at restaurant frontline service employees, and Madupalli and Poddar (2014) also looked at employees in telecommunication call centres. Also, studies have showed a direct effect of customer abusive behaviour on employee satisfaction (Akkawanitcha, Patterson, Buranapin, & Kantabutra, 2015; Bitner & Wang, 2014; Grandey, Rupp, & Brice, 2015; Kim, Kim, Lacey, & Suh, 2018; Stock & Bednarek, 2014; Wilder, Collier, & Barnes, 2014). Other studies have also identified a direct effect of employee satisfaction on turnover intentions (Ahmad, 2018; Chan, Mai, Kuok, & Kong, 2016; Chung, Jung, & Sohn, 2017; Luz, de Paula, & de Oliveira, 2018; Suifan, Diab, & Abdallah, 2017; Tongchaiprasit & Ariyabuddhiphongs, 2016; Zeffane & Melhem, 2017). This presupposes that employee motivation could have some mediating role between abusive customer behaviour and turnover intentions. However, studies on this are very limited.

From 2017 to 2018, Ghana has seen the collapse of seven banks, namely, Capital, UT Bank, UniBank, Beige, Sovereign, The Construction Bank and The Royal Bank. Naturally, it would lead to employees losing their jobs, and increasing pressure on the retained staff to be highly productive or risk losing their jobs. Despite this employment challenge in the industry, are employees still willing to voluntarily leave their banks because of customer abuse? And to what extent would their job satisfaction mediate customer abuse and the intention to quit their jobs? In meeting this research gap, the current study looked at how frontline employee job satisfaction could mediate the relationship between abusive customer behaviour and employee turnover intentions in the banking sector of Ghana.

2. Literature and theoretical review

2.1. Abusive customer behaviour

Abusive customer behaviour is a series of deliberate acts that violate widely held norms (Reynolds & Harris, 2006). Abusive customer is generally different from customer complaints and is synonymous with terms like disruptive customer behaviours (Cai, Lu, & Gursoy, 2018; Gursoy, Cai, & Anaya, 2017), dysfunctional customer behaviour (Habel, Alavi, & Pick, 2017; Kim et al., 2018), customer misbehaviour (Rummelhagen & Benkenstein, 2017), customer incivility (Han et al., 2016; Walker, van Jaarsveld, & Skarlicki, 2014), deviant customer behaviour (Reynolds & Harris, 2006) and jay-customer behaviour (Fong, So, & Law, 2017; Harris & Reynolds, 2004). Greer (2015) presented six dysfunctional customer behaviours in the financial and healthcare services. These were property abuse, fraudulence, verbal abuse, physical aggression, under-participation and over-participation. Among these dysfunctional behaviours, verbal abuse was found to be the most dominant of all (Fong et al., 2017). Ang and Koslow (2012) also broadly classified customer misbehaviour into disruption and violation of norms perspectives and indicated that disruption perspective is more managerially useful than violation of norms perspective. They explained that disruption of business is not just unlawful but could also lead to loss of well-being and company reputation. As indicated by Daunt and Greer (2015), customer misbehaviour thrives when there is a motivated offender, suitable target and the absence of capable formal guardian transpire in time and space. This means that abusive customers seek opportunities to do so when they judge a loophole in the service provision. Fong et al. (2017) further identified service failure as a major cause of abusive customer behaviour, as earlier stated by Greer, Russell-Bennett, Tombs and Drennan (2014). Daunt and Harris (2012) also presented three motives of dysfunctional customer behaviour, namely, financial gains motive, ego gain motive and revenge motive. Further, Fisk et al. (2010) identified five motivators for dysfunctional behaviours which include the potential for material gains (reward and benefits), opportunity to cheat, perceived injustice, dissatisfaction and external pressure. Past customer misbehaviour has also been found to predict future misbehaviour intent (Daunt & Harris, 2011). In a service sector where customers are involved in the process of production, abusive customer behaviour influences the overall service perception of other by-standing customers (Cai et al., 2018; Gursoy et al., 2017).

