Can workload decrease turnover intention of
auditors?
The mediating effect of job satisfaction and work-life conflict moderated by
professional identification
Student: Niek Harkema S2364344
Supervisor: Dr. D.B. Veltrop
University of Groningen Faculty of Economics and Business
9 January 2020
Master Thesis MSc Accountancy (EBM869B20)
Abstract
Audit firms increasingly have to deal with high turnover rates of their employees. One of the main reasons for auditors to leave their current job is due to high workloads. However, some studies also suggest that some auditors may appraise workload as a challenge and may therefore be less likely to leave in response to high workload. Based on previous research and theory, I expect that workload may indirectly impact auditor turnover intention through two different mechanism (i.e. mediate), namely auditor job satisfaction and auditor work-life conflict. Furthermore, building on research demonstrating the importance of auditors’ professional identity I anticipate that auditors’ professional identification moderates these indirect effects, such that for auditors who strongly identify with their profession workload increases job satisfaction, whereas auditors with a weak professional identity are more likely to experience less work-life conflict, which is ultimately associated with turnover intention. Regression analyses were used to analyze data from surveys which were completed by 213 auditors in the Netherlands. I did not find any evidence for the proposed moderating effect of auditor professional identification. The results do indicate, however, that workload indeed increases work-life conflict and that workload indirectly increases turnover intention through work-life conflict. The results further indicate that workload decreases job satisfaction, and that job satisfaction decreases turnover intention. However, no indirect effect of workload on turnover intention through job satisfaction was found. All in all, these results partly confirm my expectations and highlight the negative impact of high auditor workload for auditor turnover.
Keywords: auditors, workload, work-life conflict, job satisfaction, professional identification, turnover
2
Introduction
A rising problem for audit firms in the Netherlands as well as in other countries is a shortage of auditors, illustrated by the fact that job offerings from audit firms were open for more than three months for 55,8% of the cases (Accountant.nl, 2017). One of the main reasons for the shortage of auditors is high voluntary turnover; in the Netherlands 46% of auditors of the four biggest audit firms leave the firm within three years (Riskmagazine.nl, 2017). Audit firms want to limit the amount of voluntary turnover within their firm because it is very costly; firms can incur costs up to 200% of an employee’s annual salary (Allen, Bryant and Vardaman, 2010). In addition, employee turnover can also lead to intangible costs, such as the loss of tacit knowledge and social capital (Dess and Shaw, 2001), reduced customer satisfaction (McElroy, Morrow and Rude, 2001) and turnover contagion (Felps et al., 2009). Lastly, when senior auditors leave the firm, they are often replaced by junior auditors who lack the experience and skill of senior auditors, which likely reduces audit quality (Chi et al., 2012).
While money is still important for auditors to keep working at audit firms, audit firms find it is no longer sufficient to make auditors loyal to their firm (Taylor and Cosenza, 1998). Instead, auditors increasingly have other expectations than just earnings, such as a challenging environment, career development and personal growth and work-life balance (Huang, Lawler and Lei, 2007). One aspect of the job of an auditor which has an impact on these factors is workload. Auditors in the Netherlands often have to work an average of 60 hours per week during busy-season (i.e. the first four months of the year; Heijden, 2018). Research on the relationship between workload and turnover intention has found mixed results. In their meta-analysis Bowling et al. (2015) found a moderately positive relation between workload and turnover intention. However, other studies – based on the challenge-hindrance occupational stressor model (e.g. Cavanaugh et al., 2000) – find that workload may sometimes act as a challenge stressor, increasing learning and growth opportunities (LePine, Podsakoff and LePine, 2005). Whereas at other times it may act as a hindrance stressor, increasing strains such as fatigue and exhaustion (Webster, Beehr and Love, 2011). Interestingly, research has shown that challenge stressors will decrease turnover intention, whereas hindrance stressors will increase turnover intention (Podsakoff, LePine and LePine, 2007; Webster, Beehr and Love, 2011). The present study therefore examines why and when auditor workload increases or decreases auditors’ turnover intention.
Research demonstrates that work-life conflict is one the most important reasons for individuals to intent to turnover and to actually turnover (Greenhaus et al., 1997; Hom and
3
Kinicki, 2001; Huffman, Casper and Payne, 2013; Rubenstein et al., 2018). Individuals who experience conflict between their work life and their non-work life may choose to leave their job, in order to reduce or eliminate the conflict (Felstead et al., 2002). Research further demonstrates that auditors who experience work-life conflict are more likely to turnover (Pasewark and Viator, 2006), while auditors who feel that their work life and non-work life are in balance are less likely to leave their job (Huang, Lawler and Lei, 2007). Pasewark and Viator (2006) argue that work-life conflict occurs more frequently in professions with high client contact, such as the accountancy profession, since professionals need to adjust their schedules to the needs of the client. In addition, auditors in general have relatively high workloads, approximately five hours above the threshold, but especially during busy season workload becomes exceptionally high (Persellin, Schmidt and Wilkins, 2014). This can cause conflict between work life and non-work life, since little time remains for auditors to spend on their personal life. In addition a high workload can result in strains, such as tension, anxiety, fatigue, depression, apathy, and irritability (Greenhaus and Beutell, 1985). As a result, these strains can be carried from work life to non-work life and this causes conflict (Brief et al., 1981; Greenhaus and Beutell, 1985). It is therefore not surprising that research found evidence for a positive relationship between workload and work-life conflict among auditors (e.g. Greenhaus et al., 1997; Pradana and Salehudin, 2015).
While high auditor workload may have a negative impact on auditors by increasing work-life conflict, it may also have a positive effect by increasing job satisfaction. In accordance with the transactional theory of stress (see Cavanaugh et al., 2000; LePine, Podsakoff and LePine, 2005), workload can increase auditor job satisfaction because auditors may view it as a challenge which needs to be overcome in order to improve one’s skills and gives a sense of personal accomplishment. In this regard, although LePine, Podsakoff and LePine (2005) acknowledge that workload may have a negative effect by increasing strain (e.g. health complaints, anxiety, frustration), they also found that the positive effects may very well outweigh the negative effects. Furthermore, Webster, Beehr and Love (2011) found evidence that individuals do not always appraise workload as a challenge, but sometimes also a hindrance. Hindrance stressors have been shown to decrease job satisfaction, because it causes strains without having positive effects (Podsakoff, LePine and LePine, 2007). Prior studies among auditors largely support the argument provided by Webster, Beehr and Love, in demonstrating a negative relationship between workload and job satisfaction (Pradana and Salehudin, 2015; Persellin, Schmidt and Wilkins, 2014). Whether auditor job satisfaction
4
increases or decreases is especially relevant, since auditors who are satisfied with their job are less likely to turnover (Pasewark and Viator, 2006; Pradana and Salehudin, 2015).
