The impact of codes of ethics and experience on auditor judgments

The Authors

Gary Pflugrath, School of Accounting, University of New South Wales, Australia

Nonna Martinov-Bennie, Discipline of Accounting, The University of Sydney, Australia

Liang Chen, School of Accounting, University of New South Wales, Australia

Acknowledgements

The authors would like to acknowledge the support of the Institute of Chartered Accountants in Australia for assistance during this study, as well as colleagues at the University of New South Wales and The University of Sydney, and participants and discussant at the 2006 AFAANZ Conference, for their comments and feedback on the paper.

Abstract

Purpose – The purpose of this study is to investigate the impact of the presence of a code of ethics on the quality of auditors' judgments, within the context of the new International Standard on Quality Controls 1 (ISQC1).

Design/methodology/approach – A sample of 112 professional accountants and auditing students was employed to investigate the effect of the presence of a code of ethics (operationalised as the presence vs absence of an organisational code of conduct) on the quality of audit judgments, pertaining to an inventory writedown, using a 2 × 2 full factorial “between-subjects” experimental design.

Findings – The results of this study indicate that the presence of a code of ethics has a positive impact on the quality of the judgments made by professional accountants, but not on students. This suggests that it is the code of ethics, in the context of greater general experience that leads to higher quality of judgments.

Practical implications – The results suggest that the requirements of ISQC1 are relevant to the quality control of accounting firms and have potential to positively impact the quality of audit performance.

Originality/value – This is the first paper to examine the impact of the presence of a code of ethics within an audit context. It is the first time that the interactive effects of the code of ethics and technical competency, which together form an integral part of standard-setters' quality control standards, upon the quality of auditor judgments has been investigated.

Article Type:

Research paper

Keyword(s):

Ethics; Competences; Auditing; Standards.

Journal:

Managerial Auditing Journal

Volume:

Number:

Year:

2007

pp:

566-589

Copyright ©

Emerald Group Publishing Limited

ISSN:

0268-6902

Introduction

Recent high-profile corporate collapses, such as Enron and WorldCom, have brought into question the status and credibility of the accounting profession, especially auditors, “with allegations of accountants' violations of public trust” (Chan and Leung, 2006, p. 436). In 1999, Steve Samek, Country Managing Partner, USA, for Arthur Andersen, talking about the firm's independence and ethical standards, noted that “The day Arthur Andersen loses the public's trust is the day we are out of business”. While all of the major accounting firms have been implicated with different corporate collapses and accounting scandals, it is Andersen which has since ceased to exist. A consequence of these apparent violations has been increased government intervention in the regulation of the profession (Chan and Leung, 2006).

In an aim to restore public confidence, regulators have embarked on a number of measures, including the introduction of new quality control standards (e.g. International Standard on Quality Controls 1 (ISQC1, 2005)) for accounting firms, which prescribe both ethical and technical requirements. The importance of possessing both technical and ethical competency is well illustrated by the alleged role of auditors in a number of these recent high-profile corporate collapses, and highlights the point that auditors need to be technical and ethical experts when auditing financial reports (Gaa, 1994). This study aims to investigate the impact of one aspect of the ethical environment (i.e. the code of ethics) on the quality of auditors' judgments, within the context of the new ISQC1.

The International Auditing and Assurance Standards Board (IAASB) has recently introduced ISQC1[1], which aims to help firms establish a system of quality control for audits and reviews of historical financial information as well as other assurance and related services engagements. The standard applies to all members[2] of the International Federation of Accountants (IFAC) and requires all firms to have policies and processes in place to ensure their human resources are both technically and ethically competent (ISQC1, p. 36). The purpose of the standard is to improve auditor performance and audit quality by strengthening the ethical environment. Given the significant burden upon firms to comply with ISQC1 mandatory changes, there is need for timely research on this standard.

The only relevant study to date (Booth and Schulz, 2004) appears to provide support for ISQC1's recommendations, by finding that an overall strong ethical environment can significantly reduce the tendency of managers to continue failing investments. As Booth and Schulz (2004) is a management accounting study, the applicability of this study to the auditing field is limited. Further research on the ethical environment within an auditing context is therefore warranted.

The current study examines the impact that the presence (vis-à-vis absence) of a firm's code of ethics has on the quality of auditor judgment. The firm's code of ethics is chosen as the proxy for the ethical environment given that organisational codes of ethics have been made mandatory by ISQC1 for all accounting firms. Moreover, previous studies on codes of ethics have determined that the codes can affect auditors' ethical judgment (Jones et al., 2003; Herron and Gilbertson, 2004); however, none has extended this to consider the effect on auditors' overall audit judgment. Therefore, it is unclear how codes of ethics may affect audit judgments.

Participants (professional accountants and undergraduate auditing students) assuming the role of a senior auditor, were required to make inventory-writedown judgments. Audit quality of the participants' judgment was assessed by two measures:

The likelihood of the subject discussing the audit issue with a supervisor (“likelihood”).
The difference in the inventory adjustment amount that the subject believes to be technically appropriate under the accounting standards, and the adjustment amount that would be recommended to the supervisor (“difference”).