2.2. Abusive customer behaviour and frontline employee job satisfaction

Some studies have identified that customer misbehaviour in the form of unwarranted complaints, verbal aggression, physical aggression and/or unreasonable demand affects service quality negatively, as well as employee workplace well-being (Grandey et al., 2015; Lytle & Timmerman, 2006; Reynolds & Harris, 2006). Kim et al.’s (2018) verbal dysfunctional customer behaviour increases job stress of frontline service employees, leading to emotional fatigue. Akkawanitcha et al. (2015) conducted a study on frontline service employees and found out that customer aggression had a negative effect on employees’ physical well-being and self-esteem. In the daily operations of frontline service employees, they encounter arrogant, rude or even aggressive customers. As indicated by Ivarsson and Larsson (2010), the mantra “The Customer is Always Right” gives some sovereignty to customers, making frontline staff obligated to customer satisfaction, as they are the reason for the existence of the firm. Firms emphasise external customer satisfaction to the detriment of internal customer satisfaction (employee satisfaction), because they are of the belief that external customers bring in cash to the firm (Bamfo, 2009). Employees are thus expected to be customer oriented and manage deviant or abusive customer behaviour in such a way that the firm would not lose its profitable customers (Bitner & Wang, 2014; Wilder et al., 2014). Frontline service employees are expected to work under pressure, keep calm, remain courteous and efficient, which becomes very difficult to achieve when interacting with difficult customers (Akkawanitcha et al., 2015). Unreasonable customer demands increase frontline employees’ emotional exhaustion, which in turn negatively affects employees’ attitude and commitment towards work (Harris & Daunt, 2013; Stock & Bednarek, 2014). Based on this, the first hypothesis was thus established as follow:

H₁: Abusive customer behaviour has a negative relationship with frontline employee job satisfaction in the banking industry.

2.3. Frontline employee job satisfaction and turnover intentions

Employee turnover intention is employee’s future plans of leaving an organisation and/or industry. A number of individual, job-related, organisational (institutional) or industry factors may contribute to employee turnover intentions (Babakus, Cravens, Johnston, & Moncrief, 1999; Brashear, White, & Chelariu, 2003). Both empirical and theoretical reviews have established employee job satisfaction as a determinant of employee turnover intentions (Ahmad, 2018; Chan et al., 2016; Chung et al., 2017; Kim, Im, & Hwang, 2015; Luz et al., 2018; Mathieu & Babiak, 2016; Poddar & Madupalli, 2012; Suifan et al., 2017; Tongchaiprasit & Ariyabuddhiphongs, 2016; Zeffane & Melhem, 2017). One of the early models developed on job satisfaction as a determinant of employee turnover intentions was by Rhodes and Doering (1983). Since then, other studies have established this relationship in different fields. Ahmad (2018) found an indirect effect of employee satisfaction on turnover intentions, with employee commitment mediating them. A study by Luz et al. (2018) identified an inverse relationship between employee’s satisfaction with job and intention to turnover. Other studies also identified a negative effect of employee satisfaction on turnover intentions among employees (Chan et al., 2016; Kim et al., 2015; Mathieu & Babiak, 2016). Chung et al. (2017) found a direct effect of employee satisfaction on turnover intentions, although self-determined work motivation moderates this relationship. Suifan et al. (2017) also concluded that employee satisfaction had a direct effect on turnover intention. Zeffane and Melhem (2017) compared the satisfaction level of employees in public and private firms and concluded that public sector employees were more satisfied and were lower turnover intention compared to the private firms. Employee satisfaction was also found to have had a direct and indirect effect on turnover intention, as job stress could mediate the relationship (Tongchaiprasit & Ariyabuddhiphongs, 2016). Walsh (2011) also demonstrates that job satisfaction had a direct effect on turnover intentions. Dissatisfaction and emotional exhaustion experienced by frontline service employees may not only lead to organisational turnover but industry turnover as well (Blau, 2007; Blau, Tatum, & Ward-Cook, 2003; Carmeli, 2005; Li & Zhou, 2013). Depending on the skills of the employee and the opportunities available to him or her, he or she could leave the industry altogether. Hence, the second hypothesis was established as

H₂: Frontline employee job satisfaction has a negative relationship with employee turnover intentions in the banking industry.