This brings up the question in which situations auditor workload increases auditor turnover intention and when it decreases auditor turnover intention. This study proposes that auditors’ professional identification acts as a moderator, because auditors with high professional identification may experience less work-life conflict from high levels of workload. I expect professional identification moderates this relationship because for auditors who strongly identify with their profession, ‘work life’ is a more important part of who they are than for low identifiers (Rothbard and Edwards, 2003). As a result, they might be less likely to experience work-life conflict from high levels of workload (Allen et al., 2016). In addition, Thoits (1983) already theorized that for individuals who identify highly with their profession, professional identification functions as a buffer against strains. This is also consistent with job demands-resources theory (Demerouti et al., 2001), which categorizes professional identification as a job resource that can buffer against strains (e.g. Xanthopoulou et al, 2013). Thus, auditors with high professional identification experience less strains when workload is high, which decreases the conflict between work life and non-work life (Greenhaus and Beutell, 1985).
I similarly propose that auditors’ professional identification also moderates the relationship between workload and job satisfaction. Socially identified auditors attribute meaning and purpose to their overall life through their job, and if they perform well at their job this can increase their well-being, esteem and pride (Caza and Creary, 2016). Therefore, they will be more likely to perceive high workload as a challenge which needs to be overcome for them to grow, and this has many positive effects, including increased job satisfaction (Podsakoff, LePine, and LePine, 2007). In contrast, auditors who do not identify strongly with their profession will benefit less from accomplishments in their job, and they will also be less interested in growing as an auditor and in increasing their reputation. They will mostly perceive high workload as negative because of increased strains, and they will therefore appraise it as a hindrance, which will decrease job satisfaction (Webster, Beehr and Love, 2011).
The concepts and anticipated relationships discussed can also be summarized in the following research question: To what degree can professional identification moderate the effects of high workload on job satisfaction and work-life conflict, and how will this impact turnover intention of auditors?
5
This study will offer several contributions. First, this study will show whether, in general, auditors in the Netherlands become more or less satisfied with their job when workload is high. This is especially pressing since research regarding workload and job satisfaction has found mixed results (Podsakoff, LePine, and LePine, 2007; Webster, Beehr and Love, 2011). Second, this study responds to the request by Bakker and Demerouti (2017) to uncover conditions under which job demands are perceived as hindrances versus challenges, by testing the moderating effect of professional identification. This relationship has not been researched before, while this can further extend our understanding of the job demands-resources theory (Demerouti et al., 2001) and the transactional theory of stress (Cavanaugh et al., 2000; Lazarus and Folkman, 1984). Lastly, it can help regulators and practitioners to prevent voluntary turnover, increase job satisfaction and limit work-life conflict of auditors by implementing regulations and practices to increase auditors’ professional identification.
The remainder of this paper is divided in the following sections: First, a theoretical framework will be provided. This will be followed by the method section. Then the results of the regression analyses will be discussed. Lastly, this paper will conclude with a discussion and I will provide limitations of this study and possibilities for future research.
Theoretical framework
The Challenge-Hindrance Occupational Stressor Theory (hereafter: Stressor Theory) was first developed by LePine, Podsakoff and LePine (2005) because of a lack of consensus in research regarding the relationship between stress and performance. Based on the work of Lazarus and Folkman (1984) and Cavanaugh et al. (2000), this theory categorizes stress in two categories; challenge stressors and hindrance stressors. Challenge stressors, such as workload, are appraised as positive because they can promote personal growth and trigger positive emotions. Hindrance stressors, such as role ambiguity, are appraised as negative because they can harm personal growth and trigger negative emotions. In their study, LePine, Podsakoff and LePine intent to identify a priori which stimuli are appraised as challenge stressors and which as hindrance stressors. They argue, based on expectancy theory (Vroom, 1964), individuals appraise stressors as a challenge stressor when they perceive that it is highly likely that they are able to meet the challenge by increasing effort, and when they expect to receive a highly valued ‘reward’ in the form of personal growth and possible formal recognition. Challenge stressors also lead to strains such as fatigue and exhaustion, but ultimately they still increase organizational outcomes because the reward outweighs the increased strains (Podsakoff, LePine
6
and LePine, 2007). However, Webster, Beehr and Love (2011) argue that stimuli cannot be appraised a priori as either a challenge stressor or a hindrance stressor. For example, auditors might appraise workload as a hindrance when they do not attribute high value to the rewards that occur when they meet the demand or if they feel that no amount of effort will increase the likelihood of meeting the demand.
Workload
Workload has both objective characteristics, i.e. work hours, as well as subjective characteristics, i.e. perceived workload (Ganster, Fox and Dwyer, 2001; Spector and Jex, 1998). Simply looking at the amount of hours someone works is misleading, since auditors who have the same amount of work hours might experience a different degree of workload, e.g. because of the nature of the work that has to be done. Instead, the present study focuses on auditors’ assessment of workload, since an important principle of general stress theory (Lazarus and Folkman, 1984) as well as Stressor Theory is that perceived stressors - and not necessarily work hours - lead to changes in well-being and behavior.1
As mentioned above, workload was identified by LePine, Podsakoff and LePine (2005) as a challenge stressor. In effect, this means that auditors may value workload because it makes the job more challenging, and they have to increase their effort and find new ways of problem solving, which increases their overall skill. In addition, auditors might feel that when they successfully cope with a high workload, they will receive formal recognition in the form of promotion, a higher salary, or more responsibility. On the other hand, workload also has negative aspects since it increases strains. For example, workload can lead to exhaustion because of the hours worked, or it can lead to anxiety because auditors fear that they are not able to cope with the amount of work that needs to be done. LePine, Podsakoff and LePine (2005) argue that the positive aspects of workload will outweigh the negative aspects, however Webster, Beehr and Love (2011) argue that this is not always true. The remainder of the theoretical framework will discuss my expectations in which situations and in which manner the positive aspects of workload will outweigh the negative aspects and when the opposite is true.
Work-life conflict
Greenhaus and Beutell (1985) were one of the first to study the conflict that may arise 1 Furthermore, many auditors work four days per week and have to follow an academic study beside that. Therefore, their (perceived) workload might be just as high as auditors who work five days per week, while this is not reflected in work hours.
7
between work-life and family-life. Based on the role stress theory (Kahn et al., 1964) they defined it as “a form of inter role conflict in which the role pressures from the work and family domains are mutually incompatible in some respect. That is, participation in the work (family) role is made more difficult by virtue of participation in the family (work) role” (Greenhaus and Beutell, 1985, p.77). However, I deliberately choose to use the terminology of work-life conflict, since auditors cannot only experience an overabundance of conflicting demands with regard to family commitments, but also for other non-work commitments (Boles, Johnston and Hair, 1997). Furthermore, researchers distinguish between work life interfering with family life (WIF) or family life interfering with work life (FIW)(Amstad et al., 2011). WIF occurs when time spent at work and strains experienced at work lead to conflict in the family domain, while FIW occurs when time spent with the family and strains resulting from family life lead to conflict in the work domain. In their meta-analysis Amstad et al. (2011) found that while WIF and FIW both influence work-related outcomes, WIF was more strongly correlated than FIW to these type of outcomes (such as turnover intention). Furthermore, previous studies among auditors have also found that WIF always leads to turnover intention, while FIW does not lead to higher turnover intention (Greenhaus, Parasureman and Collins, 2001; Pasewark and Viator, 2006). Therefore, for the purpose of this study, work-life conflict is defined as a conflict in non-work life caused by demands from the job (WIF) as perceived by the auditor.
