Wibo Ypma S2944189 w.j.ypma@student.rug.nl Supervisor: D.B. Veltrop 18 January 2021 Word count: 9.668
Auditors in a Pandemic: The Effect of
COVID-19 on Work-Life Conflict and
Audit Quality Threatening Behaviour
Master Thesis MSc Accountancy
University of Groningen
Faculty of Economics and BusinessABSTRACT: While the COVID-19 pandemic turned into a worldwide crisis during 2020, inherent
lockdown measures taken by governments and audit firms led to a workspace transition for most auditors. Although digital tools seem to largely guarantee the essential role of communication during audit activities, working from home during the COVID-19 crisis involves drastic and previously unanticipated consequences for work- and non-work boundaries, threatening auditor work-life balance. Based on previous research, I expect the COVID-19 crisis to increase auditor work-life conflict, which may subsequently lead to auditors engaging in audit quality threatening behaviour (AQTB). However, I anticipate that auditor professional identification mitigates this relationship. Based on survey data from over 500 Dutch auditors, collected before and during the COVID-19 crisis, I found a positive effect between auditor work-life conflict and AQTB. However, no effect was found between the COVID-19 crisis and AQTB via work-life conflict. Surprisingly, the results indicate that auditors engage in less AQTB during the COVID-19 crisis than before. The findings have implications for both academics and practitioners. Limitations and future research opportunities are discussed.
1
INTRODUCTION
While COVID-19 transformed into a worldwide pandemic in March 2020, infecting millions of people, lockdown measures taken by governments and audit firms, in order to stop the spread, have significant consequences for the way audit services are carried out (Albitar et al., 2020). Most notable are the empty offices, as audit firms around the world were forced to obligate their auditors to work from home (PwC, 2020). This enormous workspace change involved a new way of working, as home became the new office and the use of digital communication tools became indispensable. Reviewing the first few months, the digital transformation seems to be completed with surprising effectiveness (Newrotimes, 2020). While using video conferencing and other virtual collaboration tools lead the developments, the essential role of communication during auditing activities seems to be guaranteed (Newrotimes, 2020). However, as it is the functioning of individual auditors that determines the degree of audit quality in the end (Christensen et al., 2016), it should be questioned whether working from home can replace the offices to a full extent. Nevertheless, as predictions at the end of 2020 expect the COVID-19 pandemic to continue for a significant part of 2021 (World Health Organization, 2020), inherently will working from home be proceeded.
While most auditors were used to work at the office of the audit firm in the past, nearly all of them are forced to work from home during the COVID-19 crisis (Albitar et al., 2020). This workspace transition involves drastic consequences in terms of work- and non-work boundaries of employees, as more opportunities arise for work life to spill over into non-work life, possible increasing the degree of work-life conflict experienced by auditors (Schieman, 2020; Palumbo et al., 2020). After all, working at home can give plenty opportunities for temptations and disturbance, distracting auditors from work (Delanoeije et al., 2019). Multiple factors such as parenthood, housemates and pets can make home a worse workspace compared to a relatively quiet and more structured office, which complicates the degree to which someone can focus on work (Chung et al., 2020; Olson et al., 1983). Besides, working from home may encourage people to work longer and harder, as no need to leave an office gives the opportunity and maybe even the pressure to ‘stay-on’ for a longer time (Kinman et al., 2020). Moreover, other measures taken by governments to restrict the freedom of movement during the COVID-19 crisis lead to less opportunity for other activities, such as socializing and sports. This may encourage auditors to use work as a substitution for other activities (Kinman et al., 2020). All
2
in all, this suggests the COVID-19 crisis has blurred the boundaries between work and non-work activities, in so doing inherently increasing non-work-life conflict.
Work-life conflict has been a growing subject of debate within the field of auditing last years, as both literature and practice began to recognize a deteriorated work-life balance among auditors as a possible cause for decreased audit quality (Buchheit et al., 2016; Monitoring Commissie Accountancy, 2018). Thus, although work-life conflict is a widely discussed topic in the broader debate of quality of employee working life (Crompton and Lyonette, 2006; Guest, 2002; Haider and Azmi, 2019), it is also an important topic for the quality of audit work. Previous research has demonstrated work-life conflict causes stress and emotional exhausting, leading to poor job performance (Gisler et al., 2018; Singh, 2011; Yustina and Valerina, 2018). Within the field of auditing, some studies suggest that pressures from work not only undermine work-life conflict (Hermanson et al., 2016), but also stimulate auditors to engage in audit quality threatening behaviours (AQTB) (Mohd Nor, 2011). Hence, seeking to understand how and when the COVID-19 crisis impacts audit work is not only relevant for the well-being of auditors themselves, but also for audit quality in general. AQTB is important as investors, employees, governmental authorities and other parties in society rely on the quality delivered by auditors and audit firms (Francis, 2011). Although some individual auditors are more susceptible to engage in AQTB than other auditors, Smith and Emerson (2017) state at least half of the auditors engage in some element of AQTB. AQTB can harm the validity of an audit opinion and thus the audit profession (Herrbach, 2001; Pierce & Sweeney, 2006). Besides, as stress and emotional exhausting, experienced due to work-life conflict, are indicators of AQTB as well, work-life conflict can be considered as an important driver of AQTB (Pierce & Sweeney, 2006).
The COVID-19 crisis thus presents auditors with more challenging circumstances to balance their work and non-work life, involving increased pressures for audit quality (Albitar et al., 2020). However, some auditors may be more susceptible to these deleterious consequences than others. Caza (2007) states that professionals with high professional identification, defining themselves strongly according to the norms and values of their profession, experience a relatively smaller loss of energy during challenging circumstances, as they attribute more value to the outcome of the situation. As work-life conflict depends among others on the perceived energy during work activities (Allan et al., 2007), the relationship between the COVID-19 crisis and work-life conflict might, hence, be moderated by professional identification. In
3
addition, professional identification can serve as a buffer against experienced pressures during work (Thoits, 1995). Bamber and Iyer (2002) claim auditors with high professional identification are less likely to behave in a way that could damage the profession, leading to less AQTB. Stronger professional identification will, for example, lay more emphasis on professional judgement, independency and public interest (Bamber & Iyer, 2002).
Overall, the goal of this study is to study the role of the COVID-19 crisis, through the impact on work-life conflict, on audit quality threatening behaviour (AQTB). In seeking to understand when the impact of the COVID-19 crisis is more or less severe, I examine the moderating effect of professional identification. The research question in this study is twofold and holds as follow:
(1) What is the effect of the COVID-19 crisis, through Auditor Work-Life Conflict, on Audit Quality Threatening Behaviour, and (2) what moderating role does Professional
Identification play in this respect?
This study provides insights in the effect of the COVID-19 crisis, through auditor work-life conflict, on AQTB. Research on the effects of the COVID-19 crisis on the functioning of auditors has been limited so far. Practically, this study contributes to improvement of work-life balance policies, composed by audit firms, as many concerning policies are inaccurate (Donovan, 2016; Zacharias, 2005). After all, improvement of work-life balance contributes to the necessary reformation of the audit profession (Monitoring Commissie Accountancy, 2018). Overarching, managers of audit firms can use the knowledge on the impact of the COVID-19 crisis and work-life conflict to increase the audit quality delivered by individual auditors.