Participants are expected to be positively impacted by the presence of a firm's code of ethics (Douglas et al., 2001; Booth and Schulz, 2004) and therefore make higher quality audit judgments (i.e. less susceptible to clients' preferences resulting in smaller “differences”) than those not in the presence of a code. Also, based on prior research (Noreen, 1988; Siegel et al., 1995; Wotruba et al., 2001), it is expected that participants possessing greater general auditing experience (i.e. professional accountants) will also make higher quality judgments.

The interaction between general auditing experience (operationalised by auditor rank, and measured in terms of months of experience) and the presence (vis-à-vis absence) of a code of ethics was also investigated to determine whether the level of experience impacts professional accountants in a manner which leads them to react differently to the presence of the code than students.

The results provide some support for the predictions that both the presence (vis-à-vis absence) of a code of ethics, and greater general auditing experience, have a positive impact on audit judgment quality. While the presence of a code of ethics had no significant impact on the participants' awareness of the importance of the audit issue[3], it significantly enhanced the quality of the judgments of professional accountants, who recommended a smaller inventory writedown “difference” than students, and professional accountants not in the presence of the code[4]. Students were not significantly impacted by the presence of a code of ethics. This suggests that the presence (vis-à-vis absence) of a firm's code of ethics affects auditors' judgment performance only on the condition of increased general auditing experience. One aspect of greater general auditing experience includes exposure to codes of ethics in a professional work environment (Wotruba et al., 2001), and this may be a possible driver of the difference between the quality of judgments made by professional accountants and students.

Overall, the results support the standard setter's belief that the ethical environment (specifically, the code of ethics) is an important factor in improving audit quality and provide empirical support for ISQC1's requirements. The results also suggest that accounting firms may be able to improve the quality of auditors' judgment performance by timely training and exposure of their employees to organisational codes of conduct and professional ethical principles.

Theory and hypothesis development

Audit quality has been defined in a number of ways in accounting research. DeAngelo (1981, p. 186) defines 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”. Other definitions have related the quality of an audit to:

the audit's ability to reduce noise and bias and improve fineness in accounting data (Wallace, 1980);
the probability that an auditor will not issue unqualified reports for statements containing material error (Lee et al., 1999);
the accuracy of the information reported by auditors (Titman and Trueman, 1986; Beatty; 1989; Davidson and Neu, 1993); and
the degree to which the audit conforms to applicable legal and professional standards (Aldhizer et al., 1995; Tie, 1999; Krishnan and Schauer, 2001; Francis, 2004).

Given that an audit is the product of individual auditor judgments, the different definitions of audit quality, to varying degrees, all encompass the notion of auditor competency (Watkins et al., 2004). In this study, auditor competency (equivalent to auditor expertise) is defined as the degree to which an auditor can apply, and comply with, the professional standards (i.e. auditing standards and the profession's code of ethics). Thus, auditor competency is conceptualised as possessing both technical and ethical dimensions, and audit quality is determined by both technical and ethical factors.

Ethical environment (code of ethics)[5]

This study investigates how the strength of the ethical environment, as proxied by an organisation's code of ethics, impacts upon the quality of auditor judgment.

Libby and Luft (1993) suggest that the audit environment is a determinant of decision-making performance. Since, the ethical environment is a subset of the audit environment and auditors carry out their work in a value-laden environment (Dienhart, 1995), organisational factors of the ethical environment may improve auditor judgment by helping individuals determine “(i) issue(s) organisational members consider to be ethically pertinent and (ii) the criteria they use to understand weigh and resolve these issues” (Cullen et al., 1989, p. 51).

Prior research examining the ethical environment surrounding auditors is limited, with the majority of ethical auditing studies focusing on individual ethical development and reasoning, and only on the ethical dimension of auditors' judgments (Jones et al., 2003). To date, there has been only one accounting study that examined the ethical environment as a whole. Booth and Schulz (2004) conducted an experimental study investigating the effect the strength of an ethical environment has on managers' economic decisions, as well as its interaction with the level of agency problem. Booth and Schulz (2004) varied the strength of the ethical environment between strong and weak by providing information on six of the eight ethical environmental factors considered by the Independent Commission Against Corruption (ICAC) in 1998[6]. The study found that a stronger ethical environment had a direct effect on managers' project evaluation judgments, with managers under the stronger environment making decisions more aligned with the interests of their organisation under both the presence and absence of agency-problem conditions.

In relation to the specific factors of the ethical environment, studies on codes of ethics have dominated the ethical accounting and auditing literature. Codes of ethics are important since they implicitly set limits for unethical behaviour and are intended to offer guidance in ambiguous situations. Codes of ethics can perform several organisational functions, such as making explicit the ethical values that were previously unstated or unclear, alert employees as to what actions are unethical and unpunishable, and help firms shift accountability of actions from the organisation to the individual (Gellerman, 1989).