One of the main reasons why auditors may experience work-life conflict is because audit firms are based on client service, which lead to very demanding jobs with long hours, especially during “busy season” (Church, 2014; Heijden, 2018; Levy et al., 2011; Reinstein, Sinason and Fogarty, 2012). While most audit firms recognize this problem and have formal policies to enhance balance between work and social life of employees, there is a gap between policies and practice because there is a culture within audit firms to meet professional targets, even at the cost of one’s social life (Ladva and Andrew, 2014). An additional reason for junior auditors to experience work-life conflict exists because they have to attain professional qualifications while at the same time having to complete the normal tasks of their job (Anderson-Gough, Grey and Robson, 2001).
Unsurprisingly therefore, workload is one the main antecedents for work-life conflict for auditors, and this occurs because of two reasons (Frone, Yardley and Markel, 1997). Firstly it increases work hours, and secondly it leads to higher strain (Greenhaus and Beutell, 1985; Parasuraman et al., 1996, Williams and Alliger, 1994). The long work hours make it physically impossible to attend to demands from non-work life even when one is attempting to fulfill those
8
demands (Bartolome & Evans, 1979). Consequently, the tasks that an auditor wants to perform in life outside work are not completed, which may lead to conflict. The role strain perspective for the relationship between workload and work-life conflict is based on scarcity theory (Gutek et al., 1991). Scarcity theory states that individuals only have a limited and fixed amount of resources and energy, and the fulfillment of both roles leads to depletion of available energy and resources (Geurts and Demerouti, 2003). In turn, this causes strains such as exhaustion and irritability (Allen et al., 2016) and these strains can be carried from work life to non-work life, which makes it difficult to function properly in non-work life and cause conflict (Brief et al., 1981; Greenhaus and Beutell, 1985). For example, an auditor might become irritable because of a long and demanding work day and take their frustration out on family or friends, which creates a conflict. In addition, because auditors have little energy left when workload is high, they might choose to spend their time away from work on leisure instead of the obligations demanded of them by their life outside work, which also leads to conflict (Geurts and Demerouti, 2003). This leads to the first hypothesis of this study:
Hypothesis 1: Auditor workload is positively related to auditor work-life conflict.
When auditors experience work-life conflict, they apply coping strategies, i.e. cognitive and behavioral efforts to manage stressors (Lazarus and Folkman, 1984), in an attempt to minimize or eliminate the stressors (Pearling and Schooler, 1978). One of these strategies is to leave one’s current job in order to eliminate the stress that caused the conflict in non-work life (Greenhaus, Parasuraman and Collins, 2001). Furthermore, when work demands lead to conflict in life outside work, this leads to less life satisfaction and thoughts of withdrawing from one’s job can arise (Greenhaus, Parasuraman and Collins, 2001). In addition, when an auditor cannot fulfill his or her non-work obligations, he or she may view the demands from the job as the main reason for this problem and might be inclined to switch jobs (Boyar et al., 2003). Consistent with the arguments presented above, research generally demonstrates a positive relationship between work-life conflict and turnover intention (Amstad et al., 2011). In addition, research among auditors consistently demonstrates a positive relationship between different aspects of work-life conflict and auditor turnover intention (Huang, Lawler and Lei, 2007; Pasewark and Viator, 2006). This leads to the following hypothesis:
9
As discussed I expect that workload will lead to increased work-life conflict for auditors and that work-life conflict will increase turnover intention. By extension, I therefore also expect that workload will (indirectly) increase turnover intention, through work-life conflict. Or in other words, I anticipate that work-life conflict mediates the relationship between workload and turnover intention. This is formulated in the following hypothesis:
Hypothesis 2b: Auditor work-life conflict mediates the relationship between auditor workload on auditor turnover intention.
Job satisfaction
The most commonly used definition of job satisfaction is by Locke (1976), who states that job satisfaction is “a pleasurable or positive emotional state resulting from the appraisal of one’s job or job experiences” (p. 1304). Auditors evaluate characteristics of the job differently, where some auditors might appraise certain aspects of the job as satisfactory, others might appraise the same aspects as hindering. Therefore job satisfaction is subjective and auditors can have different levels of job satisfaction for the same job (Robbins and Judge, 2015).
Research regarding the relationship between workload and job satisfaction has resulted in two contrasting views. On the one hand, Stressor Theory assumes a priori that individuals view workload as a challenge and they therefore appraise it positively which ultimately increases job satisfaction (LePine, Podsakoff and LePine, 2005). This theory is also corroborated by prior studies who find a positive relationship between workload and job satisfaction (e.g. Dwyer and Ganster, 1991). On the other hand, Webster, Beehr and Love (2011) argue that, since individuals appraise aspects of the job differently, some might also view workload as hindering. Specifically, Schlotz et al. (2004) find that individuals who regularly experience high workload will worry more (e.g. they worry that they have too much to do and too little time to do it), which increases stress and decreases job satisfaction. The ambiguity of these theories is illustrated by the fact that some studies found a positive relationship between workload and job satisfaction (e.g. Dwyer and Ganster, 1991), while others found no significant relationship (e.g. Ahuja et al., 2002) or a negative relationship (e.g. Pradana and Salehudin, 2015). Despite these ambiguous findings, research found that auditors’ workload is negatively related to job satisfaction (Pradana and Salehudin, 2015; Persellin et al., 2018). Persellin et al. (2018) argue that auditors’ job satisfaction will decrease when workload is too high because they fear audit quality will decrease when they have too little time to get
10
things done. Therefore, I similarly expect that for auditors, the increased strains in combination with lower perceived audit quality will outweigh the benefits of a high workload, and that job satisfaction will decrease when workload is high. This is formulated in the following hypothesis:
Hypothesis 3: Auditor workload is negatively related to auditor job satisfaction.