In the next part of this paper a theoretical review is given in which the hypotheses are developed. After that, the methodology is described followed by the results. I conclude with the discussion and conclusion section, in which the research limitations will be addressed and recommendations for future research will be made.
4
THEORETICAL FRAMEWORK
The theoretical framework used for this study is summarized in Figure 1. First, I develop the relationship between auditor work-life conflict and audit quality threatening behaviour (AQTB). Second, I compose the relationship between the COVID-19 crisis and auditor work-life conflict. Thereafter, I explain the mediating effect of auditor work-work-life conflict within the relationship between the COVID-19 crisis and AQTB and illustrate the moderation of professional identification on all three relationships. Ultimately, the full model is presented.
Figure 1: Theoretical Framework
Audit Quality Threatening Behaviour
Audit quality has been a largely debated subject in audit literature (Cassell et al., 2020; Francis, 2004; Knechel et al., 2013). Audit services exist as a monitoring device because of the so-called ‘agency problem’: the potential conflicts of interest between owners and managers of an organization as well as those among different classes of security holders (Benston, 1985, DeAngelo, 1981; Watts & Zimmerman, 1981). In order to protect stakeholders from these potential conflicts of interest, auditors provide an independent opinion on the financial statement of the organization in accordance with general accepted accounting principles (GAAP). If the financial statements give a ‘true and fair’ view of the financial performance of the organization, the auditor provides an unqualified opinion, counting as assurance to users of the financial statements (Knechel et al., 2013). However, the extent to which this audit opinion is provided adequately and correctly, depends on the quality of the audit (Christensen et al., 2016). Knechel et al. (2013), however, claim that audit quality is “much debated but little understood”, lacking a clear definition. As audit quality does not involve precise and critical
COVID-19 Crisis Auditor Work-Life
Conflict Audit Quality Threatening Behaviour Professional Identification
5
objective criteria, appropriate to define and measure audit quality, it is sensitive for strong ambiguity (Herrbach, 2001). DeAngelo developed the most widely accepted definition and described audit quality as “the market-assessed joint probability that a given auditor will both (a) discover a breach in the client’s accounting system, and (b) report the breach” (DeAngelo, 1981, p. 186). Other definitions of audit quality focus on earnings management indicators, restatements and auditor reporting (Knechel et al., 2013).
Regardless of how audit quality is defined exactly, ultimately it is the auditor and his or her individual capacities and behaviours that, along with those of fellow auditors on the audit, determine the audit quality (Buchheit et al., 2016; Christensen et al., 2016). Therefore, I focus in this study on how auditors carry out their audit work. More specific, I use the concept of audit quality threatening behaviour (AQTB), defined as “the poor execution of an audit procedure that reduces the level of evidence gathered for the audit, so that the collected evidence is unreliable, false or inadequate quantitatively or qualitatively” (Herrbach, 2001, p. 790). AQTB focuses on the undertaking of quality reducing acts by auditors during the audit process. Examples of AQTB include reduction of work below what would normally be considered reasonable; superficial review of audit evidence; acceptance of weak client explanations; failure to research an accounting principle; failure to pursue a questionable item and false sign-off (Herrbach, 2001; Pierce and Sweeney, 2006). If conducting AQTB leads to failing to report a breach within the financial statements, the validity of the financial statements is at stake (Paino et al., 2010; Svanström, 2016). This is of high concern to stakeholders, who rely on the auditor’s opinion on the financial statements (Francis, 2011; Smith and Emerson, 2017). Other concepts in literature that correspond with AQTB are dysfunctional audit behaviour and reduced audit quality practices (RAQP) (Gundry & Liyanarachchi, 2007; Paino et al., 2010).
Literature indicates the occurrence of AQTB on an international level at many audit firms (Pierce & Sweeney, 2006). In fact, the study of Smith and Emerson (2017) demonstrates that at least half of the auditors engage in some element of AQTB. Auditors working for Big Four audit firms, however, seem to be less engaged in AQTB than auditors working for non-Big Four audit firms, as Francis (2004) claims Big Four audit firms to produce higher audit quality. Big Four auditors have ‘more to lose’ as they have larger portfolios of clients than auditors of non-Big Four audit firms. Besides, Big Four audit firms can invest more in audit technology and training programs, increasing the quality delivered by auditors (Boone et al., 2010).
6
Although engagement in AQTB will not necessarily indicate a wrong audit opinion, it may harm the validity of the opinion (Herrbach, 2001). A systematic occurrence of AQTB could have a significant negative impact on audit firms and the audit profession (Pierce & Sweeney, 2006). Previous research has revealed that AQTB often occurs due to pressures perceived by auditors (Otley & Pierce, 1996). These pressures can, within work activities, be caused by the cost-quality conflict and the identification of auditors with clients (Pierce & Sweeney, 2004; Svanberg & Öhman, 2015). However, recent research increasingly demonstrates that AQTB can also occur due to pressures between work and non-work activities (Buchheit et al., 2016; Khavis & Khrishnan, 2017).
Auditor Work-Life Conflict
Since the 1990’s and just before, literature started to pay attention to work-life conflict as the traditional role distribution among men and women began to change (Crompton, 1999). Until then, usually men had a fulltime job while women worked within the household (Fayaz Kahn & Fazili, 2016; Haider & Azmi, 2019). After this change, more interest has been given to individual employees, who often worked many hours as an economic necessity. This workload resulted often in tensions and pressures among employees as they tried to combine employment and family responsibilities (Crompton & Lyonette, 2006). From then onwards, the importance of the need to balance work and home life became recognized and the phenomenon of work-life conflict emerged (Bernardi, 1999). Greenhaus and Beutell (1985) were one of the first to study the role conflict between work and family responsibilities. They defined work-family conflict as “a form of interrole conflict in which the role pressures from the work and family domains are mutually incompatible in some respect” (Greenhaus and Beutell, 1985, p. 77). Other studies also acknowledge similar non-work demands as conflicting with work demands, such as friends, sports, education and having time for self and recovery (Boles et al., 1997; Kossek, 2016). If there is no conflict between work and non-work domains, and thus a proper distribution of time over work and life outside of work exist, this is generally referred to as work-life balance (Guest, 2002).