2.4. Abusive customer behaviour and frontline employee turnover intentions

Most studies in the past on workplace aggression focused largely on inter-organisational factors (Cortina, 2008; Cortina & Magley, 2009), workplace bullying (Rayner, Hoel, & Cooper, 2002), abusive supervision (Tepper, 2000), intra-organisation members and job tasks (Frone, 2000), incivility (Andersson & Pearson, 1999), interpersonal conflict (Mulki & Wilkinson, 2017; Spector & Jex, 1998) and interactional justice (Bies & Moag, 1986), while paying less attention to extra-organisational factors like customer misbehaviour. A study by Grandey et al. (2004) found customer misbehaviour as a determinant of employee attitudes and behaviour. Customer misbehaviour is exhibited in many forms such as customer injustice (Rupp, McCance, Spencer, & Sonntag, 2008; Skarlicki, Van Jaarsveld, & Walker, 2008), sexual harassment (Yagil, 2008) and verbal abuse (Grandey, Kern, & Frone, 2007). Some customers also engage in unwarranted complaints, argue with the service personnel and make excessive demands (Cole & Bedeian, 2007). As indicated by Taylor (1991), frequent customer verbal abuse stimulates equally frequent and strong emotional reactions from the service employees, as these negative events typically induce strong and quick cognitive, physiological and behavioural reactions. These have been supported by studies like Karatepe, Yorganci and Haktanir (2009) and Kim, Murrmann and Lee (2009). Interpersonal conflict with customers significantly affects employee felt stress, leading to higher turnover intentions (Mulki & Wilkinson, 2017). Han et al. (2016) also found customer incivility to have a significant effect on employee turnover intention, although it is fully mediated by job burnout. Frontline service employee job is mentally and physically exhausting, as excellent service is usually equated to employees’ attentiveness to customers’ needs (Poddar & Madupalli, 2012). Employees are expected to demonstrate positive moods and enthusiasm, even when customers are obviously rude. This pretence or acting puts much increase on emotional exhaustion, trickling employee turnover. Yagil (2008) specifically identified customer misbehaviour as a determinant of employee turnover intentions, jobs satisfaction and absenteeism. Li and Zhou (2013) also found customer verbal aggression to significantly predict employee turnover intentions. Walsh (2011) also demonstrates that job satisfaction mediated the relationship between perceived customer misbehaviour and turnover intentions. Thus, the last two hypotheses are stated as follows:

H₃: Abusive customer behaviour has a positive relationship with employee turnover intentions in the banking industry.

H₄: Frontline employee job satisfaction plays a mediating role between abusive customer behaviour and employee turnover intentions in the banking industry.

2.5. Job characteristics theory

Hackman and Oldham in 1976 developed the job characteristic theory, which was later supported by other researchers like Hellriengel, Jackson and Slocum (1999), Dugguh (2008) and Steel (2012). The theory presents five job characteristics which impact job outcome. The first is “skill variety (SV)”, which represents the extent to which a job demands different variety of activities, skills and talents in executing the job (Dugguh, 2008). Frontline staff serves as the bridge between the organisation and the customer. Frontline staff are therefore expected to possess multiple skills to be very effective at their post. They are expected to be people centred and have good communication skills, good interpersonal relations, well-mannered, control over emotions and moods etc. The second job characteristic is “task identity (TI)”, which considers the extent to which job demands completion in whole or in an identifiable piece (Steel, 2012), that is, the extent of executing job from start to finish, with an identifiable outcome. The third is “task significance (TS)”, that is, the extent to which job has an impact on the lives of people (both in and out of the organisation) and the environment (Hellriengel et al., 1999; Steel, 2012). The job of frontlines staff could either elongate a customer’s relationship with the organisation or abrogate it, especially in organisations where the primary mode of service consumption is through direct contact with frontline staff. The fourth job characteristics presented by the theory was “job autonomy (AU)”. This is the extent to which the job gives employee independence, freedom and discretion in the execution of task (Dugguh, 2008; Steel, 2012). Studies (such as Harris & Daunt, 2013; Kim et al., 2018; Stock & Bednarek, 2014) have indicated that frontline job roles are emotionally more stressful compared to employees with no frequent contact with customers. Other studies have also identified job autonomy to have a positive effect on employee motivation and job satisfaction (Gözükara & Çolakoğlu, 2016; Li, Li, & Chen, 2018). Granting these employees, some level of job autonomy will help reduce this emotional stress by identifying new ways of executing task and handling challenges like abusive customers. The last job characteristic was “job feedback (FB)”. Job feedback represents the extent to which clear performance results are provided to employees. Serving employees with objective feedback helps them in improving their weakness and leveraging on their strengths.