The relationship between job satisfaction and turnover behavior has been theorized for a long time (e.g. March and Simon, 1958; Price, 1977). In their meta-analysis of 147 studies, Rubenstein et al. (2018) demonstrated that prior studies have consistently found a negative association between job satisfaction and turnover intention. This indicates that auditors who get more satisfaction from their job are significantly less likely to express turnover intention. Research suggests that when an auditor experiences low levels of job satisfaction, he or she will want to get rid of these feelings and therefore thoughts of leaving one’s job arise (Mobley, 1977). In addition, auditors’ overall life satisfaction may increase as a result of high job satisfaction (Judge and Watanebe, 1993). Consequently, auditors will not want to leave their current job, because it contributes to their overall life satisfaction. In conclusion, I argue that auditors who experience high levels of satisfaction from their job will not want to leave their job, since they fear that they will not experience the same level of satisfaction when they switch jobs. Therefore, the following hypothesis is formulated:
Hypothesis 4a: Auditor job satisfaction is negatively related to auditor turnover intention. As discussed above, I expect that a high workload will decrease job satisfaction and that lower job satisfaction will increase turnover intention for auditors. Therefore, I also expect that workload will indirectly lead to high turnover intention, through job satisfaction, i.e. that job satisfaction mediates the relationship between workload and turnover intention. This is stated in the following hypothesis:
Hypothesis 4b: Auditor job satisfaction mediates the relationship between auditor workload and auditor turnover intention.
11
Professional identification
Professional identification is derived from and builds upon social identity theory, whereby individuals “perceive themselves to be members of the same social category, share some emotional involvement in this common definition of themselves, and achieve some degree of social consensus about the evaluation of their group and of their membership of it” (Tajfel et al., 1979, p.40). An auditor’s professional identification is important, since it influences work attitudes, affect and behavior (Siebert and Siebert, 2005). Furthermore, through professional identification, auditors can attribute meaning and purpose to their overall life, and it can give them a perception of how they contribute to society (Caza and Creary, 2016). In addition, the term ‘professional’ can be described as a preferable term to identify with, since it is used in society to describe individuals who are knowledgeable in their field and have skills that allow them to excel at their job and be of service to society (Larson, 1977; Wallace, 1995). Lastly, professional identification can be categorized as a personal resource since professional identification has been linked with resilience (Caza, 2007; Hobfoll et al., 2003), and research has shown that personal resources can function as a buffer against strains such as depression and anxiety (Xanthopoulou et al., 2013).
As discussed earlier, workload can cause strains which, in turn, can lead to work-life conflict. Personal resources such as professional identification can function as a buffer against these strains in several ways. Firstly, personal resources can give auditors the feeling that they are in control and that they can impact their environment successfully (Hobfoll et al., 2003). Workload can lead to a great deal of pressure for auditors, which will become especially undesirable when they feel they are not in control (Boswell, Buchanan and LePine, 2004). Simultaneously, auditors who feel they are in control will perceive demands such as workload less as a hindrance and experience less strains as a result (Crawford, LePine and Rich, 2010). Secondly, auditors who are highly professionally identified will experience less energy loss when faced with challenging circumstances because they attribute greater value to the outcome of the challenge (Caza, 2007). Lastly, auditors who identify highly with their profession are more likely to gain social support from other auditors, which in turn can buffer against strains (e.g., emotional support, social companionship, instrumental and informational support, see Haslam et al., 2005). In turn, highly identified auditors will be less likely to experience the negative emotions caused by workload and therefore they will experience less work-life conflict (Allen et al., 2016). In addition, when auditors identify themselves highly with their profession, the relative importance of their work life in comparison to non-work life is higher and they are
12
less likely to experience work-life conflict simply because work is a more important part of who they are (Rothbard and Edwards, 2003). In conclusion, I expect that highly professionally identified auditors will experience less work-life conflict when workload is high compared to low identifiers, which leads to the following hypothesis:
Hypothesis 5a: Auditor professional identification moderates the relationship between auditor workload on auditor work-life conflict such that the positive relationship is weaker when auditor professional identification is high.
As discussed, I argue that work-life conflict will lead to turnover intention and that auditors who are highly socially identified with their profession will experience less work-life conflict when workload is high than low identifiers. Thus, I also expect that the indirect effect of workload on turnover intention through work-life conflict will be weaker when auditors’ professional identification is high than when it is low. This is formulated in the following hypothesis:
Hypothesis 5b: Auditor professional identification moderates the positive indirect effect of auditor workload on auditor turnover intention (via auditor work-life conflict), such that the positive indirect relationship is weaker when auditor professional identification is high.
Strains that are caused by workload can also decrease the degree of job satisfaction of auditors (LePine, Podsafkoff and LePine, 2005). Therefore, since professional identification can buffer against strains, I expect that highly socially identified auditors will have higher levels of job satisfaction than low identifiers. Furthermore, since job satisfaction is highly subjective, I expect that workload cannot be categorized a priori as a challenge stressor (Webster, Beehr and Love, 2011). As mentioned before, Stressor Theory categorizes workload as a challenge stressor because it is assumed that individuals value the outcomes of meeting the challenge. Meeting the challenge of a high workload can be rewarding in two different ways. Firstly, a high workload creates a challenge for which an auditor might need a different, more effective work method or strategy, and this induces learning and someone might become a better auditor as a result (Kyndt et al., 2013). Secondly, since a high workload is seen as a part of the job for auditors (Persellin et al, 2014), those individuals who can still deliver high quality work when workload is high might receive formal recognition, e.g. promotion or an enhanced reputation (LePine, Podsakoff and LePine, 2005). For auditors with high professional identification, being
13
an auditor is an important part of their identity, and they will therefore put high value on becoming a better auditor, as well as having a good reputation and position. Thus, they see workload as a challenge and their overall job satisfaction will increase when workload is high. In contrast, for auditors with low professional identification, being an auditor is not as important, and they will therefore be less interested in becoming a better auditor or having a good reputation or position. I therefore argue that low identifiers will perceive workload more as a hindrance than as a challenge and their overall job satisfaction will decrease. Furthermore, as discussed, high professional identification can increase auditors perception of control. Research suggests that workload only leads to decreased job satisfaction when an individual feels her or she is not in control of the situation, while a high workload causes job satisfaction to increase for individuals who perceive that they are in control (Dwyer and Ganster, 1991). In summary, I expect that auditors with high professional identification appraise workload as a challenge because the increased motivation will outweigh the strains, while individuals with low professional identification appraise workload as a hindrance because the strains will outweigh the increased motivation. This leads to the following hypothesis:
Hypothesis 6a: Auditor professional identification moderates the relationship between auditor workload and auditor job satisfaction such that for auditors with high professional identification the relationship is positive and for auditors with low professional identification the relationship is negative.
As discussed, I argue that high job satisfaction will lead to less turnover intention for auditors and that auditors who are highly socially identified with their profession will experience increased job satisfaction when workload is high, while a high workload will decrease job satisfaction for low identifiers. Thus, I also expect a negative indirect effect of workload on turnover intention through job satisfaction for highly socially identified auditors, and a positive indirect effect of workload on turnover intention through job satisfaction for low identifiers. This is formulated in the following hypothesis:
Hypothesis 6b: Auditor professional identification moderates the positive indirect relationship between auditor workload and auditor turnover intention (via auditor job satisfaction), such that the relationship becomes negative when professional identification is high.