Work-life conflict is primarily based on role conflict. Kahn et al. (1964) define role conflict as “the simultaneous occurrence of two (or more) sets of pressures such that compliance with one makes it more difficult to comply with the other” (Kahn et al., 1964, p. 19). Interrole conflict is a form of role conflict whereby an individual participates in multiple roles (Kahn et al., 1964). According to Kahn et al. (1964), interrole conflict is experienced when the pressures in
7
one role contradict with the pressures of the other role. An explanation for role conflict can be found in boundary theory, which claims that individuals manage the boundaries between work and personal life through processes of segmenting and/or integrating these roles (Ashforth et al., 2000; Clark, 2000; Bulger et al., 2007). The most important aspect of boundary theory is the transition between different roles in life (Ashforth et al., 2000). This includes on workdays, for example, employees moving from the role of parent to spouse at home, transitioning to employee after the commute to work, and switching back to the role of parent or spouse after the commute to home (Desrochers & Sargent, 2004). Complete segmentation of both work and non-work life exists when these roles are performed with no role overlap in terms of space, time and mind (Ahrentzen, 1990). Contradicting, complete integration of both work and non-work life exists when these roles are performed with thorough overlap in terms of space, time and mind (Ahrentzen, 1990). Complete integration creates higher opportunity for work-life conflict, as overlap could cause disturbances between different roles (Kahn et al., 1964).
Work-life conflict can have a significant impact on individuals and organizations. Previous research has explicitly highlighted the link between work-life conflict and both mental and physical wellbeing of employees (Allan et al., 2007; Keyes, 2002). An imbalanced distribution between work and non-work activities not only reduces the mental and physical wellbeing, but also decreases job satisfaction and productivity (Baltes et al., 1999). In addition, a well-distributed balance between work and non-work activities can lead to more happiness and higher quality of employees’ family and overall life (Crompton & Lyonette, 2006). Within audit firms specifically, work-life balance not only leads to individual working effectiveness, but it also leads to higher organisational performance and organisational commitment (Allen et al., 2000; Crompton & Lyonette, 2006; Haworth & Lewis, 2005). Furthermore, evidence has revealed that work-life policies, designed and set up by a firm in order to stimulate employees’ work-life balance, can trigger behaviour that is beneficial to this firm (Donovan, 2016; Allan et al., 2007).
Most important, and in line with above, Yustina and Valerina (2018) found a negative relationship between work-life conflict and auditor job performance. Auditors who experience work-life conflict tend to feel emotional exhausted, resulting in negative consequences such as poor job performance (Yustina & Valerina, 2018). Besides, also non-auditing research claim reduced job performance when someone’s job does not correspond with personal desires and values (e.g. work and family), and thus results in conflicts between different domains (Cordes
8
& Dougherty, 1993; Jones et al., 2010). Similarly, other studies have showed a positive relationship between work-life balance, the counterpart of work-life conflict, and job performance. Johari et al. (2018), for example, found this relationship among teachers, Abdirahman (2018) among administrative staff and Gayathiri et al. (2013 in the context of hospital personnel. Thus, as work-life conflict has been found to generally undermine work performance, I anticipate that work-life conflict may lead auditors to engage in AQTB. Therefore, I expect a positive relationship between work-life conflict and AQTB, drawing my first hypothesis:
H1: Auditor work-life conflict is positively related to audit quality threatening behaviour
COVID-19 Crisis
After the new Coronavirus (COVID-19) was detected in China in December 2019, it spread rapidly across the globe leading to measures all around the world in an attempt to decrease the danger for human health (Welch, 2020). Many lockdown measures taken by governments and audit firms have resulted in an enormous workspace change, moving people from their offices to their home. This enormous workspace change involves an abrupt shift in work- and non-work boundaries (Schieman, 2020), resulting in rising concerns regarding the non-work-life balance and mental well-being of employees (Kulik, 2020; O’Donnell, 2020; Delanoeije et al., 2019). Although the consequences of the COVID-19 crisis for work-life conflict has been a small research subject so far, it is irrefutable that new pressures have arisen within work and non-work life, as physical boundaries between both domains have almost entirely disappeared (Davis & Green, 2020).
Undoubtedly, an important consequence of the COVID-19 crisis for auditors, is that they are required to almost fully work from home during this crisis (Albitar et al., 2020). Although little research has been done on working from home, some studies give insight in the effects of this phenomenon. Harpaz (2002) and Dockery and Bawa (2018) defined more personal time, self-management and better household relationships as a positive consequence of working from home. Contradicting, Gatrell et al. (2013) and Choudhury et al. (2020) found isolation, a lack of social contribution and inefficient working as negative consequences for employees. The latter is reinforced in case of parenthood (Gatrell et al., 2013). In addition, home life is not necessarily pleasant for all and might be fraught with tensions for many people (Moore, 2002), particularly when working from home is not a voluntary choice but required.
9
Although an unambiguous perspective regarding the effect of working from home has not been established yet, it is important to note that the positive consequences often apply to working from home in a voluntary context (Dockery & Bawa, 2018; Harpaz, 2002; Schieman, 2020). While more integration of different roles can have positive consequences for individuals when this integration is an outcome of their own personal decision (Ashfort et al., 2000; Dumas & Sanchez-Burks, 2015; Kossek et al., 2006), the opposite effect might occur when choices about the extent and timing of role integration are constrained (Schieman, 2020). In addition, an important detriment from working from home during the COVID-19 crisis, is that auditors can also be faced with other family members, including young children, working from home at the same time. This is important, since the earlier mentioned role theory of Kahn (1964) holds that individuals should keep roles separate in order to successfully conduct these roles. It is safe to say that not all homeworkers successfully negotiate the social, personal, temporal and physical transitions between the boundaries of home and work (Nippert-Eng, 2008). Separating work and personal life reduces the opportunity of distractions in non-work activities during work activities, resulting in less blurred boundaries and role conflict (Rau & Hyland, 2002). Therefore, role conflict might be a serious consequence from auditors working from home during the COVID-19 crisis (see also Schieman, 2020). Hence, I predict that levels of work-life conflict increase during the COVID-19 crisis, shaping the second hypothesis:
H2: The COVID-19 crisis is positively related to auditor work-life conflict
Moreover, and as a result of hypotheses 1 and 2, I expect the COVID-19 crisis to be positive related to AQTB via work-life conflict. This indirect effect of the COVID-19 crisis on AQTB leads to the third hypothesis:
H3: The COVID-19 crisis is positively related to audit quality threatening behavior, through auditor work-life conflict.
Auditor Professional Identification
Professional identification can primarily be related to the social identity theory of Tajfel et al. (1979). This theory states that “individuals who 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
10
membership of it” (Tajfel et al., 1979, p.40). This social consensus occurs because every person is motivated by a need for positive self-imaging, resulting in one’s desire to belong to certain social groups (Garcia-Falières & Herrbach, 2015). Furthermore, Mael and Ashforth (1992) claim that, to a certain extent, individuals define themselves according to the work they perform and according to the characteristics generally attributed to the people who perform that type of work. This type of identification is called professional identification and claims that individuals recognize themselves in the norms and values of their profession (Garcia-Falières & Herrbach, 2015). Professional identification is closely related to and an element of the overarching concept of professional commitment, which not only includes identification with the profession, but also dedication to a professional career and the acceptance of a profession’s goals and values (Sorensen & Sorensen, 1974).
