How to mitigate auditors� conscious and unconscious bias by 2va3rBK

VIEWS: 21 PAGES: 37

									      How to mitigate auditor’s conscious and unconscious bias?
                       An experimental study

                                                     Abstract

Increasing evidence of auditors’ failure to provide an independent opinion on true and fair view of financial
position of firms during the financial crisis has reopened the debates about effective measures to ensure
auditor’s independence. The purpose of this paper is to examine two prominent determinants of auditor’s biased
opinion – financial incentives and personal relationship with a client that cause conscious and unconscious
auditor’s bias. We hypothesise that contingent financial incentives elicit predominantly conscious bias, whereas
personal relationship induces mainly unconscious bias. We conjecture that oversight, intended to mitigate
biases, is effective only for conscious bias. To analyze the hypotheses we conduct a two period two-by-two
between and with-in subject experiment with 176 students involving a choice task. We find that financial
incentives have a significant effect on an auditor’s choice. While personal relationship appears to have no such
overall effect, we find it significantly related to unconscious bias. Our results show that oversight can effectively
mitigate auditor’s bias, because most of it is conscious. Unconscious bias is persistent throughout both rounds of
the experiment and is unaffected by oversight. The results suggest that while oversight may effectively mitigate
conscious bias it does not affect the unconscious one. The findings contribute to the recent regulatory
discussions on measures to increase auditors’ independence. They show that in addition to oversight alternative
solutions are needed to mitigate the unconscious bias. Audit firm rotation may prevent the effects of
unconscious bias by terminating the long-term auditor-client relationship.



Key words: auditing, financial incentives, personal relationship, conscious bias, unconscious

bias, oversight, rotation.




         1. Introduction

         At the wake of the global financial crisis the audit profession has failed to fulfill the

principle purpose of the audit: to enhance the degree of confidence of intended users in the

financial statements. Sikka (2009) reports that many distressed financial institutions in

different countries received unqualified audit opinions on their financial statements

immediately prior to the public declaration of financial difficulties. This has drawn attention

to the age-old debates related to auditor independence and audit oversight. Impaired auditor




                                                         1
independence is generated by inter alia economic dependence and long tenure (Dart, 2011).

Both factors were abundant at the audit market in the aftermath of the financial crisis.

       Recently, accounting profession has undergone a profound regulatory reform in the

U.S. (Sarbanes-Oxley Act of 2002), the EU (Directive 2006/43/EC) and elsewhere. One of

the most radical measures of the last reform was the introduction of public oversight,

established to control the audit quality beyond professional self-regulation. As auditor

independence is central to the integrity of the audit process (Chen, Elder and Liu, 2005), it

was hoped that an efficient public oversight would mitigate the negative effects of impaired

auditor independence. Evidence supports the conclusions that the quality of auditing and the

quality of financial reporting have improved after the passing of the Sarbanes-Oxley Act

(Cohen, Dey and Lys, 2008) and that auditors appear to be more conservative (Lobo and

Zhou, 2006). In spite of these developments, continued audit failures suggest that the newly

implemented oversight mechanism failed to mitigate a vital element of auditor’s conflict of

interest. In addition, drivers of the audit failure in the financial crisis remain the same as those

of the large financial scandals at the turn of the century. This leads to the conjecture that the

regulatory changes in the auditing profession did not effectively succeed to enforce auditor

independence.

       Some leading scholars believe that to restore the integrity of the auditing function,

“auditing firms (not just lead auditors) should work on a contract for fixed number of years

that cannot be terminated by the client or renewed” (Bazerman and Moore, 2011, p. 310).

This issue was revisited by the European Commission Green Paper, where the Commission

recognizes that “situations where a company has appointed the same audit firm for decades

seem incompatible with desirable standards of independence” (Green Paper “Audit policy:

Lessons from the crisis”, 2010, p. 11). The new proposal draft of the European Commission




                                                 2
for a Regulation on specific requirements regarding statutory audit of public interest

companies is for the first time about to introduce mandatory audit firm rotation.