The theory presents an index on the likelihood of job affecting employee’s behaviour, called the Motivating Potential Score (MPS). The formula is as follows:

$\mathrm{M}\mathrm{P}\mathrm{S}=\frac{\mathrm{S}\mathrm{V}+\mathrm{T}\mathrm{I}+\mathrm{T}\mathrm{S}}{3}\times \mathrm{A}\mathrm{U}\times \mathrm{F}\mathrm{B}$

It has been established that jobs which highly motivate employee potentials should also be high in at least skill variability (SV), task identity (TI) or task significance (TS). It must however be high in both autonomy (AU) and feedback (FB). Job characteristics theory assumes that a high Motivating Potential Score (MPS) implies that job satisfaction and motivation will have negative outcomes such as absenteeism and turnover. That is to say, there is an inverse relationship between MPS and variables such as absenteeism and turnover (Steel, 2012).

3. Methodology

The study adopted an explanatory research design as it established relationships among variables (Saunders, Lewis, & Thornhill, 2009). The variables were abusive customer behaviour, employee job satisfaction and employee turnover intentions. Survey research approach was also adopted for the study, where banks in the Ashanti Region of Ghana were sampled for the study. The data collection was primarily the administration of structured questionnaire based on some past studies on the subject (Li & Zhou, 2013; Madupalli & Poddar, 2014; Poddar & Madupalli, 2012; Yoo, 2017). The items measuring the three main constructs were to be responded to on a Likert scale of 1—strongly disagree, 2—disagree, 3—neutral, 4—agree and 5—strongly agree.

The study was limited to 2 main frontline staff in the banking sector, which were tellers and customer service personnel of 10 selected banks in the Ashanti Region of Ghana. These personnel were selected as they have frequent touch with customers. The number of employees sampled from each bank was determined by the size of their frontline staff (influenced by the number of branches in the region). For the banks with more than five branches in the region, 25 staff were selected, and 15 staff selected from banks with less than 5 branches in the region. In all, 5 banks selected had more than 5 branches, giving 125 staff sampled from there (5 × 25). Five (5) banks also had less than 5 branches, giving 75 staff sampled (5 × 15).

After the determination of the expected sample, the convenient sampling technique was used in sampling respondents for the study. By this approach, respondents who were available and willing to respond to the questionnaire were finally selected for the study (Saunders et al., 2009). Two hundred questionnaires were distributed in all; however, after the exercise, 186 were appropriately filled and returned. The data analysis was therefore conducted using 186 questionnaires. By reason of the target group (tellers and customer service personnel), females dominated the study by constituting 62% of the respondents. Most of the respondents were aged 25–35 years and held first-degree certificates. The majority of the respondents had also worked for their respective banks for more than 5 years, giving them much experience to offer reliable information towards the study.

The data analysis software used was STATA. The main method of data analysis was structural equation modelling (SEM); however, a confirmatory factor analysis (CFA) was first conducted to check for the validity and reliability of the research instrument. After the factor extraction procedure, each of the constructs had four factors left and was subsequently used for the SEM. A mediation analysis was finally conducted and confirmed using Sobel test.

4. Reliability and validity analysis

To test the reliability and validity of the measures, the study used a CFA. A number of items were deleted from the model because they either loaded poorly or cross loaded on other items, as was guided by Bagozzi and Yi (2012). The CFA as presented in Table 1 contained the retain items used in the model estimation. From the analysis, all the items under abusive customer behaviour, employee job satisfaction and employee turnover intentions had a factor loading of greater than 0.5. The minimum expected Cronbach’s alpha is 0.7; however, the alpha score for all the three dimensions was greater than 0.8, indicating a high internal consistency. Composite reliability was also expected to be minimum 0.7; however, the scores presented indicate that all the three dimensions had a score of greater than 0.8, also supporting the fact that there is high internal consistency among items measuring a specific construct. As proposed by Fornell and Larcker (1981), Average Variance Extracted (AVE) score greater than 0.5 is an indication of convergent validity. As could be seen from Table 1, all the AVE scores were greater than 0.5, indicating a convergent validity among the items measuring specific constructs.