14
Methodology
Participants
This study relies on data gathered from auditors of Dutch audit firms who were personally approached by one of eight master students doing an internship at an audit firm. Convenience sampling was used to approach auditors. Auditors were personally contacted with the request to fill out the questionnaire. The sample consists of auditors with at least two years of experience. The web-based survey was distributed through mailings. Between April 2019 and November 2019 239 auditors received the invitation link to the survey. One week after the initial mail was sent out, they received a first reminder and a second reminder was sent out after two weeks. Of 239 auditors who initially agreed to participate, 218 (91%) completed the survey. Measures
All of the measures used in this study are adapted from scales validated in prior research. Data was collected through a survey with seven-point Likert scale questions. A seven-point Likert scale is preferable to a five-point Likert scale because it has been found to be more reliable (Allen and Seaman, 2007) and participants rate it as the most accurate and easiest to use (Finstad, 2010). In addition to the survey a cover letter was added with information regarding the topic of the survey, the questions which the participants could expect and the approximate time it takes to answer all the questions. Furthermore, participants were informed about the confidentiality of their answers.
15
Auditor turnover intention.
To measure auditor turnover intention three items were used which are adapted from Dalton, Davis and Viator (2015). This scale is preferred over the two-item scale developed by Viator (2001), because of its’ slightly higher consistency. Recently Cohen, Dalton and Harp also used this scale in a study among 184 auditors in the United States and they reported a Cronbach’s α of 0.94. The response scale ranged from strongly disagree (1) to strongly agree (7). An example item includes ‘I often think about leaving the firm I work for’. The average of the three items was calculated to produce the measure of auditor turnover intention (Cronbach’s α = 0.889).
Auditor workload.
To measure auditor workload five items from the Quantitative Workload Inventory were used, developed by Spector and Jex (1998). This scale measures perceived workload in favor of work hours, because prior research has found that subjective measures rather than objective measures are the primary reasons of behavioral outcomes (e.g. Perrewé and Zellars, 1999). In addition, auditors often work four days per week and attend classes or are otherwise occupied with their education, which means that they may work less hours per week, but their workload may still be relatively high. The response scale ranged from never (1) to very often (7). Example questions are ‘How often does your job require you to work very hard?’ and ‘How often does your job leave you with little time to get things done?’. The average of the five items was calculated to produce the measure of auditor workload (Cronbach’s α = 0.822).
Auditor job satisfaction.
To measure auditor job satisfaction three items were used which are adapted from the job characteristic model (Hackman and Oldham, 1976). The response scale ranged from strongly disagree (1) to strongly agree (7). An example item is ‘I am generally satisfied with feeling of worthwhile accomplishment I get from doing this job’. The average of the three items was calculated to produce the measure of auditor job satisfaction (Cronbach’s α = 0.850).
Auditor work-life conflict.
To measure auditor work-life conflict five items were used which are adapted from Netemeyer, Boles and McMurrian (1996). Since the items used by Netermeyer, Boles and McMurrian concerned conflict between work and family, the items were adapted to reflect more home rather than family situations. Recently, Ribeiro, Bosch and Becker (2016) similarly used
16
this scale to measure work-family conflict of woman auditors in South Africa and reported a Cronbach’s α of 0.96. The response scale ranged from strongly disagree (1) to strongly agree (7). Example items include ‘The demands of my work interfere with my private life’ and ‘My job produces strain that makes it difficult to fulfill the duties of my private life’. The average of the five items was calculated to produce the measure of auditor work-life conflict (Cronbach’s α = 0.897).
Auditor professional identification.
Auditors’ professional identification was measured using six items adapted from Bamber and Iyer (2002). Bamber and Iyer rephrased the items which were developed by Mael and Ashforth (1992) regarding organizational identification, to a professional orientation to measure professional identification. The response scale ranged from strongly disagree (1) to strongly agree (5). Example items are ‘When someone criticizes the audit profession, it feels like a personal insult’ and ‘When I talk about the audit profession I usually say ‘we’ rather than ‘they’’. The average of the six items was calculated to produce the measure of auditor professional identification (Cronbach’s α = 0.824).
Control variables
For this study I considered age, tenure, seniority, gender and marital status as control variables based on their known relationships with turnover intention and/or work-life conflict (Benson, Finegold and Mohrman, 2004; Elvira and Cohen, 2001; George and Wallio, 2017; Pasewark and Viator, 2006; Rubenstein et al., 2018). Age and tenure are both measured in years, seniority can take on a value from starter (1) to partner (6), gender is a dummy variable where 0 is a woman and 1 is a man, marital status is also a dummy variable where 0 is unmarried and 1 is married.
Data Analysis
Data analyses were performed to determine the direction and extent of the relationships of auditors’ workload, job satisfaction, work-life conflict, professional identification and turnover intention. Data analysis began with an examination of the sample characteristics (table 1). Multiple regression analyses were run to test the hypotheses, controlling for relevant variables (age, tenure, position, gender and marital status). Hypotheses 1, 2a, 3 and 4a were tested using measures of simple association and an ordinary least squares (OLS) regression. Next, to test the hypotheses 2b, 4b, 5a, 5b, 6a and 6b I used the PROCESS application, an
17
application that was developed to handle dichotomous moderating, mediating, independent, dependent and control variables (Hayes, 2017).
Results
Demographics of the sample
Table 1 presents the distribution of the characteristics of the sample. The majority of the auditors who participated in the study were male (71.8%). Furthermore, the sample consists mostly of auditors aged 30 or younger (62.9%), and they are mostly unmarried (73.2%). Lastly, most auditors of the sample had a tenure either between 2 and 4 years (44.1%) or 5 and 15 years (42.7%) and the majority of auditors have a medior or senior position in their firm (72.9%). While women only represent 28.2% of the sample, this is representative for the audit profession as a whole in the Netherlands (Accountant.nl, 2019). Overall, the sample shows that while auditors are relatively young, they have sufficient experience to give an opinion on the questions of the survey, while also providing an adequate distribution of characteristics among the auditors.
Table 1
Distribution of demographic sample
Characteristic Frequency Percentage Characteristic Frequency Percentage
Gender Tenure Male 153 71.8 <2 years 10 4.7 Female 60 28.2 2-4 years 94 44.1 Age 5-15 years 91 42.7 20-25 43 20.2 16+ years 18 8.5 26-30 91 42.7 Position 31-35 35 16.4 Starter 4 1.9 36-50 36 16.9 Junior 25 11.7 51+ 8 3.8 Medior 79 37.1
Marital Status Senior 55 25.8
Unmarried 156 73.2 Very Senior 23 10.8
Married 57 26.8 Partner 27 12.7
n=213
Table 2 Descriptive Statistics
Variable Mean Median Standard
Deviation Minimum Maximum
Workload 4.916 5.000 .882 2.0 7.0 W-L conflict 4.047 4.200 1.207 1.0 7.0 Job satisfaction 5.214 5.333 0.974 1.3 7.0 Professional ID 4.246 4.200 1.163 1.0 6.8 Turnover Intention 3.476 3.667 1.580 1.0 7.0 Gender .282 .000 .451 .0 1.0 Age 30.981 28.000 7.704 20.0 60.0 Marital Status .268 .000 .448 .0 1.0 Tenure 6.863 5.000 6.201 0.0 35.0 Position 3.700 3.000 1.242 1.0 6.0 n=213
18
Descriptive statistics and correlation analysis
Table 2 shows the mean, median, standard deviation, minimum and maximum of all variables of the model. In addition to the information given in table 1, it shows that the average age of the sample was almost 31 and that the average tenure of the sample was almost 7 years. More importantly, table 2 shows that the average auditor in my sample perceives that workload is relatively high (M=4.916), but the average auditor is also relatively satisfied with his or her job (M=5.214). Furthermore, the average work-life conflict and professional identification are both just above the midpoint (M=4.047 and M=4.246 respectively). Lastly, the average auditor expresses limited intention to turnover (M=3.476). Table 3 presents the correlation between all the variables. The inter-scale correlations show the expected direction of association and are all significant at the p < 0.01 level.