As professional identification influences personal characteristics like behaviour, attitudes and affections, auditors within the entire profession experience the consequences of it (Siebert & Siebert, 2015). More in general, employees who are more identified with their profession will incorporate more commitment to comply with the role expectations of their profession (Haslam, 2001; f & Creary, 2016). This results in a higher individual’s self-esteem and higher employee satisfaction (Ashforth & Mael, 1989; Dutton et al., 1994; Caza & Creary, 2016). Besides, the term ‘professional’ is often considered as a positive term to be identified with, as it describes an individual who has a lot of skills and expertise in his or her field, requiring long and special training and education (Wallace, 1995). This means professional identification can give auditors insight into their contribution to society, adding meaning and purpose to life (Caza & Creary, 2016). Therefore, professional identification can be a source of energy, inspiration and strength (Haslam, 2001).
Moreover, Casa (2007) claims that professionals having more professional identification will, in a situation of challenging circumstances, experience a relatively smaller loss of energy, as they attribute more value to the outcome of the situation. Also, the positive relation between professional identification and an employees’ self-esteem, satisfaction and level of energy and strength, can serve as a buffer against experienced pressures during work and therefore reduce work-life conflict (Thoits, 1995). As the COVID-19 crisis and working from home will lead to more situations of challenging circumstances and involve new pressures (Welch, 2020), professional identification might mitigate these pressures, thereby mitigating the impact of the COVID-19 crisis on work-life conflict. Additionally, Rothbard and Edwards (2003) argue that
11
auditors having high professional identification attribute more relative importance to work life than to non-work life, because work is an important part of who they are. Therefore, these auditors again suffer less from challenging circumstances and pressures, whether or not caused by the COVID-19 crisis (Rothbard & Edwards, 2003). Hence, I expect a weaker positive relationship between the COVID-19 crisis and work-life conflict for auditors with higher professional identification, resulting in the fourth hypothesis:
H4: The positive relationship between the COVID-19 crisis and auditor work-life conflict is moderated by professional identification such that the positive relationship is weaker when
professional identification is high
In addition to limiting the antecedents for work-life conflict, professional identification can reduce the negative consequences of work-life conflict as well. Sorensen and McKim (2014) claim that employees with high professional commitment suffer less from the negative consequences of work-life conflict, as their ability to manage these consequences is higher. This is based on these employees’ willingness to make more sacrifices in their job, as they feel a strong sense of oneness with their profession (Sorensen & McKim, 2014). Reversely, employees with low professional identification are less able to manage work-life conflict consequences, as their willingness to make sacrifices for their job is lower (Caza, 2007). On the other hand, within auditing research, Lord and DeZoort (2001) claim auditors having higher identification with their profession are less likely to perform in a way that could damage the profession, as this goes against the professional identity. After all, individuals with high professional identification will, in a desire to facilitate the profession and achieve corresponding goals in the best possible manner, behave more according to values and goals of that profession than individuals with weak professional identification (Meyer, 1993). Therefore, auditors with stronger professional identification will place more emphasis on professional judgement, independence and public interests (Bamber & Iyer, 2002). In line with the arguments above, Broberg et al. (2018) found in their study that auditors with stronger professional identification are less likely to buckle under pressure and will have higher commitment to uphold professional standards, decreasing the engagement in AQTB. That means, auditors’ oneness with their profession can mitigate the impact of stressors, caused by work-life conflict, and prevent certain AQTB (Thoits, 1995).
12
For these reasons, I expect higher professional identification will create a stronger buffer against incentives for AQTB, such as emotional exhaustion, caused by work-life conflict. Therefore, I predict professional identification to weaken the positive relationship between auditor work-life conflict and AQTB, leading to my fifth hypothesis:
H5: The positive relationship between auditor work-life conflict and audit quality threatening behaviour is moderated by professional identification such that the positive relationship is
weaker when professional identification is high
All variables included in one research assumption, results in the last hypothesis of this study:
H6: The indirect effect of the COVID-19 crisis on audit quality threatening behaviour (through auditor work-life conflict) is moderated by professional identification
13
METHODOLOGY
Data Collection
The data used in this study was collected by a web-based survey, distributed among auditors working for different audit firms in The Netherlands. These auditors were required to have a minimum of two years of working experience as an auditor, as it is assumed that after two years of work experience, one can form an opinion on relevant occupational issues (Collins, 1993; Masihabadi, et al., 2015). Convenience sampling was used in order to approach auditors personally to participate in the study. Ultimately, the distribution of the survey occurred by e-mail. The e-mail, with invitation link to the survey, was sent out in November 2020 to 144 auditors. Awaiting their responses, auditors who did not fill in the survey yet received a reminder after one week and two weeks, respectively. A total of 127 usable responses was received, creating a response rate of 88%.
The survey existed mainly of questions based on a seven-point Likert scale. By offering seven response categories, higher reliability, consistency and validity is created (Preston & Colman, 2000). Along with the survey, a letter was sent to the respondents including information about the subject and goal of the survey, the type of questions they could expect and the duration of the survey. Besides, a statement guaranteeing the confidentiality of the answers provided was included in both the survey and the survey invitation. All questions that were used in order to measure the variables are illustrated in Appendix A.
In order to measure the impact of the COVID-19 crisis, the results of the survey were merged with the results of the same survey, conducted several times before the COVID-19 crisis. These results were collected among different auditors in April 2019, November 2019, and March 2020. As the governmental measures against COVID-19 were introduced mid-March 2020, the results of the survey in March 2020 were included to the responses collected before the COVID-19 crisis. In total, this created a sample of 515 responses, existing of a subsample of 127 responses collected during the COVID-19 crisis and a subsample of 388 responses collected before the COVID-19 crisis.
14 Measurement
Dependent variable
Audit Quality Threatening Behaviour (AQTB) is the poor execution of an audit procedure that reduces the level of evidence gathered for the audit, so that the collected evidence is unreliable, false or inadequate quantitatively or qualitatively (Herrbach, 2001). When measuring AQTB, a social desirability bias could arise. While this type of bias can never be completely ruled out when collecting this type of data, measures such as guaranteeing anonymity and confidentiality and involving a research supervisor in the distribution of the survey, minimize the likelihood of the bias. AQTB was measured by using the questions in the study of Herrbach (2001), Otley and Pierce (1995) and Pierce and Sweeney (2004). The questions represent eight audit quality threatening behaviours that may be encountered during an audit engagement. One of these behaviours reads “Acceptance of weak client explanations”. The respondents could indicate how often they were involved in the behaviours by answering according to a seven-point scale ranging from 1 (never) to 7 (very often). The Cronbach’s α was .802.
Mediator variable
Work-life conflict is a form of interrole conflict in which the role pressures from the work and non-work domains in life are mutually incompatible in some respect (Greenhaus and Beutell, 1985). Work-life conflict can be seen as an adapted concept of the traditional work-family conflict. In the past, especially in the US, life outside work was mainly focused on family. However, employees nowadays might not only feel family demands as conflicting relative to work demands, but also other non-work demands in life such as friends, sports, education and having time for self and recovery (Boles et al., 1997; Kossek, 2016). Therefore, auditor work-life conflict was measured using an adapted version of the work-family conflict scale of Netemeyer et al. (1996), in which the emphasize on family situations was replaced by home situations. A sample item from the scale is: “My job produces strain that makes it difficult to fulfill duties in private life”. Respondents were offered a seven-point scale, ranging from 1 (strongly disagree) to 7 (strongly agree), to answer the questions. The Cronbach’s α was .911.