Table 1. Confirmatory factor analysis

Variables					Factor loadings	Cronbach’s alpha		Composite reliability	Average variance extracted
AC						0.8692		0.877	0.544
Some of the customers blame me rather than the company policies when I am unable to resolve their problems (ac1)					0.6955571
Some customers accuse me of stealing their money (ac2)					0.7202168
Some customers are extremely repetitive without giving me a chance to speak (ac3)					0.8090290
Some customers are very rude (ac4)					0.7958084
ES						0.8596		0.868	0.527
My job is satisfying (es1)					0.5487928
My job gives me a sense of accomplishment (es2)					0.8061396
This current job fits my ideal job (es3)					0.8255997
I feel good about working at this company (es4)					0.7478079
TI						0.8226		0.865	0.520
I intend to leave this company within a short period of time (ti1)					0.7620420
I do not think I will spend all my career with this company (ti2)					0.6904007
I intend leaving this banking (sector) industry all together (ti3)					0.8496411
I am looking at some other jobs now, in the same field (ti4)					0.6077142
Variables	x^2	Df	p-Value	x^2/df	RMSEA	CFI	TLI	SRMR	CD
Abusive customer	2.65	2	0.27	1.33	0.033	0.998	0.993	0.014	0.774
Employee satisfaction	3.95	2	0.14	1.98	0.057	0.996	0.987	0.014	0.851
Turnover intention	0.51	2	0.77	0.26	0.000	1.000	1.024	0.007	0.706

AC: Abusive customer; ES: employee satisfaction; TI: turnover intention.

Regarding the goodness-of-fit indices presented, Chi-square (x²) divided by the degree of freedom (df) is expected to be less than 2 (Hu & Bentler, 1999). For all the three constructs, x²/df was less than 2. In similar manner, the p-values for the x² are expected to be statistically insignificant (Hu & Bentler, 1999; Newsom, 2012). And this was so for all the constructs. The root mean square error of approximation and the standardized root mean square residual are expected to be less than 0.08 (Brown, 2006; Kline, 2005). The comparative fit index and the Tucker–Lewis index are also expected to be above 0.95 (Brown, 2006; Kline, 2005). Meeting these assumptions means that the model is fit for estimation. And as could be seen from Table 1, the results meet all the threshold, implying that any estimation will be valid and reliable. The coefficient of determination for all the constructs were greater than 0.7, indicating observed items explained more than 70% of the latent variable.

In assessing the discriminant validity, the square root of the AVE values was expected to be compared with the correlations. As a rule of thumb, the square root of AVE is expected to be greater than the correlation values. From Table 2, square root of the AVEs of all the three constructs was greater than the intercorrelation scores, implying that the observed items distinctively measured the constructs they were expected to measure.

Table 2. Discriminant validity

Variables	1	2	3
Abusive customers (1)	0.738
Employee satisfaction (2)	−0.559***	0.726
Turnover intentions (3)	0.545***	−0.634***	0.721

***p < 0.01 (two-tailed). Italic value represents the difference between the Square root (AVE) values and the other presented in the table. 

Notes: Square root (AVE) on the diagonal and construct correlations below the diagonal.

5. Hypotheses testing

The main focus of the study was to ascertain the mediating effect of employee job satisfaction on the relationship between abusive customer behaviour and employee turnover intentions. But a number of relationships need to be established before a conclusion could be made on the main objective of the study. This led to the establishment of three hypotheses (H₁, H₂ and H₃) prior to the main one (H₄).

The first hypothesis ascertained the effect of abusive customer behaviour on frontline employee job satisfaction in the banking industry. From Table 3, the relationship between them was negative and statistically significant. By implication, more abusive customer behaviour reduces employee job satisfaction. For example, the more customers blame frontline staff when unable to resolve or handle their complaints; instead of blaming the systems which are not allowing staff to perform, their job satisfaction reduces. In a similar manner, frontline staff who have much encounter with rude and extremely repetitive customers were less satisfied with the banking job. Prior studies (such as Akkawanitcha et al., 2015; Bitner & Wang, 2014; Grandey et al., 2015; Wilder et al., 2014) also came to similar conclusion. The first hypothesis was therefore accepted.