       The reason why audit firm rotation debates have re-gained relevance lies in the

essence of the auditor’s independence threat. Extant research points out that auditors are

inclined to serve the client’s interest (Hackenbrack and Nelson, 1996; Kadous, Kennedy and

Peecher, 2003; Blay, 2005; Kadous, Magro and Spilker, 2008; Moore, Tanlu and Bazerman,

2010) if accounting choices are ambiguous. As argued by Callao and Jarne (2010) the scope

for ambiguity has rather increased with the adoption of IFRS in Europe. The tendency to

serve client’s preferences decreases with increasing risks for an auditor associated with

incurring high explicit or implicit costs (such as loss of reputation, litigation costs or license

withdrawal). On one hand, this tendency arises from implicit auditor’s dependency on a

client’s service satisfaction, which ensures prolongation of a contract and guarantees future

rents. On the other hand, biased judgment may arise also from interpersonal auditor-client

relationship, which can make an auditor hesitant to act with professional rigor not to impair a

relationship with the client (Johnstone, Sutton and Warfield, 2001).

       Both types of incentives create so-called directional goals that elicit motivated

reasoning (Kunda, 1990). The theory of motivated reasoning predicts that individuals with

directional goals process information in a biased manner, achieving seemingly objective

support for the desired goal (Kunda, 1990, Blay, 2005). Through motivated reasoning people

adopt self-serving attributions in information processing that permit them to reach the

conclusion they want to reach (Kunda, 1990). One of the controversies of motivated

reasoning lies in the neural processes underlying it – it seems to be unresolved whether it

occurs through cold cognition (i.e. analytical, conscious) or through hot cognition (i.e.

intertwined with emotion, unconscious). This is a pertinent question in the context of

auditor’s independence threat as unconscious biases may be persistent to measures intended



                                                3
to mitigate them. While Kunda (1990) and Nelson (2005) propose that people are not aware

that their information processing is biased by their goals, empirical evidence in the

accounting settings suggests that professionals are sensitive to high practice risks (Blay,

2005; Kadous et al., 2008). Such sensitivity suggests that this bias is not unconscious.

       As the reasons for the bias are not explicitly addressed in the literature, the studies do

not distinguish which of the two conditions – financial incentives or personal relationship

predominantly accounts for which type of bias and do not tackle the question whether these

effects are intertwined. Literature indicates that financial incentives are closely related to

auditor independence (DeAngelo, 1981). The effect of financial incentives on auditor’s

choice was examined by Farmer, Rittenberg and Trompeter (1987), Lord (1992), Blay (2005)

and Moore et al., (2010). The empirical evidence that financial incentives induce auditors to

support client-preferred choices is relatively consistent.

       Although the bias arising from personal relationship has also been investigated in

psychology (Milgram, 1974; Neuberg and Fiske, 1978, Kunda, 1990), corporate governance

(Morck, 2008) and auditing (Bamber and Iyer, 2007; Moore et al., 2010), it is much less

understood. To our best knowledge, the only study that has simultaneously analysed the

effect of financial incentives and personal relationship on auditor’s choice is the Moore et al.

(2010) study. Moore et al. (2010) argue that personal relationship elicits unconscious bias.

Although the results of their experimental study do not confirm the hypothesis that in the

presence of financial incentives personal relationship affects biased auditor’s opinion in

public reports, they show that auditors are susceptible to their role to serve the client and that

they can not entirely correct for the bias even if in a different role – for a private purpose.

       Our study extends the Moore et al. (2010) study by re-examining the effect of bias

under direct (future rents) and indirect (personal relationship) incentives by strengthening the

measurement of personal relationship. We measure unconscious bias by comparing an



                                                 4
individual’s decision in the role of an auditor and that of an expert (c.f. Moore et al., 2010)

and we analyse which incentive accounts for unconscious bias. The limitation of the Moore et

al. (2010) experiment is that the participants did not face a threat of punishment. To examine

how subjects react to such a threat, we introduce the oversight risk and compare choices in

the model with and without the oversight risk. Assuming that individuals are not aware of

unconscious bias, oversight should mitigate only their conscious bias, while unconscious bias

should remain unaffected.