Independent variable
The COVID-19 crisis in The Netherlands started mid-March 2020 and was going on when the survey of November 2020 was sent out. Therefore, all questions within this survey were answered during the COVID-19 crisis. Auditors were emphasized to keep this situation in mind while completing the survey. Ultimately, the results of this survey were compared to the results
15
of the same survey, conducted before the COVID-19 crisis, in order to measure the difference in auditor functioning before and during the COVID-19 crisis. Therefore, the COVID-19 crisis is presented as a dummy variable, where 0 and 1 means the response to the survey was provided before or during the COVID-19 crisis, respectively.
Moderator variable
Professional identification is the extent to which individuals define themselves according to the work they perform and the characteristics generally attributed to the people who perform that type of work (Mael & Ashforth, 1992). Auditor professional identification was measured using the scale of Bamber and Iyer (2002), who rephrased five items in their organizational identification scale to a professional orientation in order to measure professional identification. This is in line with the study of Russo (1998). Examples of the rephrased items include “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 response scale ranged from 1 (strongly disagree) to 7 (strongly agree). The Cronbach’s α was .810.
Control variables
From January until March of each year, in the so-called “busy season”, auditors are faced with higher workloads, leading to more work-life conflict and AQTB (Buchheit et al., 2016). Therefore, I control for busy season using a dummy variable, such that 0 means the response is collected outside busy season and 1 means the response is collected inside busy season. Besides, previous research has demonstrated gender might influence the ethical decision-making of auditors (Duff, 2004; Brown et al., 2016). As, consequently, ethical decision-decision-making is an important indicator of audit quality (Beattie et al., 2013), I control for gender using a dummy variable. Herewith, men are represented as 0 and women as 1. My third control variable is seniority, as Ghosh et al. (2005) illustrated higher seniority is often associated with higher audit quality.Seniority is measured using six categories, varying from starter (1) to partner (6). Lastly, I control for parental statusas having children living at home results in stronger work-family strains (Schieman, 2020). Parental status is presented as a dummy variable, where 0 means the auditor is a parent and 1 means the auditor is not a parent. For both seniority and parental status, one missing value in the sample was replaced by the mean. Initially, I also controlled for age, tenure and contract (i.e. part-time or fulltime). However, as age and tenure measure almost the same construct as seniority and contract had little effect on other variables, these constructs have been deleted from the analyses.
16 Data Analysis
Data analyses were performed to determine the direction and the extent of the relationships among the COVID-19 crisis, auditor work-life conflict, AQTB and professional identification. The data analyses started with an examination of the sample characteristics. Then, a correlation analysis was run to get an indication whether the variables are related or not. Thereafter, multiple regression analyses were conducted to test the hypotheses, while controlling for relevant variables (busy season, gender, seniority and parental status). Hypotheses 1 and 2 were examined using ordinary least squares (OLS) regression. Consequently, hypotheses 4 and 5 were examined by estimating moderated OLS regression models. Within these analyses, to make the variables included in the interaction term interpretable at a meaningful value, I mean-centered the independent variables before calculating the interaction term (cf. Aiken & West, 1991). The interaction terms are the products of these standardized variables. In addition, I controlled for the mediating effect of work-life conflict using both OLS and the regression PROCESS tool of Hayes (2018). By using the PROCESS tool, direct and indirect bootstrapped effects could be generated in a mediation model (Hayes, 2018). To test the full model, I used the PROCESS tool as well.
17 RESULTS
Demographic Sample Characteristics
Table 1 presents the distribution of the demographic sample characteristics. The sample consisted of substantially more male auditors (69.3%) than female auditors (30.7%). Although women are underrepresented in the sample, this is in line with the distribution of auditors’ gender in the Dutch auditing industry (accountant.nl, 2020). Moreover, most auditors participating in the study do not bear parentship (76.5%) and has a medior or senior position in the audit firm (36.3% and 27.4%, respectively). Lastly, more than half of the auditors participated in the study during busy season (58.8%). All in all, as the full sample contains two or more years of working experience as an auditor, it is expected that all participants are able to form a well-considered opinion on relevant working matters and provide an adequate distribution across the total population of Dutch auditors.
TABLE 1
Distribution of Demographic Sample Characteristics
Characteristic Frequency Percentage Characteristic Frequency Percentage
Gender Position
Male 357 69.3 Starter 4 0.8
Female 158 30.7 Junior 58 11.2
Parental status Medior 187 36.3
Parent 121 23.5 Senior 141 27.4
Not a parent 394 76.5 Very senior 57 11.1
Busy season Partner 68 13.2
Inside busy season 303 58.8 Outside busy season 212 41.2
N = 515
Descriptive Statistics and Correlation Analysis
In addition to the demographic characteristics given in Table 1, Table 2 provides the descriptive statistics of all variables. It confirms that, on average, the participant has a position in the audit firm between medior and senior, most closely to senior (M = 3.763). Besides, Table 2 shows that 24.7% (M = 0.247) of the sample participated in the study during the COVID-19 crisis. The average auditor in the sample perceives work-life conflict almost similar to the scale midpoint (M = 3.970), while experiencing professional identification just above the scale midpoint (M = 4.397). Furthermore, the auditors claim to be engaged in audit quality threatening behaviour (AQTB) relatively few (M = 2.375, SD = 0.794).
18
Table 3 presents the correlations among the independent, mediating, dependent, moderating and control variables. Contrary to my expectation of a positive correlation, the table shows that the COVID-19 crisis is negatively correlated with work-life conflict (r = -.142, p < .01). This means, instead of increasing work-life conflict, the COVID-19 crisis might decrease work-life conflict. Moreover, the COVID-19 crisis is also negatively correlated with AQTB (r = -.301, p < .01), which is also contrary to my expectation of a positive correlation. Work-life conflict is, in line with my prediction, positively correlated with AQTB (r = .203, p < .01). Furthermore, Table 3 shows a significant correlation between busy season and AQTB (r = .208, p < .01), seniority and AQTB (r = -.191, p < .01) and parentship and AQTB (r = .129, p < .01). Work-life conflict is significant correlated to busy season (r = .197, p < .01) and seniority (r = -.133, p < .01). With regard to the moderating variable, there is a significant correlation between professional identification and gender, seniority and parentship. Among the control variables, significant correlations exist between seniority and gender, parentship and gender and parentship and seniority.