Table 3. Structural equation model estimation

Hypotheses	Estimates	Std. err.	Z-score	Results
H1: Abusive customer → employee satisfaction	−0.4937143	0.068077	−7.25***	Supported
H2: Employee satisfaction → turnover intentions	−0.7468008	0.14995	−4.98***	Supported
H3: Abusive customer → turnover intentions	0.3694597	0.10771	3.43***	Supported

***p < 0.01.

The second hypothesis assessed the effect of frontline employee job satisfaction on turnover intentions in the banking industry. Results indicate a negative and statistically significant relationship between the variables. This means that more satisfied frontline staff has less propensity to leave their firm. Frontline staff who find their jobs satisfying, who feels their jobs give them a sense of accomplishment, who feel their current job fits their ideal job and those who feel good about their company are less likely to leave, and vice versa. This finding was in conformity with past studies (such as Ahmad, 2018; Chan et al., 2016; Chung et al., 2017; Kim et al., 2015; Luz et al., 2018; Mathieu & Babiak, 2016; Poddar & Madupalli, 2012; Suifan et al., 2017; Tongchaiprasit & Ariyabuddhiphongs, 2016; Zeffane & Melhem, 2017). This meets priori assumption and therefore, the second hypothesis was also accepted. Ahmad (2018) for example identified an indirect effect of employee satisfaction on turnover intentions, and Luz et al. (2018) also found an inverse relationship between employee’s satisfaction with job and intention to quit.

Finally, the last hypothesis looked at the effect of abusive customer behaviour on frontline employee turnover intention in the banking industry. Results presented in Table 3 indicate that customer abusive behaviour has a positive and significant effect on frontline employee turnover intention in the banking industry. This implies that more abusive customer behaviour leads to a rise in frontline employee turnover intention in the banking sector, and vice versa. Frontline employees like cashiers who are wrongfully accused of stealing money, employees who are spoken to rudely by customers and employees who are blamed unduly for inability to meet customer demand are more likely to either leave the organisation or leave the banking sector altogether. Poddar and Madupalli (2012) and Yagil (2008) also concluded alike during their empirical study. Han et al. (2016) also found customer incivility to have a significant effect on employee turnover intention. The third hypothesis was thus accepted.

The main thrust of the work, and the final hypotheses for that matter, considered the mediating role of frontline employee job satisfaction on the relationship between abusive customer behaviour and employee turnover intentions in the banking sector. The results as presented in Table 4 indicate that there was a significant and direct effect of abusive customer behaviour on frontline employee turnover intentions in the banking industry. The indirect effect (that is, passing through employee job satisfaction) was also statistically significant. Similarly, the total effect was also significant. Going by the Baron and Kenny (1986) approach, it was concluded that employee job satisfaction had a partial mediating effect. This was because both the direct and indirect effects were statistically significant. The Sobel’s test which considers both the coefficients and the standard errors of the indirect relationships was also conducted. The result was also significant, implying frontline employee job satisfaction, mediated the relationship between abusive customer behaviour and employee turnover intentions in the banking industry. The final hypothesis was thus accepted. Figure 1 also presented the diagrammatic form of the SEM among the variables.

Table 4. Mediation analysis

Path	Direct effect	Indirect effect	Total effect	R^2	Sobel test	Results
H4: Abusive customer → employee satisfaction → turnover intentions	0.369***	0.369***	0.738***	0.828	4.106***	Partial mediation

***p < 0.01.

Figure 1. Structural equation model.

6. Conclusion and managerial implications

It was concluded that abusive customer behaviour has a negative and significant effect on frontline employee satisfaction in the banking sector of Ghana. Frontline staff who experienced much abusive behaviour from customers were much dissatisfied with work. Frontline employee satisfaction also has a negative and significant effect on employee turnover intention in the banking industry. The more dissatisfied frontline employees were, the higher the tendency to leave the organisation or even the industry as a whole. Abusive customer behaviour also has a positive and significant effect on frontline employee turnover intentions in the banking sector. The more frontline employees experienced abusive behaviour from customers, the more likely they are to quit their jobs. The significant relationships among these variables imply that frontline service employee satisfaction has a partial mediating effect on the relationship between abusive customer behaviour and employee turnover intentions in the banking industry.