Survey studies have established that codes of ethics are thought to be relevant and important to organisations and their employees (Martinov, 2004; Lamberton et al., 2005). Lamberton et al. (2005) found that US firms that were more ethically concerned were more likely to possess a code of ethics and 90 per cent of the surveyed managers believed codes of ethics to be an important factor in resisting unethical behaviour. This echoes the earlier findings of Siegel et al. (1995), where surveyed internal auditors thought of the profession's code of ethics as a tool of ethical guidance, whose effectiveness increased as the auditors gained experience.

Brief et al. (1996) specifically examined the effect the presence of a firm's code of conduct[7] had on the frequency of fraudulent-financial reporting by executives and financial controllers and found no significant results. The subjects were given seven role plays requiring a decision of whether or not to misrepresent their firm's financial statements and found the occurrence of fraudulent reporting to be high regardless of the presence (vis-à-vis absence) of a code of ethics. The lack of positive relationship between the presence of a code of ethics and individuals' ethical behaviour is also reported by Laczniak and Inderrieden (1987) and Pater and Van Gils (2003). Laczniak and Inderrieden (1987) found the presence and endorsement of a firm's code of ethics to have little impact on the behaviour of managers (MBA students), and Pater and Van Gils (2003) reported a negative relationship between management consultants' ethical behaviour and their organisations' possession of a written code of ethics.

In contrast, studies by Adams et al. (1995) and Barnett and Vaicys (2000) report findings that the presence of a code of ethics has a positive impact on ethical behaviour. Adams et al. (1995) surveyed certified practising accountants (CPAs) as to their personal beliefs and ethical actions they would take for three scenarios relating to confidentiality. The majority of the CPAs would follow the CPA code of professional conduct when asked to resolve ethical dilemmas, even when the code might not reflect their own personal beliefs. This may be interpreted as a potential limitation of codes of ethics as they hinder the development of ethical reasoning of auditors (Jones et al., 2003), but compliance with codes of ethics should not be regarded as a negative consequence as codes are implicitly designed to create ethical behaviour which conform to a pre-determined set of ethical values. This is suggested by Barnett and Vaicys (2000), who found marketeers to be likely to engage in behaviour that is considered ethical from an organisational though not personal level when there is an organisational code of ethics.

The results of research to date on the code of ethics and its impact on the ethical dimension of auditor's judgment are mixed, with no clear indication of how a code of ethics, as part of the ethical environment, may impact auditors' judgments. The mixed findings may be due to the lack of distinction made by prior studies between the presence, and the reinforcement, of a code of ethics. The reinforcement of a code of conduct was found to be significant on the judgment of managers (Booth and Schulz, 2004), while an earlier study found no impact if the code was just presented without reinforcement (Brief et al., 1996). Furthermore, no auditing research to date has addressed the effect of the presence and reinforcement of a code of ethics on auditor's overall judgment quality, despite explicit acknowledgment that auditor judgment comprises both ethical and technical dimensions.

The aim of standard-setters is to improve audit quality through an enhanced-ethical environment. This aim, supported by the several research findings, suggests that the presence of the code of ethics will have a positive impact on the quality of judgment. This is because a code of ethics can act as an ethical framework, and help ensure individuals will behave within organisational and social expectations. Thus, it is expected that the presence of a code of ethics will have a favourable impact on the quality of auditors' judgments:

H1. The presence of a code of ethics will elicit higher quality auditor judgments.

Technical competency (general-auditing experience)

Technical competency is an individual ability and has been shown to be an important determinant of decision-making performance and judgment quality. Based on research to date, common proxies for technical competency are experience (general and task-specific) and knowledge. General-auditing experience has been found to be positively related to auditors' judgment performance when complexity of the audit task is taken into consideration (Abdolmohammadi and Wright, 1987; Libby and Tan, 1994; Shelton, 1999; Lin et al., 2003). For low-complexity tasks with well-defined routine solutions (i.e. structured tasks), auditors need to only exercise minimal judgment, and in such situations, general auditing experience is not a significant determinant of performance.

However, for more complex tasks requiring greater exercise of judgment (i.e. semi-structured and unstructured tasks), experience can improve performance by providing the necessary skills and/or knowledge required to complete these tasks (Anderson and Wright, 1988; Bonner and Lewis, 1990; Anderson and Maletta, 1994). Furthermore, task-specific experience has been shown to be able to provide additional improvement to the quality of auditors' judgments for semi-structured and unstructured tasks (Bonner and Lewis, 1990; Pincus, 1991; Libby and Tan, 1994; Wright, 2001; O'Reilly et al., 2004).

The other major determinant of technical competency is knowledge (Bonner and Lewis, 1990; Libby and Luft, 1993; Tan and Libby, 1997). Knowledge has been found to be a stronger explanatory factor of performance than experience because experience tends to impact indirectly on performance through knowledge (Libby and Luft, 1993; Libby and Tan, 1994). Task-specific knowledge is particularly helpful in semi-structured and unstructured-audit tasks (Bonner, 1990; Bonner and Lewis, 1990).