       Our baseline model results confirm prior findings that contingent fees (i.e. expectation

of future business with a client) are significantly associated with the auditor’s decision. We,

however, also don’t find a direct support for the hypothesis that personal relationship affects

the auditor’s decision. In a further investigation, in which we analyze whether the participants

were aware of the bias, we break up the observations of the bias to conscious and

unconscious one. We find that conscious bias prevails. Furthermore, we observe that personal

relationship is the only significant predictor of unconscious bias. As its frequency in our

experiment is very low, this may explain why personal relationship could not be detected as

significant in a more robust test of auditor’s decision.

       Our analysis provides the following contributions to the literature. Our paper adds to

the theory of motivated reasoning and the relatively scarce empirical evidence of bias that

arises in personal relationships, after controlling for financial incentives. In particular, we

examine the question whether bias is intentional (conscious) or unconscious. We measure the

latent phenomenon by the difference in decision upon changing the role of a subject and

introducing of oversight. The finding that oversight reduces only part of the bias contributes

to the regulatory discussion regarding how to mitigate auditor’s biases more efficiently.

       The paper is further structured as follows. Section two provides theoretical

background and develops the hypotheses. Section three focuses on the design of the



                                                5
experiment. Section four provides the results of the analyses. Section five concludes with the

discussion of the results, possible implications for audit regulation and limitations of the

study.



         2. Hypotheses development

         Auditor independence is considered a vital determinant of audit quality. Lee and Gu

(1998, p. 534) define it as “the absence of collusion between the auditor and the manager of

the client firm”. Different threats to auditor independence and their impact on earnings

management have been thoroughly investigated in the accounting and auditing literature.

While agency theory stresses intentional distortions, behavioural literature points out to

auditor bias due to cognitive limits (Blay, 2005; Kadous et al., 2008). The neural processes

underlying auditor’s bias are hard to be discerned as they lead to same consequences.

         Kadous et al. (2003) explain the mechanisms of auditor’s decision-making with the

theory of motivated reasoning. According to this theory, when a person expects that his or her

outcome depends in some way on another person, such outcome dependency creates a

directional goal. A directional goal, in turn, influences the cognitive process of reasoning by

affecting which information will be considered, how it will be evaluated and interpreted. The

theory of motivated reasoning stems from psychology (Kunda, 1990). As the author of the

theory explains, the confusing fact is that a person is able to provide evidence to support his

or her biased decision, without realizing the bias: “The objectivity of this justification

construction process is illusory because people do not realize that the process is biased by

their goals, ... and that they might even be capable of justifying opposite conclusions on

different occasions.” (Kunda, 1990, p. 486).

         Related to the auditing setting, the more the auditors aspire to benefit from their

support to client’s preferences, the more likely they will find sufficient evidence and interpret



                                                6
it in such a way that it agrees with client’s choices. A necessary condition for motivated

reasoning is the ambiguity of the choice. If the preferred choice cannot be seemingly

objectively justified, individuals will not reach it, regardless of their commitment to a

directional goal (Kunda 1990). This theory has been supported by convincing empirical

evidence (Kadous et al., 2003; Blay, 2005; Kadous et al., 2008). In a similar vein, Cloyd and

Spilker (1999) analyse confirmation bias, i.e. biased information search by tax professionals

to support the client’s desired outcome.

       However, prior experimental studies unanimously show that auditors and tax-advisors

while serving to their client are sensitive to the variation of risks that might threaten their own

interest (Farmer, Rittenberg & Trompeter 1987; Lord, 1992; Hackenbrack and Nelson, 1996;

Cloyd and Spilker, 1999; Kadous et al., 2003; Blay, 2005; Kadous et al., 2008). This blurs

the distinction of whether this cognitive process is deliberate or unconscious. Although

motivated reasoning research seems to minimize the importance of this division, the rationale

why we believe it is important is because different biases may only be efficiently tackled with

different regulatory measures.