Table 3 (continued Correlation Analysis Variable 6 7 8 9 1. Workload 2. W-L conflict 3. Job Satisfaction 4. Professional ID 5. Turnover intention 6. Gender 7. Age -.155* 8. Marital Status -.096 .603** 9. Tenure -.154* .649** .477** 10. Position -.177** .796** .592** .625**
Scale reliabilities (Cronbach’s α) appear in parentheses on the diagonal
** p < .01, * p < .05 Based on two-tailed tests
prediction Table 3 Correlation Analysis Variable 1 2 3 4 5 1. Workload (.822) 2. W-L conflict .602** (.897) 3. Job Satisfaction -.209** -.358** (.850) 4. Professional ID -.057 .006 .321** (.824) 5. Turnover intention .325** .383** -.575** -.200** (.889) 6. Gender .081 .099 -.077 -.102 .102 7. Age -.225** -.130 .187** .089 -.239 8. Marital Status -.243** -.166* .267** .058 -.189** 9. Tenure -.132 -.034 .166* .199** -.163* 10. Position -.160* -.098 .217** .114 -.190*
Scale reliabilities (Cronbach’s α) appear in parentheses on the diagonal ** p < .01, * p < .05
19
Hypothesis testing
To test the hypotheses, multiple regression analyses were run. Tables 4, 5 and 6 present the results for the regression analyses of work-life conflict, job satisfaction and turnover intention respectively. Each analysis consists of different models, which will be explained hereafter. In addition, while unreported, I have also ensured that no multicollinearity exists within any of the models. Multicollinearity becomes problematic when the variance inflation factor (VIF) is greater than 10 (Myers, 1990) or the tolerance value is below 0.2 (Menard, 2002). None of the models had a VIF greater than 10 or a tolerance value lower than 0.2, which suggests that multicollinearity is not problematic for this study.
Control Variables
Models 1, 4 and 7 of tables 4, 5 and 6 respectively present the results of the baseline model which include only the effect of the control variables. Firstly, none of the control variables had a significant effect on work-life conflict. Secondly, marital status is the only control variable which has a significant effect on job satisfaction in the fourth (b = .483, p = .011), fifth (b = .444, p = .015) and sixth (b = .440, p = .016) model. This provides relatively strong evidence that married auditors experience higher levels of job satisfaction than unmarried auditors. Lastly, age is the only control variable which has a significant effect on turnover intention, and only in the tenth model (b = -.049, p = .019). This gives some, although weak, evidence that when auditors get older they are less likely to intent to turnover.
Direct effect of workload
To test hypotheses 1 and 3, regression analyses were run, which also include all control variables and professional identification, and are presented in model 2 (table 4) and model 5 (table 5) respectively. Hypothesis 1 stated that workload increases work-life conflict for auditors. In support of hypothesis 1, results indicate that workload has a significant positive effect on work-life conflict (b = .816, p < .01). Hypothesis 3 stated that workload has a negative effect on auditors job satisfaction. In support of hypothesis 3, results indicate that workload has a significant negative effect on job satisfaction (b = -.157, p = .031). Interestingly, results also indicate a significant positive direct effect of professional identification on job satisfaction (b = .252, p < .01).
While no hypothesis was formulated for the direct effect of workload on turnover intention of auditors, regression analyses were run to test this effect, and results are presented
20
in models 8 and 11 (table 6). First an analysis was run with only workload and the control variables (model 8) and then an analysis was run with workload, the control variables and all other independent variables (model 11). Evidence was found for a positive effect of workload on turnover intention in model 8 (b = .503, p < .01), while no significant effect was found in model 11. This indicates that workload increases turnover intention for auditors.
Direct effect of work-life conflict
Hypothesis 2a predicts that work-life conflict positively influences turnover intention for auditors. To test this hypothesis, regression analyses are run with only work-life conflict and all control variables and with work-life conflict, all control variables and all other independent variables, and results are presented in model 9 and model 11 (table 6) respectively. Evidence was found in support of hypothesis 2a in model 9 (b = .465, p < .01), while no significant relationship was found in model 11.
Direct effect of job satisfaction
Hypothesis 4a predicts that job satisfaction decreases turnover intention for auditors. To test this hypothesis, regression analyses are run with only job satisfaction and all control variables and job satisfaction, all control variables and all other independent variables, and results are presented in model 10 and model 11 (table 6) respectively. Evidence was found in support of hypothesis 4a in model 10 (b = -.911, p < .01) and in model 11 (b = -.798, p < .01).
Table 4
Regression analysis on Work-Life Conflict
Model 1 Model 2 Model 3
Intercept 4.402** -.174 4.010** Controls Gender .237 .156 .169 Age -.020 .001 .000 Marital status -.436 -.121 -.113 Tenure .022 .016 .015 Position .042 -.022 -.017 Main effects Workload .816** .809** Professional ID .036 .027
Two way interaction
Workload x Professional ID .074
R2 .044 .372 .376
ΔR2 .044 .328** .004
Based on two-tailed tests ** p < .01, * p < .05 n = 213
21
Moderating effect of professional identification
Hypothesis 5a predicts that professional identification moderates the relationship between workload and work-life conflict for auditors. To test this hypothesis an interaction term of workload and professional identification was created and included in model 3 (table 4). No evidence was found in support of this hypothesis (b = .074, p = .275). Therefore, professional identification does not moderate the relationship between workload and work-life conflict for auditors. Nevertheless, a positive relationship is still found between workload and work-life conflict.
Hypothesis 6a states that professional identification moderates the relationship between workload and job satisfaction for auditors. To test this hypothesis an interaction term of workload and professional identification was created and included in model 6 (table 5). No evidence was found in support of this hypothesis (b = -.037, p = .559). Therefore, professional identification does not moderate the relationship between workload and job satisfaction for auditors. Furthermore, in this model no significant effect is found of workload on job satisfaction.