TABLE 2 Descriptive Statistics
Variable Mean Median Standard Deviation Minimum Maximum COVID-19 Crisis 0.247 0.000 0.431 0.0 1.0 Work-Life Conflict 3.970 4.000 1.315 1.0 7.0 AQTB 2.375 2.250 0.794 1.0 5.5 Professional ID 4.397 4.400 1.128 1.0 7.0 Busy Season 0.588 1.000 0.493 0.0 1.0 Gender 0.307 0.000 0.462 0.0 1.0 Seniority 3.763 4.000 1.212 1.0 6.0 Parentship 0.765 1.000 0.424 0.0 1.0 N = 515 ** p < .01. * p < .05 Based on two-tailed tests
TABLE 3 Correlation Analysis Variable 1 2 3 4 5 6 7 8 1. COVID-19 Crisis 1 2. Work-Life Conflict -.142** 1 3. AQTB -.301** .203** 1 4. Professional ID 0.015 -0.003 -0.085 1 5. Busy Season -.684** .197** .208** 0.061 1 6. Gender .108* 0.052 -0.065 -.109* -0.060 1 7. Seniority -0.029 -.133** -.191** .181** 0.048 -.169** 1 8. Parentship 0.010 .090* .129** -.112* 0.050 .101* -.552** 1
19 Regression Analyses
Multiple regression analyses were performed in order to test the hypotheses. The results of the regression analyses for AQTB and work-life conflict are presented in Table 4 and Table 5, respectively. Both analyses consist of different models. Model 1 of both Table 4 and Table 5 includes the control variables. Model 2 of both tables contains both the control variables and the independent variables. Model 3 presents the regression analyses of all variables, including the control variables, independent variables and the interaction term of the moderator. The analyses will be explained hereafter.
Control variables
Model 1 of Table 4 and Table 5 is the baseline model and includes only the four control variables (busy season, gender, seniority and parentship). The results show that seniority had a significant negative effect on AQTB in the first (β = .-.137, p < .001), second (β = -.109, p < .01) and third (β = -.109, p < .01) model of Table 4. This means that auditors engage in less AQTB when their position in the audit firm is higher and vice versa. Besides, model 1 of Table 4 shows a significant positive effect of busy season on AQTB (β = .342, p < .001). This indicates that auditors are more engaged in AQTB during busy season. Gender, in addition, is significant negatively related to AQTB in model 1 (β = -.153, p < .05). This provides tentative evidence that women are less engaged in AQTB than men.
Within Table 5, seniority has a significant negative effect on work-life conflict in model 1 (β = -.147, p < .01), 2 (β = -.149, p < .01) and 3 (β = -.149, p < .05). This means that auditors experience less work-life conflict when their position in the audit firm is higher and vice versa. Moreover, busy season is significantly positive related to work-life conflict in model 1 (β = .550, p < .001), 2 (β = .512, p < .01) and 3 (β = .519, p < .01) as well. This indicates that auditors suffer more from work-life conflict during busy season.
Direct effect of work-life conflict
Hypothesis 1 posits that auditor work-life conflict is positively related to audit quality threatening behaviour (AQTB). To test this hypothesis, a regression analysis was run on AQTB including all independent variables and control variables. The results of this analysis are
20
presented in model 2 of Table 4. I find a significant positive relationship between auditor work-life conflict and AQTB (β = .087, p < .01). Therefore, hypothesis 1 is supported.
Table 4
Regression Analyses on Audit Quality Threatening Behaviour (AQTB)
Model 1 Model 2 Model 3
Intercept 2.719*** 2.747*** 2.796*** Controls Busy Season .342*** -.014 -.015 Gender -.153* -.136 -.137 Seniority -.137*** -.109** -.109** Parentship .023 .055 .055 Main effects COVID-19 Crisis -.520*** -.225*** Work-Life Conflict .087** .114** Professional ID -.038 -.042 Two-way interaction
Work-Life Conflict x Professional ID -.011
R2 .091 .158 .158
∆R2 .091*** .067*** .000
*** p < .001, ** p < .01. * p < .05 Based on two-tailed tests
N = 515
Direct effect of the COVID-19 crisis
Hypothesis 2 suggests a positive relationship between the COVID-19 crisis and auditor work-life conflict. To test this hypothesis, a regression analysis was run on work-work-life conflict including the COVID-19 crisis dummy, professional identification and all control variables. The results are presented in model 2 of Table 5. I did not find a significant positive relationship between the COVID-19 crisis and work-life conflict (β = -.060, p = .740). Thus, hypothesis 2 is not supported.
21 Table 5
Regression Analyses on Work-Life Conflict (1)
Model 1 Model 2 Model 3
Intercept 4.163*** 4.119*** 4.170*** Controls Busy Season .550*** .512** .519** Gender .116 .124 .117 Seniority -.147** -.149** -.149* Parentship .003 .008 .022 Main effects COVID-19 Crisis -.060 -.024 Professional ID .019 .022 Two-way interaction
COVID-19 Crisis x Professional ID .060
R2 .061 .061 .063
∆R2 .061*** .000 .002
*** p < .001, ** p < .01. * p < .05 Based on two-tailed tests
N = 515
Moderating effect of professional identification
Hypothesis 4 posits that professional identification moderates the relationship between the 19 crisis and work-life conflict. It was expected that the positive effect of the COVID-19 crisis on work-life conflict would be weaker if auditors experience higher professional identification. In order to test this hypothesis, an interaction term of the COVID-19 crisis and work-life conflict (COVID-19 crisis x professional ID) was created and included in model 3 of Table 5. However, no evidence was found to support the hypothesis (β = .060, p = .304). This means professional identification does not moderate the relationship between the COVID-19 crisis and work-life conflict.
Hypothesis 5 predicts that professional identification moderates the relationship between work-life conflict and AQTB. I expected the positive effect of work-work-life conflict on AQTB to be weaker if auditors experience higher professional identification. To test this hypothesis, an interaction term of work-life conflict and AQTB (work-life conflict x professional ID) was created and included in model 3 of Table 4. Again, no evidence was found to support the hypothesis (β = -.011, p = .723). Thus, professional identification does not moderate the
22
relationship between work-life conflict and AQTB. However, in line with results earlier provided in this study, I still found a positive relation between work-life conflict and AQTB.
Mediating role of work-life conflict
Hypothesis 3 posits that the positive relationship between the COVID-19 crisis and AQTB is mediated by work-life conflict. To test this mediation, I conducted two separate regression analyses. First, Table 6 presents the results of a regression analyses on all paths in the mediation model. All control variables were included in the analyses. Path A of the mediation model, which tested the regression of the COVID-19 crisis on work-life conflict, was not significant (β = -.055, p = .763). Path B showed the regression of the mediator, work-life conflict, on the dependent variable AQTB and was significant (β = .086, p < .01). As only one path is significant, there is no partial mediation. Moreover, while controlling for the mediator work-life conflict, the COVID-19 crisis was still a significant predictor of AQTB (β = -.532, p < .001). Therefore, there seems to be no full mediation.
Table 6
Mediating Effect of Work-Life Conflict (1)
Dependent variable AQTB Work-Life
Conflict
AQTB
Step 1 Step 2 Step 3
Controls Busy Season .019 .517** -.026 Gender -.118 .120 -.128 Seniority -.127*** -.146* -.115*** Parentship .059 .006 .058 Main effects COVID-19 Crisis -.537*** -.055 -.532*** Work-Life Conflict .086** R2 .136 .061 .155 *** p < .001, ** p < .01. * p < .05 Based on two-tailed tests
N = 515
Secondly, a regression analysis was run using the PROCESS tool with 5.000 bootstraps (Hayes, 2018). Within this analysis, I used the COVID-19 crisis as the predicting variable, work-life conflict as the mediating variable and AQTB as the outcome variable. The results are provided in Table 7 and reveal the bootstrapped 95% confidence interval includes zero [-.0387, .0274]. This means, despite the effect of the COVID-19 crisis on both work-life conflict and AQTB,
23
there is no effect of the COVID-19 crisis on AQTB through work-life conflict. Therefore, hypothesis 3 is not supported.