       In auditing the major two conditions that create directional goals are contingent

financial incentives and personal relationship with the client (Nelson, 2005, 2006; Blay,

2005; Moore et al., 2010).

       Threat to auditor’s independence, stemming from financial incentives, relates to the

possibility of the loss of the client and related future quasi-rents in case of issuing a qualified

audit opinion (Krishnan and Krishnan, 1996). Management’s ability to influence auditor

reporting decisions is increased if the incumbent auditor perceives the client as the source of

perpetual annuity (Ruiz-Barbadillo, Gomez-Aguilar and Carrera, 2009). To guarantee the

independence and disconnect the auditors from the interests of the management, auditors

receive a fixed fee for their services. Nevertheless, implicitly their financial incentive may be



                                                7
variable: the nature of the auditor-client relation can create a variable fee structure in a sense

that “opposition” to managements’ representation of financial statements can lead to the loss

of future business.

       In other words, the better the performance of the auditor matches the preferences of

the client, the more likely the stipulation of the contract in the future years. Moreover, the

longer the audit tenure, the higher the labour productivity as the auditor is increasingly

familiar with the client’s business and needs to perform fewer check-ups. Johnstone et al.

(2001) suggest that contingent fees (i.e. implicit promise of the future rents dependent on the

auditor’s choice) directly threat auditor’s independence. Hence, we propose to test the

following hypothesis:



       H1: Financial incentives increase the probability that the auditor supports the client’s

preference.



       Alternative and/or complementary venue that affects biased decision-making is the

nature of the relation between the auditor and the client. Existing literature suggests different

drivers of personal relationship on decision choice. Threat to auditor’s independence, related

to personal relationship with the client has been accentuated already by Mautz and Sharaf

(1961) suggesting that long audit tenure can have a detrimental effect on auditor’s

independence because auditor’s objectivity regarding a client is reduced with the passage of




                                                8
time1. Similarly, Rennie, Kopp and Lemon (2010) indicate that the endurance of the auditor –

client relationship results in closeness between auditors and their clients. Drawing from

Stickel and Gamroth (2006) who report that individuals tend not to be aware of the drivers of

their trust of another individual until they are in a circumstance that tests their trust, Rennie et

al. (2010) infer that in auditing a disagreement between an auditor and a member of a client

management represents such a circumstance. Using actual auditor – client disagreement

situations, they report that the length of the auditor – client relationship is positively related to

the auditor’s trust of a client representative.

        Other studies reveal that closeness between auditors and their clients compromises

independence (Moore, Tetlock, Tanlu and Bazerman, 2006) and creates stronger auditor’s

identification with the client (Bamber and Iyer, 2007). Bamber and Iyer (2007) additionally

report that auditors who identify more with a client, are more likely to consent to the client-

preferred position. Moreover, in a review of existing research of professional scepticism in

auditing, Nelson (2009) considers identification with a client to be similar to low professional

scepticism. Likewise, Johnstone et al. (2001) suggest that interpersonal relationships might

affect the auditor’s ability to exercise a proper level of professional scepticism.




1
  Deis and Giroux (1992) report a negative relation between auditor tenure and audit quality in the public sector.
On the other hand, several studies provide evidence for a positive relation between audit firm tenure and audit
quality (Myers, Myers and Omer, 2003; Johnson, Khurana and Reynolds, 2002; Jackson, Moldrich and
Roebuck, 2008; Geiger and Raghunandan, 2002; Carcello and Nagy, 2004; Ruiz-Barbadillo et al., 2009;
Cameran, Prencipe and Trombetta, 2009). The prevailing argument behind this finding is the learning process
resulting from longer audit tenure. However, these studies have not addressed auditor’s biases.



                                                        9
          In the paradigm of motivated reasoning the belief that one’s outcome depends on

another person creates a directional goal, which may bias trait and ability evaluations of the

target person (Kunda, 1990). People tend to see others as more likable if they expect to

interact with them. Neuberg and Fiske (1987) report that outcome dependency enhances

liking of a target person. Liking somebody introduces emotions into decision-making process

that is no longer based purely on rational maximisation of utility which suggests that personal

relationship may induce bias beyond financial incentives being expected from the target

person.