The task used in this study is a semi-structured task, for which higher quality judgments would be expected from those participants who have greater levels of general-auditing experience within a work environment. Also, it is a task that all participants would be expected to be able to complete, from a technical perspective. With this in mind, we have operationalised-technical competency as general auditing experience. Task-specific experience and task-specific knowledge have been measured to provide further insight into possible reasons for any significant differences, and to confirm the assumption that all participants are technically competent in terms of the requirements of the task. Therefore, it is posited that greater general-auditing experience will impact positively upon the quality of auditor's judgments:

H2. Greater general-auditing experience will elicit higher quality auditor judgments.

Interaction between the code of ethics and general auditing experience

While a stronger ethical environment may be created by the presence of a code of ethics, there has been no research examining the relationship between the ethical environment and technical competency (general-auditing experience).

Libby and Luft (1993) suggest that there are two types of interaction between an environmental factor and performance – an environmental factor may alter the requirements of the task (i.e. the ability, knowledge and/or effort required) and/or alter the amount or allocation of effort that decision makers are willing to employ to fulfil those requirements (i.e. changes motivation). In the first instance, a stronger ethical environment would impose additional guidelines by which the auditors must abide, so that their judgments are based on technical and ethical considerations. In the second instance, a stronger ethical environment may increase the effort auditors are willing to expend on judgment-making, due to such concerns as accountability and risk.

General-auditing experience is expected to be the dominant factor in determining auditor performance in a semi-structured task, such as the task employed in this study, because semi-structured tasks provide limited alternative choices under accounting and auditing standards. This may restrict the permissible technical and ethical options available to the auditor.

However, Noreen (1988) suggests that ethical codes of conduct will only be effective if they are bilaterally internalised by individuals as the monitoring of breaches of codes is difficult and costly. Furthermore, Siegel et al. (1995) suggest that the effectiveness of a code of ethics as a tool for ethical guidance, improves as auditors gained experience. Wotruba et al. (2001) proposes a positive relationship between content familiarity of codes of ethics and their usefulness to managers. Familiarity with good corporate governance practices (including the existence of codes of ethics) in organisational settings leads to a higher level of internalisation, whether it be consciously or unconsciously. These results are contrasted with studies pertaining to those people who do not have a high level of audit experience; namely students. Fulmer and Cargile (1987) note that while exposing students to professional codes of conduct within their accounting courses leads them to perceiving ethical issues more frequently, it has no significant impact on their actions. Furthermore, Wynd and Mager (1989) reported that students' attitudes towards ethical decisions do not significantly change after having undertaken a business ethics course. Indeed, in discussing ethics education, Armstrong et al. (2003) make the point that students' behaviour is unlikely to change simply through exposure to ethical issues, albeit a necessary part of the process.

Consistent with these findings, the following hypothesis is posited:

H3. The presence of the code of ethics will have a significantly greater favourable impact on the quality of auditor judgments for professional accountants than auditing students (i.e. in the presence of greater audit experience).

Research method

A 2 × 2 full factorial “between-subjects” design was used to test the hypotheses. Two independent variables were:

general auditing experience of the subjects (high or low); and
the presence of a code of ethics (presence or absence).

The design yielded four experimental cells.

Participants

The study had two groups of participants, auditing students and accounting professionals. The participants (120 in total)[8] were randomly allocated to experimental cells.

Auditing students were chosen as subjects, as prior studies have used this group to proxy for low-level auditing experience when studying auditors' technical competency and performance (Krogstad et al., 1984; Abdolmohammadi and Wright, 1987; Anderson and Wright, 1988; Marchant, 1989; Bonner and Lewis, 1990; Libby and Frederick, 1990; Tubbs, 1992; Anderson and Maletta, 1994; O'Reilly et al., 2004). In addition, their familiarity with codes of conduct within the work environment context is much less that that of professional accountants. Depending on the extent to which professional accountants have internalised and are familiar with the codes of conduct, the impact of the presence (vis-à-vis absence) of a code of conduct is expected to differ between the students and professional accountants. The students were in the final stages of an auditing and assurance course when the experiment was administered. The majority of students (51 of 60) had no practical auditing experience, and on average, the students had less than one month of general-auditing experience and virtually no task-specific experience. However, importantly for this study, students had the necessary task-specific knowledge to complete the task (Table I), and throughout their studies had been exposed to ethical professional requirements, including their importance and relevance in an auditing context. Recent studies suggest that exposure to ethics and ethical instruction does make students more sensitised to ethical issues, thus more readily being able to perceive ethical issues than if they have not received such instruction (Armstrong et al., 2003; Allen et al., 2005).