Table 7
Mediating Effect of Work-Life Conflict (2) Boot indirect
effect
Bootstrap SE Bootstrap 95% confidence interval Lower bound Upper bound
Work-Life Conflict -.0047 .0162 -.0387 .0274
N = 515
Bootstrap sample size is 5.000
Full Model
Hypothesis 6 suggests that professional identification moderates the effect of the COVID-19 crisis, through work-life conflict, on AQTB. In order to test this moderated mediation effect, a regression analysis was conducted using the PROCESS tool with 5.000 bootstraps (Hayes, 2018). I used the COVID-19 crisis as the predicting variable, work-life conflict as the mediating variable, professional identification as the moderating variable and AQTB as the outcome variable. For all variables, the standardized coefficients were taken. Table 8 provides the conditional indirect moderated effect of the COVID-19 crisis on AQTB, through work-life conflict. The effect is shown at the standardized mean level of professional identification (M) and one standard deviation above (+ 1SD) and below (- 1SD) the mean. The bootstrapped 95% confidence interval includes zero for all levels of professional identification [.0763, .0292; -.0403, .0277; -.0415, .0523]. Therefore, hypothesis 6 is not supported.
Table 8
Conditional Indirect Moderated Effect of the COVID-19 crisis on AQTB through Work-Life Conflict
Conditional indirect effect at Professional Identification = Mean +/- 1 SD Level of Professional
ID
Boot indirect effect
Bootstrap SE Bootstrap 95% confidence interval Lower bound Upper bound
-1.1279 (M - 1SD) -.0186 .0256 -.0763 .0292
0 (M) -.0049 .0166 -.0403 .0277
1.1279 (M + 1SD) .0064 .0220 -.0415 .0523
N = 515
24 Additional Analysis
Although I did not find a significant positive relationship between the COVID-19 crisis and auditor work-life conflict in model 2 of Table 5, this might differ per level of seniority (i.e. starter, junior, medior, senior, very senior and partner). Previous research demonstrated that, similar to my findings, auditors with a higher position in the audit firm experience less work-life conflict than auditors with a lower position in the audit firm (Buchheit et al., 2016). Auditors with lower seniority often combine their work with study activities, which can result in more challenges to find work-life-study balance (Monitoring Commissie Accountancy, 2018). More important, the degree of knowledge and skills needed to work independent from home during the COVID-19 crisis is lower in case someone has less working experience (Yang et al., 2020). This could cause auditors with lower seniority to work longer as it takes more time to perform tasks independently, possibly leading to work-life conflict. Therefore, I did an additional analysis, testing the moderating effect of seniority on the relationship between the COVID-19 crisis and work-life conflict. An interaction term of the COVID-19 crisis and seniority was created (COVID-19 crisis x seniority) and included. The results of this analysis are shown in Table 9. However, no evidence was found to determine the moderating effect of seniority on the relationship between the COVID-19 crisis and work-life conflict (β = .035, p = .554). Nevertheless, a significant negative relationship was again found between seniority and work-life conflict (β = -.173, p < .05).
Table 9
Regression Analyses on Work-Life Conflict (2)
Intercept 3.615*** Controls Busy Season .518** Gender .120 Parentship .019 Main effects COVID-19 Crisis -.022 Seniority -.173* Two-way interaction
COVID-19 Crisis x Seniority .035
R2 .062
∆R2 .001
*** p < .001, ** p < .01. * p < .05 Based on two-tailed tests
25
DISCUSSION AND CONCLUSION
While the COVID-19 pandemic turned into a worldwide crisis during 2020, inherent lockdown measures taken by governments and audit firms led to a workspace transition for most auditors. Although digital tools seem to largely guarantee the essential role of communication during audit activities (Luo & Malsch, 2020), working from home involves drastic consequences in terms of work- and non-work boundaries, threatening auditor work-life conflict (Schieman, 2020). As, ultimately, work-life conflict is associated with audit quality (Khavis and Krishnan, 2017), the purpose of this study was to examine the effect of the COVID-19 crisis on audit quality, through work-life conflict. Thereby, audit quality was measured by engagement in audit quality threatening behavior (AQTB). Multiple hypotheses were formulated during this study and regression analyses were used to test these hypotheses. Auditor professional identification was considered as a moderator for all relationships. Based on survey responses from over 500 auditors, collected across multiple moments in time, I found no relationship between the COVID-19 crisis and AQTB via work-life conflict. These results have implications for both academics and practitioners and include limitations, on which can be built in future research.
Firstly, the results demonstrated a significant positive relationship between auditor work-life conflict and AQTB. This is in line with the studies of Khavis and Krishnan (2017, 2020), who found in both studies a negative relationship between auditor work-life balance and audit quality. As work-life balance is the counterpart of work-life conflict, the negative relationships found by Khavis and Krishnan (2017 & 2020) correspond with the positive relationship found in this study. Also, Yustina and Valeria (2018) found a negative relationship between work-life conflict and auditor job performance, which is comparable to my finding. Therefore, my finding not only strengthens the results of the aforementioned studies, but also reacts to the request of Knechel et al. (2013) for extension of research on the antecedents of audit quality.
Secondly, I did not find a significant relationship between the COVID-19 crisis and auditor work-life conflict. As there is high integration of different roles in life when working from home during the COVID-19 crisis, this finding is in contrast with the role theory of Kahn (1964), which claims that individuals should keep roles separate in order to successfully conduct these roles. However, while role integration in terms of space might be high during
26
the COVID-19 crisis, the degree to which working from home leads to integration of different roles in life in terms of time and mind could be mild, weakening the total impact on work-life conflict. Besides, Harpaz (2002) and Dockery and Bawa (2018) define more personal time, self-management and better household relationships as positive consequences of working from home, indicating the workspace transition due to the COVID-19 crisis can also have benefits for work-life proportions. As, in general, there is little known about the consequences of the COVID-19 crisis and working from home for work-life conflict, more research is needed to establish an unambiguous perspective on this topic.
Thirdly, I found no significant evidence for either a partial or a full mediating effect of work-life conflict on the relationship between the COVID-19 crisis and AQTB. Although I expected a mediating effect based on related literature, literature did not pay specific attention to this mediation yet, making it an underexposed research topic. Moreover, my finding is in line with the findings discussed in the paragraphs above, where a positive relationship was found between auditor work-life conflict and AQTB but no effect was found between the COVID-19 crisis and auditor work-life conflict. However, a surprising direct significant negative effect was found between the COVID-19 crisis and AQTB. This indicates that the COVID-19 crisis decreases the degree to which auditors are engaged in AQTB, and that there is presumably a mechanism other than auditor work-life conflict mediating this effect. While one would expect a positive effect between the COVID-19 crisis and AQTB based on pre-COVID-19 research, recent studies, however, increasingly start to recognize the positive consequences of working from home for employees’ productivity. Bloom (2020), for example, found out communication software has improved during the COVID-19 crisis, while, at the same time, employees report a quieter and more convenient working environment, with less distraction caused by co-workers. Alipour et al. (2020) found working from home employees spent 1.4 more days working each month due to flexible working hours. These findings might be an indicator for the decrease of AQTB of auditors during the COVID-19 crisis.