          An alternative explanation of the influence of personal relationship on choice can be

drawn from the field of social psychology, the work of Milgram (1974) in particular. On the

basis of a psychological experiment, the author reports that humans have an instinctive

predisposition for loyalty, an impulse to obey authority. Building on the work of Milgram

(1974), Morck (2008) points out to the predisposition of individuals towards excessive

loyalty (obedience) to the principal in the field of corporate governance. Excessive loyalty is

dependent on the nature (strength) of the personal relationship between the agent and the

principal. Applied to the lasting relationship between independent directors and executive

managers, Morck (2008) spotlights the phenomenon of agentic shift according to which

independent directors become excessively subservient to executive managers due to different

effective distribution of power than formally defined. Although this reasoning relates to

lasting relationship between independent directors and executive managers, parallels can be

drawn to the relation between the auditor (agent) and executive manager (who may be seen as

a principal).

          Overall, we propose to test the following hypothesis:




                                                10
       H2: Personal relationship increases the probability that the auditor supports the

client’s preference.



       The two venues inducing auditor’s bias, however, are essentially different. To be able

to provide some suggestions regarding the policies that affect auditors’ independent opinion,

it is important to distinguish between the two. According to the agency theory (Gavious,

2007), Moore et al., (2010) and as implied from the literature of motivated reasoning in

auditing (Blay, 2005; Kadous et al., 2008), we propose that financial incentives create

predominantly conscious bias of the auditor. On the other hand, we view personal

relationship as a mechanism that elicits primarily unconscious bias of the auditor. As it stems

from the literature investigating the antecedents and consequences of interpersonal

relationship, developing affection (or liking, or subservience) for a target person may be

rather unconscious. To test this conjecture, we propose the following hypothesis:



       H3: Personal relationship elicits unconscious bias in the auditor’s decision-making.



       A further cue of the neural underpinning of bias may be obtained by varying auditor’s

practice risk, by the introduction of a measure such as oversight. The oversight may bring

about litigation costs, licence withdrawal and negative reputation effects. If financial

incentives were the only driver of bias, the independence could be assured by regulating the

structure of fees, the maximum length of the audit tenure and by the introduction of

oversight. Unconscious bias may, however, be more persistent. Milgram (1974) has

demonstrated that excessive loyalty in personal relationship can be substantially reduced by

the introduction of a so-called “dissenting peer”. More specifically, having the auditor action

being challenged or reviewed by a different auditor would create such an effect and could



                                              11
reduce the impact of bias on the auditor’s choice of action. If auditor’s decisions were

affected by unconscious bias, we argue that independence could not be completely assured by

the introduction of oversight. Instead, mandatory auditor rotation could be more effective as

it would break personal ties and eliminate the bias. If, on the other hand, the oversight is

effective in personal relations, this would imply that also personal relationship may to a large

extent induce conscious bias and that the auditors perceiving high practice risk adapt to it. In

sum, we expect that the oversight as an ex-post verification of the audit performance has a

significant mitigating effect on conscious bias arising from financial incentives as it threatens

the auditor’s directional goal. Moreover, we expect that the oversight has a less strong effect

in personal relationship because the auditor is not aware of the bias. We propose to test the

following two hypotheses:



       H4: Oversight has a stronger effect on the auditor’s choice in a contingent financial

incentive scheme as compared to a scheme with a fixed fee.



        H5: Oversight has a weaker effect on the auditor’s choice in a personal auditor-

client relation as compared to an anonymous relation.