Professional accountants were selected to proxy for high general-auditing experience, based on prior studies employing audit tasks similar to the current study (Krogstad et al., 1984; Pincus, 1991). This group comprised both auditors and non-auditors who had been attending the final focus session in the Institute of Chartered Accountants in Australia program (CA program)[9], with a minimum of two years accounting experience. A majority of these participants were seniors, with an average of 20 months general-auditing experience, being significantly greater than that of auditing students (Table I). In addition, the professional accountants have greater inventory writedown task-specific experience than students (mean = 0.9615 vs 0.0667, respectively, based on a five-point scale measuring the number of times they had dealt with the particular type of issue being examined).

Both participant groups achieved a similar task-specific knowledge score (6.3269 for accountants, 6.1667 for students, out of a maximum of 10), indicating that both groups have sufficient technical knowledge to undertake the inventory writedown case, which strengthens the ability of the experiment to discern the impact that the code is having on judgments. All participants were volunteers and no incentives were given.

Research instrument

The research instrument is an adapted version of the “Babyboomers Inc.” case (Cohen and Trompeter, 1997) tailored by Martinov (2004) for the Australian-auditing context. The case is suitable for students and professional accountants, and is reflective of the business environment of public accounting. The audit issue in the case study deals with a potential material writedown of inventory. Prior research (e.g. Reckers and Wong-On-Wing, 1991) has identified write off of stock and stock obsolescence issues to be judgment based.

The case material consists of a background description of the client, including key-financial information[10], details of an audit issue involving potential material writedown of inventory, and two proposals for the treatment of this inventory, which had been discussed with the client. The background information also included audit materiality of $2.5 million to ensure consistent interpretation of the materiality of the potential writedown adjustment by subjects. All writedown adjustments based on the two proposals (other than the client's current treatment of nil writedown) are material.

Pilot testing

The research instrument was tested and reviewed by three senior audit practitioners and an accounting academic. All three practitioners are technical directors/consultants for their respective organisations, with two also managing audit engagements in addition to their technical role. The accounting academic was an ex-audit manager with seven years of experience in a Big-4 firm. Questions were asked to elicit responses as to the realistic nature and relevance of the background information and clarity of the instructions and the questionnaire. Two audit research academics later reviewed the revised instrument as an additional quality check.

Independent variables

There are two independent variables:

The existence of the code of ethics is manipulated by varying the presence/absence of an organisational code of conduct. The code of conduct used was based upon the Joint Code of Professional Conduct of the ICAA and CPA Australia (Joint Code of Professional Conduct), which is the official code recognised by the AUASB[11].
General auditing experience of the participants is operationalised firstly by audit rank – auditing student or professional accountant. It is also measured by the length of the auditing experience (months).

Four additional variables were measured. The first variable is the subjects' task-specific experience with inventory writedown (months); and the second, task specific knowledge (ten multiple choice questions, qualitatively similar to those employed by previous audit expertise studies such as Bonner and Lewis (1990)). These were measured in order to assess that all subjects had the requisite technical knowledge to complete the task.

The last two variables were measured as prior studies found these to be significant in affecting judgment performance. These variables include the subjects' assessment of the risk of material misstatement in relation to the given inventory account (seven-point Likert scale). Prior studies suggest that inexperienced auditors and students tend to be more risk-averse relative to experienced auditors (Anderson and Wright, 1988; Anderson and Maletta, 1994). Also, Martinov (2004) found risk assessments to explain much of the variation in auditor judgment (partners and managers) in the “Babyboomers Inc.” case utilised in this study. Finally, ethical awareness of the participants was also measured as prior studies found this variable to moderate the effect of codes of conduct on auditor behaviour (Cohen et al., 1993; Harrington, 1996).

Dependent variables

The quality of auditors' judgment is measured by two dependent variables:

The likelihood that the participant would bring the inventory writedown issue to the supervisor's attention (“Likelihood” measured on a seven-point Likert scale, ranging from 1=“Highly Unlikely” to 7=“Highly Likely”). Competent auditors should bring this issue to the attention of their supervisor because an adjustment to the inventory valuation is a significant audit issue and could have potential material impact on the financial statements. Furthermore, technically competent auditors would realise from the given financial information that the minimal adjustment required ($2.6 million) was beyond the materiality threshold ($2.5 million).
The difference between the inventory writedown amount that the participant believes to be technically appropriate under AASB (2005) 102 “Inventories” and the amount that would be recommended to the supervisor in light of client's preference for nil writedown (“Difference” ranging from “$0 million” to plus or minus “$15 million”). Expert consensus opinion suggests that no difference should be found between the “appropriate” and “recommended” adjustment amounts. Of particular concern are judgments where the “recommended” adjustment is less than what the subject believed was technically appropriate, as this would suggest reduced audit quality according to DeAngelo's (1981) definition[12].

Procedure

This experiment used an “in-basket” approach similar to that employed by Brief et al. (1996) who examined the impact a code of conduct has on fraudulent reporting in a managerial context.