Furthermore, the results of this study show professional identification neither moderates the effect of the COVID-19 crisis on auditor work-life conflict nor the effect of auditor work-life conflict on AQTB. The former finding is rather novel, as no studies in literature included the moderating effect of professional identification on the relationship between the COVID-19 crisis and work-life proportions yet. The latter finding is rather surprising, as it is in contrast with the findings of Lord and DeZoort (2001) who state auditors with high professional
27
identification are less likely to perform in a way that could damage the profession. An explanation for my finding could be that professional identification functions as a buffer against certain incentives that provokes AQTB, but not against the incentives released by work-life conflict. Thus, professional identification does not decrease the positive relationship between work-life conflict and AQTB.
In an additional analysis, I explored whether the effect of the COVID-19 crisis on work-life conflict differed for auditors with a lower position in the audit firm compared to auditors with a higher position. However, using seniority as a moderator, I found no evidence indicating that the effect of the COVID-19 crisis on work-life conflict is stronger for auditors with a lower position in the audit firm than for auditors with a higher position in the audit firm. This is contrary to my expectation, where, based on Buchheit et al. (2016), I assumed less senior auditors to suffer more from the COVID-19 crisis in terms of work-life conflict as their ability to work independently from home might be lower and they often have additional study obligations. Despite, the degree to which auditors experience work-life conflict as a result of the COVID-19 crisis is not dependent on auditors’ position in the audit firm.
Practical Implications
Besides the theoretical implications, this study includes practical implications. In the past, the Dutch audit industry has repeatedly received criticism on both the delivered audit quality and the workload of (young) auditors. The Monitoring Commissie Accountany (MTC) and the Commissie Toekomst Accountancysector (CTA), respectively, composed reports in which the current state of audit quality was criticized and recommendations were made to improve the industry. Both reports claimed that substantial progress still has to be made. Therefore, and due to the research on the consequences of the relatively new COVID-19 crisis for the audit profession, this study provides hands-on guidance for audit firms.
The results of this study show that auditors experiencing higher work-life conflict are more engaged in AQTB than auditors who experience lower work-life conflict. This means AQTB will be lowered, and thus audit quality will increase, when auditors perceive a balanced work-life situation. Therefore, it is valuable to audit firms to monitor the work-work-life balance of their employees and to reduce work-life conflicts. This can, for example, be done by implementing or redesigning the work-life balance policy. Another way to improve the balance between working life and private life, and in line with the recommendations of the CTA (2020), is
28
reducing the workload and implementing training programs. After all, the CTA, together with the MCA (2018), plead for a more appealing working environment to maintain the attractivity of the profession.
Moreover, this study is one of the first to measure the consequences of the COVID-19 crisis for the audit profession. As nearly all auditors are enforced to work from home during this pandemic, the impact on the profession is substantial. However, so far less evidence exists about the effects of the workspace change on the functioning of the auditor. As I did not find a relationship between the COVID-19 crisis and auditor work-life conflict, the impact of the shift of working at the office towards working from home can assumed to be limited for auditors’ overlap in work and private life. In addition, this study suggests auditors are less engaged in audit quality behavior during the COVID-19 crisis. Therefore, working from home seems to increase the quality delivered by auditors as well. Both findings strengthen the degree to which working from home can replace the audit offices, while Luo and Malsch (2020) already proved that digital tools guarantee the essential role of communication during audit activities. These findings are beneficial to the audit profession, which seems to become suitable for home working during a pandemic more and more.
Furthermore, this study indicates that professional identification might be of low importance compared to other factors when performing high-quality audits. As such, my findings show that professional identification neither moderates the effect of the COVID-19 crisis on auditor work-life conflict nor the effect of auditor work-life conflict on AQTB. This implicates attempting to increase auditor professional identification might not be an appropriate strategy to prevent AQTB caused by work-life conflict.
Limitations and Future Research
This study contains several limitations which could influence the findings. First of all, it has limitations inherent to survey research. Although surveys can easily capture individual’s emotions and attitudes (Spector, 1994), outcomes might be subjective due to self-report measurement. Besides, respondents could tend to provide socially desirable responses. To address this possible bias, the survey was carefully designed and anonymity was guaranteed to the participants. A disadvantage of this anonymity is the impossibility to test for non-response bias as it is unknown which response belongs to which respondent.
29
Secondly, although I control for the effects of busy season, gender, seniority and parentship, some biases could exist regarding the survey sample. To measure the effect of the COVID-19 crisis, the survey was merged with the same survey distributed before the pandemic across three moments in time and among different participants. As the total amount of survey responses from the surveys conducted before the COVID-19 crisis is equal to 388, while the survey conducted during the COVID-19 crisis resulted in 127 responses, there might be a disproportionate distribution of both subsamples. Besides, as different participants were used each time the survey was distributed, a bias might exist when comparing the responses of the three surveys before the 19 crisis with the responses of the survey during the COVID-19 crisis. Therefore, future research could use a more equal amount for both the ‘COVID-COVID-19 crisis subsample’ and the ‘non-COVID-19 crisis subsample’, including the same participants. Moreover, the total sample consisted primarily of medior and senior auditors (36.3% and 27.4%, respectively). This unequal proportion could lead to an undercoverage bias as junior and very senior auditors are under-represented. Other researchers could include a more proportional sample in terms of seniority when replicating this study.
Thirdly, this study has been conducted using Dutch auditors working for audit firms located in The Netherlands. During the COVID-19 crisis, measures taken by government and audit firms, and inherently the degree to which auditors work from home, differ throughout countries based on the pace at which the virus spreads (Anderson et al., 2020). The difference per country regarding the degree to which auditors work from home, may lead to a weaker or stronger effect on auditor work-life conflict and AQTB when this degree is lower or higher, respectively. In addition, the perception of work-life conflict differs for auditors throughout countries as a result of cultural differences. Auditors in other countries might value the distribution of multiple roles in life differently than Dutch auditors due to a higher or lower focus on working life compared to non-working life (McGinnity & Whelan, 2009). Therefore, the results of this study cannot directly be generalized to auditors living in other countries in the world. Future research could replicate this study by using a sample of auditors from different countries.
Finally, the COVID-19 crisis is a relatively new and still proceeding phenomenon, where a lot remains to be discovered. My findings suggest a surprisingly negative relationship between the pandemic and audit quality, without mediation of auditor work-life conflict. Future research should further study this negative relationship by implementing different mediating
30
mechanisms and, in general, explore other consequences of the COVID-19 crisis for the audit profession as well.