      3. Experimental setting

      Participants. We analysed the hypotheses experimentally, with a two-by-two within

and repeated-subjects design that involved a choice task. We recruited 182 students of the

University of Ljubljana, the Faculty of Economics. The participation in the experiment was

voluntary. To motivate them, they could earn compensation in the amount between 3 to 9

EUR. On average, their compensation amounted to 4.4 EUR for one hour of activity, which

roughly represents the average hourly rate of student work. The amount paid ensured that



                                               12
students considered it high enough to be attractive. We recruited undergraduate (46%) and

graduate (54%) students of accounting and finance to be more acquainted with the decision

problem. Their average age is 22.1 years, 70% of them are female, their average work

experience (including part-time student work) is 3.5 years. Subjects assumed the roles of

auditors (118 students) or clients, i.e. Chief Financial Officers (64 students). Our analysis was

focused on decision-making of auditors. Subjects in the role of clients were used to create the

atmosphere of auditor-client familiarity.

      Procedure. Auditors were either seated alone (in the condition without the client) or

matched in pairs (i.e. seated with the client). In the materials the auditors were presented a

task of approving a valuation of an investment proposed by clients, or choosing the

alternative, not in the interest of the client. Clients were presented the same scenario with the

task to persuade the agents to approve the valuation in their interest. Two alternative values

of the investment were both measured with a valuation model. In the first model a more

realistic assumption about future cash flow growth was used which produced a lower value of

the investment. In the second model valuation was based on a rather aggressive assumption

about future growth, in favour of the client. In the first five minutes of the experiment, agent-

principle pairs discussed their personal matters to get to know each other better. Then they

read the case and their task. In the next 5-10 minutes they were discussing the valuation.

After the discussions the auditors indicated their decision in the provided form. Not only was

their decision observed by the client, but the client’s reward was calculated on its basis as

well. In the subgroup without the client the auditor took their decisions without any

interference with others.

       After the first round of the experimental task, we proceeded with the second round. A

new package of instructions was handed to the participants. It contained changed

circumstances, mentioning the possibility of oversight and an altered regime of financial



                                               13
incentives. Another discussion took place between the auditors and the clients in the

condition where they were matched. The auditors with unknown clients took a new decision

solely on the basis of the new piece of information.

       Finally, the participants completed an exit questionnaire with manipulation and

understandability checks and demographic questions. In addition, we measured a control

variable for subjective norms, and the extent to which subjects believed the correctness of the

decision they took in the role of auditors. An important determinant of an individual’s

behaviour is subjective norm. Subjective norm is an individual’s perception of the opinion of

others about the behaviour in question (Carpenter and Reimers, 2005). It refers to the

perceived social pressure to perform or not to perform certain behaviour (Ajzen, 1991).

According to the theory of planned behaviour the more an individual perceives that important

others think she (he) should engage in a behaviour, the more likely she (he) is going to do so

(Ajzen, 1991). Normative beliefs constitute the underlying determinants of subjective norms.

Carpenter and Reimers (2005) suggest that subjective norms of an agent may be, for example,

affected by the opinion of a respected professor by »making disapproval of violating

accounting standards as a part of the students’ education« (p. 119). To control for subjective

norms we measured whether participants have concluded the course of Auditing. Namely, in

the course of Auditing, conducted by one of the authors of this study, a strong emphasis is

given to the ethics of auditors with an intention to influence normative beliefs of students.

       To measure bias in decision-making we designed a question in which the task was

repeated, but the role of the participants changed. According to Lord, Lepper and Preston

(1984) subjects cannot undo their bias even if they are told to be objective, but only if they

are asked to consider another perspective. With the change in decision in different roles we

assessed the degree to which the subjects believed that their decision was affected by the role

of auditors (see the questionnaire in the Appendix).



                                               14
       We excluded six observations (agents) because of the incomprehension of the task

checked with the exit questions. The final number of subjects in the role of agents was 112.

      Design. The experiment had a two (personal relationship: anonymous vs. personal) by

two (financial incentive: fixed fee vs. future business) between-participants factorial design.