Participants were randomly assigned to one of two relevant cells (presence/absence of the code) and given a package containing a set of instructions, two envelopes, case material and a questionnaire with two separate parts. The instructions asked participants to assume the role of an audit senior, who was about to go on leave. They were also told that prior to their return from leave; an audit on which they are currently working would be signed off (i.e. Babyboomers Inc.).

Participants were instructed to go through their in-tray in the short amount of time they had before their departure. The in-tray consisted of a number of documents: the “Babyboomers Inc.” case; a brief one page memorandum containing amended definitions of ethical principles (reproduced from the Code of Professional Conduct, Section B) contained in the firm's code of conduct (for participants in the “presence of code” cells); and two filler items[13] (an expenses reimbursement form and an internal memorandum discussing the firm's Christmas party). The filler items were used to create a sense of realism and to disguise the experimental manipulations (Ueker et al., 1981; Connelly et al., 2004). Subjects were not required to make an immediate response to any of the items except for the “Babyboomers Inc.” case. This provided the subjects the choice of whether to respond to the items in the in-tray, which helped to enhance the realism of the study. Schwartz (2001) found that the majority of employees in four large corporations only skimmed through their firm's code of conduct, with only a few having read the entirety of the document. Therefore, strong emphasis on the code of conduct would have been inappropriate, giving greater support for the “in-basket” approach that has been adopted.

In practice, senior auditors are often faced with time deadline pressure (Solomon and Brown, 1992; Kelley et al., 1999); therefore, to mimic this pressure and enhance realism, all participants were informed that the tasks should take no more than 15-20 min of their time. This was to create a realistic urgency for the completion of the tasks, with subjects required to balance the need for competent completion of tasks and the upcoming leave. The majority of the participants did complete the task within 15 min, with time spent by participants ranging from 10 to 20 min.

The questionnaire was organised into two parts. Part one contained the “Babyboomers Inc.” case, the code of conduct and the filler items. Following the completion of Part one, the material and questionnaire were sealed in an envelope. Part two elicited subjects' demographic information, including their general and task-specific auditing experiences. Subjects' level of knowledge of inventory valuation and their ethical awareness[14] were also measured. A manipulation check question was asked.

Results

The three hypotheses are discussed in terms of each dependent variable. The results are presented in Tables II-IV. Panel A of these tables shows the descriptive statistics including the mean, standard deviation and the number of participants. Panel B shows the results of the ANOVAs, while Panel C in Tables III and IV show the results of contrasts analyses.

Likelihood of discussing the inventory writedown

The results of ANOVA (Table II, Panel B) show a significant main effect for auditor rank (F=4.837, p=0.039), with professional accountants more likely to discuss the issue with the supervisor (mean = 5.817) than the auditing students (mean = 5.292). This provides partial support for H2. However, the presence (vis-à-vis absence) of the code of conduct had no significant effect on judgment (F=0.111, p=0.740). Therefore, H1 is not supported.

Other main (i.e. general-auditing experience) and interaction effects tested were not significant. Participants' assessment of the risk of material misstatement (F=3.176, p=0.078) was significant, with participants more likely to discuss the issue with the supervisor if they assess the risk of material misstatement to be high.

Difference between the technically “appropriate” and “recommended” inventory writedown[15]

Figure 1 shows that the majority of subjects did not reduce or increase the technically “appropriate” inventory writedown amount when they provided a recommendation to the supervisor (i.e. difference is zero). However, for the majority of participants who made adjustments the difference was negative (i.e. the “recommended” writedown was less than what they believe to be technically “appropriate”).

The ANOVAs (Table III, Panel B) report a significant main effect for the presence (vis-à-vis absence) of a code of conduct (F=6.027, p=0.016), when general-auditing experience is measured in terms of length of experience (rather than operationalised by auditor rank), with subjects in the presence of the code having a smaller difference in judgments (mean=−1.592) than those without the presence of a code (mean=−2.579). This provides partial support for H1. General-auditing experience was also significant (F=1.710, p=0.046), with a smaller difference between the two judgments for more experienced subjects, providing support for H2. General-auditing experience, operationalised as auditor rank, was not significant.

While the interaction between auditor rank and the presence/absence of the code did not show as being significant using ANOVA, descriptive statistics (Table III, Panel A) highlight that the “differences” for professional accountants in the presence of a code vary greatly from the other three cells. As discussed in Buckless and Ravenscroft (1990), contrasts analysis is a more appropriate form of analysis than ANOVA, when the results (and hence expected interactions) are in the form hypothesised (and shown) in this study. Contrasts analysis (Table III, Panel C) shows that these “differences” (mean=−0.536) are significantly smaller than the “differences” reported for the other three cells (t=−2.344; p=0.021). This supports H3. Refining the analysis in order to compare individual cells shows that the “differences” reported by professional accountants in the presence of the code, are significantly smaller than the “differences” reported by:

professional accountants in the absence of a code (F=3.735, p=0.059); and
auditing students in the presence of a code (F=6.343, p=0.015).