Mediating role of work-life conflict
Hypothesis 2b states that work-life conflict mediates the relationship between workload and turnover intention. Mediation occurs when three conditions are met: (1) X (workload) must predict M (work-life conflict); (2) M (work-life conflict) must predict Y (turnover intention);
Table 5
Regression analysis on Job Satisfaction
Model 4 Model 5 Model 6
Intercept 4.975** 4.744** 5.033** Controls Gender -.091 -.025 -.031 Age -.006 -.007 -.006 Marital status .483* .444* .440* Tenure .003 -.007 -.007 Position .084 .085 .082 Main effects Workload -.157* -.154* Professional ID .252** .257**
Two way interaction
Workload x Professional ID -.037
R2 .079 .187 .194
ΔR2 .079** .108** .007
Based on two-tailed tests ** p < .01, * p < .05 n = 213
22
and (3) the relationship between workload and turnover intention must be smaller when work-life conflict is included than when it is not (Kenny, Kashy and Bolger, 1998). To formally test whether work-life conflict mediates the relationship between workload and turnover intention, a regression was run in SPSS with the PROCESS tool (Hayes, 2017), with workload as the predicting variable, work-life conflict as the mediating variable, turnover intention as the outcome variable, and gender, age, marital status, tenure and position as control variables. Results show that workload is positively related to life conflict (b = .814, p < .01), work-life conflict is positively related to turnover intention (b = .385, p < .01), and that there was a significant indirect effect of workload on turnover intention through work-life conflict (b = .314, 95% BCa CI [.153, .484]. These results indicate mediation and therefore hypothesis 2b is accepted.
Mediating role of job satisfaction
Hypothesis 4b states that job satisfaction mediates the relationship between workload and turnover intention for auditors. To test this hypothesis the same regression analysis has been run as for hypothesis 2b, with the exception that job satisfaction is the mediating variable instead of worklife conflict. Results show that workload is negatively related to job satisfaction (b = -.172, p = .025), and job satisfaction is negatively related to turnover intention (b = -.861, p < .01). However, there is no significant indirect effect of workload on turnover intention through job satisfaction, since the confidence interval includes zero (b = .148, 95% BCa CI [-.026, .321]. Since the third condition is not met, no mediation occurs, and hypothesis 4b is rejected.
Table 6
Regression analysis on Turnover Intention
Model 7 Model 8 Model 9 Model 10 Model 11
Intercept 4.646** 1.923* 2.600** 9.180** 6.808** Controls Gender .240 .186 .130 .158 .098 Age -.043 -.031 -.034 -.049* -.040 Marital status -.272 -.079 -.069 .169 .269 Tenure -.001 -.003 -.011 .002 -.001 Position .046 .007 .026 .122 .090 Main effects Workload .503** .237 W-L conflict .465** .155 Job satisfaction -.911** -.798** Professional ID -.037 R2 .065 .138 .186 .356 .399 ΔR2 .065* .073** .121** .290** .334**
Based on two-tailed tests ** p < .01, * p < .05 n = 213
23
Full model
Hypothesis 5b states that professional identification moderates the effect of workload on turnover intention through work-life conflict, and hypothesis 6b states that professional identification moderates the effect of workload on turnover intention through job satisfaction. To test these moderated mediation relationships, regressions were again run in SPSS with the PROCESS tool (Hayes, 2017), with workload as the predicting variable, work-life conflict and job satisfaction as mediating variables, professional identification as the moderating variable, turnover intention as the outcome variable, and gender, age, marital status, tenure and position as control variables.. Results show that in the full model workload is significantly related to work-life conflict (b = .809, p < .01), but the interaction term of workload and professional identification is not significantly related to work-life conflict (b = .074, p = .275). Furthermore, work-life conflict is not significantly related to turnover intention (b = .149, p = .115). Table 7 presents the indirect moderated effect of workload on turnover intention through work-life conflict. The effect is shown at the mean level of professional identification and one standard deviation above and below the mean. At each level of professional identification, the confidence interval includes zero. Therefore, hypothesis 5b is rejected.
Results also show that there is a significant relationship between workload and job satisfaction (b = -.154, p = .036), but the interaction term of workload and professional identification is not significantly related to job satisfaction (b = -.037, p = .559). Further, job satisfaction is significantly negatively related to turnover intention (b = -.815, p < .01). Table 8 presents the indirect moderated effect of workload on turnover intention through job satisfaction. The effect is shown at the mean level of professional identification and one standard deviation above and below the mean. At each level of professional identification, the confidence interval includes zero. Therefore, hypothesis 6b is rejected.
action
Table 7
Indirect moderated effect of workload on turnover intention through WL-conflict
Level of Prof ID Effect SE LLCI ULCI
3.083 .107 .077 -.022 .278
4.246 .120 .080 -.027 .286
5.409 .133 .086 -.028 .308
Table 8
Indirect moderated effect of workload on turnover intention through job satisfaction
Level of Prof ID Effect SE LLCI ULCI
3.083 .091 .135 -.162 .373
4.246 .125 .085 -.038 .293
24
Discussion and Conclusion
Audit firms increasingly face a problem of a shortage of auditors, which is primarily caused by high turnover rates (Riskmagazine.nl, 2017). One of the reasons why auditors decide to leave their job, is because of high workload, especially during busy season (Persellin, Schmidt and Wilkins, 2014). Prior studies among auditors have found that thoughts of leaving their jobs arise when workload is high because it causes a conflict in their social life or because their job satisfaction decreases (e.g. Greenhaus et al., 1997; Pradana and Salehudin, 2015). On the other hand, prior studies have also found that some individuals appraise workload as a challenge, which increases job satisfaction and ultimately decreases turnover intention (LePine, Podsakoff and LePine, 2005). Hence, the purpose of this study was to examine under which circumstances workload increases or decreases turnover intention of auditors. In order to find out what those circumstances are, several hypotheses have been formulated and based on those hypotheses, regressions have been run to study the effect of workload on turnover intention, with job satisfaction and work-life conflict as mediators, and how professional identification moderates these relationships.
Firstly, I found a significant positive effect of auditor workload on work-life conflict. This finding is in line with prior studies among auditors, who found that high workload leads to time constraints and strains, which cause a conflict in one’s personal life (e.g. Greenhaus et al., 1997; Pradana and Salehudin, 2015). In addition, I found that work-life conflict directly influences turnover intention of auditors and that workload indirectly influences turnover intention through work-life conflict. These findings are in line with Greenhaus et al. (1997), who argue that auditors express an intent to leave their organization in order to eliminate the conflict at home that was caused by a high workload. In contrast, Pradana and Salehudin (2015) did not find a significant relationship between work-life conflict and turnover intention for auditors. I argue that differences in findings of the current study and that of Pradana and Salehudin (2015) has to do with the sample. Firstly, Pradana and Salehudin studied junior auditors who mostly had little or no experience, while the majority of the auditors in the sample of the current study had at least 2 years of experience and had a medior or higher position within their firm. It might be that auditors with little experience are less inclined to express turnover intention, since they expect that they will be better able to find a balance between work life and non-work life later in their careers, while auditors with more experience feel that the conflict in their personal life can only be reduced or resolved by switching jobs. Secondly, Pradana and Salehudin conducted their study in Indonesia, while I conducted my study in the Netherlands.