The task was repeated in two subsequent rounds – without and with auditors’ oversight –

which created a repeated-measures design. Participants were randomly assigned to four

groups. We manipulated the experimental conditions as follows: In the condition of

anonymous principal, auditors received instructions in a written form and submitted their

decision in writing to the experimenter. In the condition of personal relationship with

principal, we matched auditors and clients. We asked the students to create pairs with

colleagues that they know best in class. We believe that this way we improved the

manipulation of the relationship condition compared to the study of Moore et al. (2010)

where auditors and clients spent only a few minutes to get to know each other, before they

started working together.

       In the fixed fee condition, the auditors were paid a fixed fee of 2 EUR regardless of

their decision. In the future business condition the agents received a 2 EUR of fixed fee if

they supported the valuation, which was not in the interest of the client. Conversely, they

could earn an additional 1 EUR as a present value of future business with the company if they

supported the valuation that the client preferred. In total, for this decision they received 3

EUR. Clients received 2 EUR of compensation if the auditor disagreed with their valuation

and 3 EUR as a result of higher bonus if the auditor conceded to the valuation based on the

aggressive assumption of growth.

      Client’s compensation scheme was designed to correspond to the auditor’s

compensation scheme, but also to make them eager to convince agents to support the

valuation in their interest. With the introduction of oversight the financial incentives changed.



                                               15
Oversight was manipulated in the following way: in the first round subjects were assured that

no oversight was possible. In the second round they were told that the Law on Auditing has

changed and that it has become possible that their audit is going to be subject to regulatory

oversight. The probability of the oversight is 50%. If they approve the lower value of the

investment, it would be highly unlikely that oversight authority opposes their opinion. If, on

the other hand, they approve the higher value of the investment, it is certain that the oversight

authority would disagree with their decision.

      The financial incentives in the oversight condition have changed only for the future

business condition in which the future business bonus increased from 1 EUR to 4 EUR for

approving the higher value of the investment. In total, the auditor could earn 6 EUR for this

decision. Under both incentive schemes - fixed fee and future business, the compensation

would be zero if the oversight authority disapproves the auditor’s decision. In the future

business condition the expected value of the total compensation in the oversight condition

was kept constant to the expected value of the compensation in the no-oversight condition.

The client in this round received 2 EUR if the auditor insisted on the lower value of the

investment and 4 EUR if the auditor supported the client’s valuation. Yet, holding constant

expected value of incentives prior to oversight and with the oversight was not feasible for the

fixed fee condition. In the fixed fee condition there is a trade-off in the experimental design

between oversight having a financial effect and keeping constant the expected value of

financial incentives. We decided that the former option is more realistic.

      As soon as the agent made a choice in the second round, a random number drawn

by the computer mimicked whether oversight took place (by taking up the values of 0 or 1).

Oversight authority opposed to the auditor’s opinion if it was congruent with the client’s

preference. This determined the value of auditor’s compensation for the decision taken.




                                                16
Clients were paid on basis of valuation outcome, but not penalized if oversight disapproved

with the auditor’s decision.



       4. Results

       To test the hypotheses 1 and 2 we first performed a two-way ANCOVA test (two

independent variables and a covariate). In the hypothesis 1 and 2 we investigate whether

financial incentive and personal relationship with the principal affect the decision taken by

the agent in the first round. Descriptive statistics show that only 33% (Std. Error = 0.062) of

subject within fixed fee condition decided to approve valuation in favour of the principal,

whereas in future business condition 52% (Std. Error = 0.066) of them decided to do so.

Furthermore, 38% (Std. Error = 0.077) of the subjects in the condition of anonymous

principal accepted the decision in favour of the principal, whereas in the condition of

personal relationship 47% (Std. Error = 0.075) of them accepted such a decision. The results

of ANCOVA in Table 1 confirm a significant effect of financial incentive on the auditor’s

choice (F=4.363, p=0.039). Despite the sign of personal relationship is such as expected, the

effect has not been found significant. Subjective norms of the subjects (proxied by the course

on Auditing) also had a significant effect on the choice (F=5.92, p=0.017).