This is shown graphically at Figure 2.

Absolute values of the “difference” between the technically “appropriate” amount and the “recommended” judgments were also analysed. Absolute values were calculated because both positive and negative differences were found in the data, and when analysed, the positive and negative differences would have cancelled each other. ANOVAs (Table IV, Panel B) report a significant main effect for the presence (vis-à-vis absence) of a code of conduct (F=6.612, p=0.012), when general-auditing experience is measured in terms length of experience (rather than operationalised by auditor rank), thereby providing partial support for H1. Furthermore, there was a significant main effect for auditor rank (F=3.225, p=0.075) and the general auditing experience (F=1.735, p=0.042), providing strong support for H2. In general, the results suggest that general-auditing experience improves the quality of auditors' judgments, with professional accountants (mean = 1.525) reporting smaller “differences” than auditing students (mean = 2.968).

Consistent with the previous discussion on “differences” contrasts analysis (Table IV, Panel C) shows that the “absolute values of differences” reported for professional accountants in the presence of a code (mean = 0.679) are significantly smaller than for the other three cells (t=2.436; p=0.16). Specifically, they are smaller than the “differences” reported for professional accountants in the absence of a code (F=3.589, p=0.064), and those reported for auditing students in the presence of a code (F=6.183, p=0.016) (Figure 3). These results support H3.

Therefore, the presence of a code improved the judgment quality of the professional accountants, but had little impact on the students, who had similar results in both the presence and absence of the code (Tables III and IV, Figures 2 and 3).

Additional analysis

Auditors and non-auditors

Using only the data from the professional accountants, the judgments of auditors (n=24) and non-auditors were compared (n=28). For the judgment on the “likelihood” of discussion of the audit issue, there were no significant differences between the two groups. The presence (vis-à-vis absence) of a code of conduct and its interactions with general-auditing experience was not significant.

For the “difference” dependent variable, the code of conduct had a significant main effect (F=4.480, p=0.040) on judgment. The average difference in judgments was smaller for non-auditors for both treatments, but was not significant. The results of the analysis of the absolute differences are qualitatively the same. This indicates that the presence of the code of ethics has a favourable impact on the judgments of professional accountants, regardless of whether they are audit, or non-audit, professionals.

Big 4 and Non-Big 4

To test whether participants from Big 4 firms provide better quality audit judgments than those from Non-Big 4 firms, the professional accountants group was divided into the Big 4 (n=30) and Non-Big 4 (n=16). There was no significant main or interaction effect for the “likelihood” variable.

The results for the “difference” variable in the presence of the code was significantly smaller than those not in the presence of the code (F=5.612, p=0.023). A similar result is reported for the absolute value of “differences” (F=5.594, p=0.023). This is consistent with the results shown for the previous analysis of audit, and non-audit, professionals. Furthermore, there was a marginally significant interaction between the code of conduct and the accounting firm (F=2.927, p=0.094) for the absolute “difference” variable. This result appears to be driven by the much larger “differences” for Non-Big 4 participants in the absence of a code[16].

Treatment of the code of conduct

The research instrument also asked participants who received a code of conduct how they dealt with the code. There were no significant correlations (using one-tailed Pearson correlations) between the use of the code and the “difference” variable for all participants. However, there was a significant correlation (p=0.041) for professional accountants, but none for students. This suggests that the greater attention paid to the code by professional accountants may contribute to the improvement in judgment quality.

Discussion and conclusion

The aim of this study was to examine whether the presence (vs absence) of code of ethics has an impact on audit quality. This was motivated by the recent introduction of the new ISQC1, which requires all accounting firms to implement policies and processes to ensure employees' technical and ethical competence. Professional accountants and auditing students (as proxies for high and low general-auditing experience) were asked to make a number of audit judgments in relation to an inventory writedown task, with aggressive client preferences already indicated. The results of the current study suggest that the presence (vis-à-vis absence) of a code of conduct can improve the quality of auditors' judgments.

The presence (vis-à-vis absence) of a code of conduct had no significant effect on the participants' decision on the likelihood of discussing the inventory writedown with a supervisor. Rather, the quality of this auditor judgment was explained by the general-auditing experience of the subjects (auditing rank) and the subjects' assessment of risk of material misstatement. The professional accountants were significantly more aware of the seriousness of the audit issue and thus were more likely to discuss the issue. Furthermore, high-risk assessments were associated with greater likelihood of discussion.

However, the presence of a code of ethics appears to have a significant influence on the audit judgments of professional accountants. The professional accountants were sensitive to the presence (vis-à-vis absence) of a code, whereby the difference between the recommended and appropriate inventory writedown was significantly smaller for judgments made in the presence of the code. In contrast, auditing students appear to be generally unaffected by the presence of a code. Furthermore, the presence of the code appears to significantly improve the judgments made by those with greater general-auditing experience. These differences in effects may be due to differences in the degree of familiarity that professional accountants and auditing students (and also those with greater general auditing-experience, and those without such experience) have with codes of ethics (Wotruba et al., 2001), as the work environment of the professional accountants would provide them with significantly greater contact with codes of ethics than the work and study environment of the auditing students.