25
Indonesia is much less individualistic than the Netherlands (Hofstede Insights, 2020), and research has shown that work-life conflict is more strongly related to negative work outcomes in countries that are more individualistic compared to countries that are more collectivistic (Billing et al., 2014). These findings contribute to prior research regarding turnover intention of auditors, since it shows that workload can indeed lead turnover intention through work-life conflict more than twenty years after Greenhaus et al. (1997). In addition, it demonstrates that findings regarding this relationship apply to auditors in the Netherlands, while this is not necessarily the case for auditors in all countries.
Secondly, I found a significant negative effect of auditor workload on job satisfaction. This finding is in line with prior studies among auditors (Pradana and Salehudin, 2015; Persellin, Schmidt and Wilkins, 2014). However, it is not in line with traditional Stressor Theory (LePine, Podsakoff and LePine, 2005), but rather it suggests that, in general, auditors appraise workload more as a hindrance rather than as a challenge (Webster, Beehr and Love, 2011). This finding suggests that while auditors might view workload as a challenge, the increased strains in combination with perceived reduction of audit quality outweigh the benefits, which ultimately decrease job satisfaction. Since prior studies have found ambiguous results regarding this relationship, the current study contributes to existing literature by showing that for auditors in the Netherlands, job satisfaction reduces rather than increases when workload is high. Further, I found that job satisfaction strongly effects auditors intent to turnover. This finding is in line with the meta-analysis of Rubenstein et al. (2018), who found that job satisfaction is the best predictor of turnover intention. However, in contrast to my expectations, I found no evidence that job satisfaction mediates the effect of workload on turnover intention, since there was no significant indirect effect. This adds to existing literature, by showing that while workload has a negative effect on job satisfaction, this effect does not indirectly lead to more turnover intention for auditors in the Netherlands.
Thirdly, I found no evidence of a moderating effect of professional identification on either the relationship between workload and work-life conflict, or on the relationship between workload and job satisfaction. These findings suggest that professional identification does not buffer against strains that are caused by a high workload. An explanation of the findings may be that not all personal resources buffer against strains, but that it depends on the nature of the personal resource whether, and to what degree, it can buffer against strains. For example, Xanthopoulou et al. (2013) found that self-efficacy, but not optimism, buffers against strains. They argue that self-efficacy does buffer strains, since self-efficient individuals are more
26
proactive in dealing with emotional demands, while optimism does not buffer against strains, since optimistic employees may think that whatever happens, the outcome will be positive and they are therefore less proactive. Similarly, professional identification may increase resilience and the feeling of being in control for auditors, but they might still fear that audit quality will decrease when workload is high, and they assign greater importance to high audit quality than low identifiers. The fear of diminishing audit quality may outweigh the benefits of an increase in a feeling of being in control and as a result, workload still leads to an increase in strains for highly identified auditors. In turn, these strains will lead to work-life conflict, and it also outweighs the increased motivation, and therefore job satisfaction will also decrease for auditors with high professional identification.
Practical implications
Turnover is a significant problem for audit firms, not only because it leads to a shortage of auditors, but also because it leads to intangible as well as tangible costs. Hence, in addition to having theoretical implications, the current study also has practical implications, since it can give information to audit firms regarding factors which can lead to turnover intention of auditors.
The results of the current study show that auditors who experience high job satisfaction are less likely to express turnover intention, while auditors who experience work-life conflict are more likely to express thoughts of leaving their job. Audit firms should therefore monitor employees’ job satisfaction and seek ways to increase it. Simultaneously, audit firms should monitor to what degree employees experience work-life conflict and actively seek ways to reduce the conflict. Results of my study also show that this might be done by decreasing perceived workload, since a high workload leads to an increase in work-life conflict, a decrease in job satisfaction as well as a direct increase in turnover intention. Additional (unreported) analyses also show that hours worked only have a significant positive relation with work-life conflict, and no significant relation with either job satisfaction or turnover intention. Therefore, perceived workload is indeed more relevant than actual hours worked, and audit firms should focus more on decreasing perceived workload than work hours.
Lastly, the results of the present study indicate that professional identification does not moderate the relationship between either workload and job satisfaction or workload and work-life conflict. This indicates that increasing auditors’ professional identification will not help audit firms as a strategy against the detrimental effects of high workload. However, results also
27
show that professional identification increases job satisfaction, and an additional (unreported) analysis shows that professional identification indirectly decreases turnover intention, through job satisfaction. Therefore, auditing firms should also try to increase professional identification of their employees to both increase their job satisfaction and to limit their turnover intention. Limitations and future research
The present study has several limitations. Firstly, as with all surveys, there is a possibility of bias, since it is based on subjective outcomes. In order to address this issue, I have carefully managed and designed the survey and I have ensured all participants that they would remain anonymous. Nevertheless, future studies could reinforce the findings of the present study by incorporating an experimental research design.
Secondly, the cross-sectional nature of the data entails that it is difficult to inhibit causal claims to the researched relationships. Future research may extend and deepen the findings of the present study with a field experiment with a longitudinal design.
A third limitation of this study is that it only consists of auditors in the Netherlands with at least two years of experience, and might therefore not be generalizable across countries. As discussed, Pradana and Salehudin (2015) did not find a significant effect of work-life conflict on turnover intention in their study. Future research might therefore study this effect, or other relationships that were found in this study, cross nationally. Furthermore, future studies might include auditors with less than two years’ experience as well as auditors with at least two years’ experience in their sample in order to compare the two groups.
Another limitation of this study is that the data was gathered at two different moments in time, once during busy season (i.e., the second quarter of the calendar year), and once when it was not busy season. Simultaneously, that auditors who completed the data during busy season were mostly auditors who worked for the Big Four audit firms, while the auditors who completed the survey outside busy season mostly worked for non-Big Four firms. Therefore it was impossible with the current sample to examine whether changes between these two groups resulted from these auditors working for Big Four firms, or whether they were due to the effects of busy season. Future research could gather data of both Big Four and non-Big Four firms during both busy season and non-busy season in order to examine whether effects are different for these groups.
28
occupations than the audit profession. It might for example be that for other professions, highly professionally identified employees do appraise workload more as challenge than as a hindrance. Hence, future studies could examine a similar model with a sample of employees of another profession, to study the possible moderating effect of professional identification across different professions.
Lastly, the model that is used in the present study might oversimplify reality. Future studies could include additional variables to deepen our understanding to what ultimately leads to turnover intention. Especially given the fact that some prior studies have found that workload can also increase job satisfaction and decrease turnover intention, future studies might examine different situations in which this is the case. Since Xanthopoulou et al. (2013) have found that some personal resources can buffer against strains, future studies could examine other personal resources which moderate the effect of workload on job satisfaction.