Abusive customer behaviour must therefore be a major concern to businesses, particularly service firms, where there is high employee–customer interactions. Employee turnover has always been costly to organisations. The cost could come in the form of time resources to recruit and select new employees, monetary cost in recruiting and selecting new employees, work overload on remaining employees, loss of employee morale, one turnover sparking chains of turnovers in the organisation etc. Management of banking institutions and service firms for that matter should pay much attention to abusive customer behaviour, as it could have a detrimental effect on their employee turnover.

Service firms must embark on relentless assessment of customer perception after service consumption. Knowing customers’ level of satisfaction or dissatisfaction with service will help in identifying whether or not they have or are likely to engage in customer misbehaviour in the future interaction with service personnel. If possible, knowing which service provider (teller) served him or her could also give a clue as to which employees are likely to be dissatisfied as a result of customer interactions. This will help firms in identifying which frontline staff would need to be given more counselling or training, so as to reduce the turnover intention that would result from job dissatisfaction.

Considering how emotionally exhaustive it is for frontline staff in handling difficult customers, job rotation, if possible, should be conducted in service firms. Staff must not only be made to spend all their working life in a particular job role. In the banking industry for example, tellers could be made to work in operations once a while, just to have some break from meeting difficult customers.

Firms must also have systems in place to collect feedback from staff on customer interaction, not just focusing on how external customers were satisfied with the service. Employees are the first customers of a firm and must also be given similar attention as granted external customers. Abusive customer behaviour has a great toll on employees, and firms should have a supportive climate to help give employees the resilience needed to handle difficult customers. Just as is done for the internal customers (employees), external customer must also be given some code of conduct. This must be posted at the various branches of the banks or service firms. In Ghana for example, it is legibly written at the waiting room of Kotoka International Airport that customer abusive behaviour on employees shall not be tolerated in any manner. This could be emulated by other service firms, just to put some limitations on deviant customers.

7. Limitations of the study

The major limitation of this study was access to data. Banks have very strict rules on confidentiality, which greatly affects the nature of information they are willing to part with. Although permission was sought from management of the banks prior to the study, the data gathering process was still not a smooth one. Two hundred questionnaires were distributed, but 186 were retrieved after the process. Employees were also sceptical about giving out information regarding their turnover intention, as some feared management may use it against them. This was particularly a challenge with the banks which decided to distribute the questionnaires themselves to staff.

8. Compliance with ethical standards

First and foremost, management of the various selected banks was contacted for approval prior to the study. The management of some banks decided to distribute the questionnaires themselves to staff and made them available to the researcher when they were filled.

Second, the respondents were made aware of the study and its intended purpose. Staff were informed that the study was purely for academic purpose, and anonymity was highly maintained. Details that would make it possible to link a questionnaire to a particular respondent after the data collection were avoided. This was also to enhance the reliability of data collected from respondents.

Correction

This article was originally published with errors, which have now been corrected in the online version. Please see Correction (http://dx.doi.org/10.1080/23311975.2018.1574379).

Additional information

Funding

The authors received no direct funding for this research.

Notes on contributors

Bylon ABeeku Bamfo

Bylon Abeeku Bamfo is a senior lecturer in the School of Business of the Kwame Nkrumah University of Science and Technology (KNUST), Ghana. He is a PhD holder from the School of Management of Southampton University in the United Kingdom. He is also a member of CIM, UK. His research areas include advertising, consumer behaviour, entrepreneurship and small business management, among others.

Courage Simon Kofi Dogbe

Courage Simon Kofi Dogbe is a PhD candidate at Jiangsu University, China. Prior to his enrolment, he was the executive director of Docks Global Consult, Ghana. He holds an MBA in International Business, and BSc. Business Administration (Marketing), all from KNUST, Ghana. He is a business research consultant.

Harry Mingle

Harry Mingle is the branch manager at Standard Chartered Bank, KNUST Branch, Ghana. He holds an MBA in Strategic Management and Consulting, and BSc. Planning, all from KNUST. He is currently pursuing a professional course with Chartered Institute of Bankers (CIB).