Overall, these results support the findings of Booth and Schulz (2004), and contradict the arguments of Dillard and Yuthas (2001) who argue that codes of ethics are only marketing tools used to legitimise organisations and cannot influence individual behaviour.

Additional analyses of the judgments of the professional accountants suggest that both auditors and non-auditors are equally affected by the presence (vis-à-vis absence) of a code of conduct. Both groups had significantly smaller differences between the technically “appropriate” and the “recommended” inventory writedown amounts when in the presence of a code of conduct. When the professional accountants are partitioned into Big 4 and Non-Big 4 groups, both groups reacted positively to the presence of a code, with those in the presence of a code having on average a smaller difference between judgment amounts than those not in the presence of a code. There was an interaction between the code of conduct and audit firm type, as Non-Big 4 accountants had on average a larger difference between the two judgment amounts than Big 4 accountants in the absence of a code of conduct[17]. This is consistent with prior research that uses Big 4 as a proxy for higher audit quality. Big 4 firms are likely to have stronger ethical environments in terms of the presence of formal codes of conduct and relevant training, resulting in increased familiarity of codes for staff within a Big 4 firm. In turn, this leads to a presumption of higher quality judgments in the absence of the code for participants from Big 4 firms, compared to those from the Non-Big 4 firms.

Overall, the results of the study provide support for the introduction of written codes of conduct at the firm level, as required by ISQC1. The fact that auditing students, unlike the professional accountants, were not affected by the presence (vis-à-vis absence) of a code of conduct suggests that the usefulness of codes of conduct depends upon users' familiarity with such codes. Therefore, accounting firms interested in improving audit quality should consider formally exposing their employees to codes of conduct through formal training, as well as audit aids such as the inclusion of checklists and/or audit program steps on all engagements. Tertiary education auditing courses, and CA and CPA programs, should also consider broadening students' exposure to codes of conduct by strengthening the emphasis on ethical principles and their significance in the auditing process.

The findings of this study must be considered in the light of the following limitations. Firstly, the use of an experimental design reduces the external validity of the study as not all information normally considered by auditors was presented in the case study. However, the internal validity of the study is strengthened by the experimental method and by use of an existing research instrument (developed by Cohen and Trompeter (1997), adapted to, and tested in, the Australian auditing context by Martinov (2004)). Secondly, the study did not control for firm-specific differences (e.g. training, decision tools and audit approach) which may influence the judgment performance of the professional accountants (Bonner, 1990). Differences in corporate cultures between firms may impact upon the judgments made by professional accountants, for as noted by Chen et al. (1997), what is perceived as being “correct” or “moral” is often defined by the environment within which decisions are made. A final limitation relates to the independent-knowledge test used to measure task-specific knowledge, which had to be created specifically for this study. However, the test is similar to those used in previous studies measuring task-specific knowledge (Bonner and Lewis, 1990; Libby and Tan, 1994; Tan and Libby, 1997), and had been pilot tested.

This appears to be the first study to date to examine the effect the presence of a code of ethics has on the quality of auditors' judgments. Given the significant findings of this study, future research into other factors of the ethical environment suggested by ISQC1 (e.g. leadership and management influence) could be explored. Another avenue of research is to extend the current study and investigate the impact of reinforcement, in addition to the presence, of a code of conduct on auditors' judgment quality. Previous studies have suggested that reinforcement of codes of conduct by firms may produce additional significant effects (Laczniak and Inderrieden, 1987; Brief et al., 1996). Lastly, similar studies could be undertaken for a wider range of auditing experience, such as audit managers and partners.

Figure 1Frequency of “differences” between inventory writedown believed to be technically appropriate and recommended

Figure 2Means of “differences” between inventory writedown believed to be technically appropriate and amount of writedown recommended

Figure 3Means of absolute value of “differences” between inventory writedown believed to be technically appropriate and amount of writedown recommended

Table IParticipants. Panel A: descriptive statistics for participants' technical competency proxies

Table II“Likelihood” of discussing the issue (what is the likelihood of the auditor discussing this issue with the supervisor before going on leave? 0, highly unlikely; 7, highly likely)

Table III“Difference” between inventory writedown believed to be technically appropriate, and amount of writedown recommended (recommended inventory writedown less technically appropriate inventory writedown 0, no difference; 15, increase in suggested amount; −15, decrease in suggested amount)

Table IVAbsolute value of the “difference” between inventory writedown believed to be technically appropriate and amount of writedown recommended (Recommended inventory writedown less technically appropriate inventory writedown 0, no difference; 15, difference in suggested amount)

References

Abdolmohammadi, M., Wright, A. (1987), "An examination of the effects of experience and task complexity on audit judgments", The Accounting Review, Vol. 62 No.1, pp.1-13.