                       <INSERT TABLE 1 AND FIGURE 1 ABOUT HERE>



       Table 2 presents the results of between-subject effect – i.e. variables that explain the

difference in the decision across subjects in the two rounds. In confirmation of hypothesis 1

based on the findings of the first round, financial incentives exhibit a significant effect

(F=6.134, p=0.015; r=0.690). Similar as before, no significant effect is found for the personal

relationship (hypothesis 2), while subjective norms (Auditing course) exhibit significant



                                              17
effect (F=10.406, p= 0.002, r=0.892). This means that financial incentives and subjective

norms persistently explain the auditor’s decision in both rounds with a large observed power.



                                 <INSERT TABLE 2 ABOUT HERE>



       Next, we investigated the role and determinants of conscious and unconscious bias in

the decision-making. We compared the decisions of the subjects taken in the role of auditors

and in the role of a valuation expert, consulting a different client – an investor who wants to

purchase the firm being valuated. By comparing the decisions in two different roles we were

able to measure the magnitude of the bias on an ordinal scale.



                                 <INSERT TABLE 3 ABOUT HERE>



       The results in Table 3 reveal that in the first round 53.6% of subjects made conscious

bias by deciding for the higher value of an investment in the role of experts after the second

round of the experiment. In the second round in which we introduced oversight, the percent

of the subjects that made a conscious bias decision declined to 42.8%. Within biased decision

we also observe the reduction in the bias magnitude from the first to the second round.

Among the subjects that did not change their decision, there were eight subjects who

persistently decided for the value of the investment in favour of the client which they served

as auditors. We consider these answers as an indication of unconscious bias. This analysis

shows that bias in our sample is predominantly conscious as only 7.1% of the subjects made

biased decision unconsciously.

       We test hypothesis 3 by analysing determinants of the unconscious bias. In Table 4

we report the results of logistic regression in which we regress financial incentives, personal



                                              18
relationship and subjective norms (Auditing course) on bias, defined as conscious and

unconscious (0 and 1, respectively).



                                <INSERT TABLE 4 ABOUT HERE>



       In line with our expectations and hypothesis 3, we find that personal relationship

significantly predicts the probability for the occurrence of unconscious bias (b=2.036, Wald

statistics=4.703, p=0.03), whereas the effect of financial incentives is not significant. In other

words, unconscious bias is more likely to occur in the personal relationship. Logistic

regression correctly classifies 88.2% of the subjects. The results are robust for bias in both

rounds. We also regressed subjects’ answers on the exit question what impacted their

decision on unconscious bias and the results show that neither self-reported financial

incentive nor personal relationship significantly explain their decision. This suggests that

subjects are indeed not aware of the unconscious bias.

       Finally, we analyse the effect of oversight. Hypotheses 4 and 5 predict the effect of

the oversight on the decisions taken by the auditors in different conditions. We performed

repeated-measures two-way ANCOVA in which the treatment variable was oversight and the

dependent variable was the agent’s decision in the first and the second round. With the

analysis we test whether oversight has significantly affected the change in the decision in two

rounds, and the interaction effects between oversight, financial incentives and relation

(hypotheses 4 and 5, respectively).

       In Figures 2 and 3 we represent the effect of oversight under different financial

incentives and personal relation conditions. In Figure 2 we observe that the introduction of

the oversight induces subjects to be more conservative in their decisions, since larger

proportion of subjects chose the lower valuation option. H4, is however, not confirmed: the



                                               19
effect of oversight did not differ between auditors that were exposed to the financial

incentives and those that were not. In Figure 3 we similarly observe the positive effect of the

oversight in personal relationship conditions on the decision in the second round – overall, in

both conditions more auditors chose the lower valuation options. The figure suggests that the

effect of oversight on the decision is more pronounced in the condition of anonymous

relation, which might imply that the personal relation mitigates the positive effect of the

oversight on bias.



                           <INSERT FIGURES 2 AND 3 ABOUT HERE>



                                 <INSERT TABLE 5 ABOUT HERE>



       The results in Table 5 confirm a significant main effect of oversight (F=5.891,

p=0.017), which is relatively large (r = 0.672). More importantly, the effects of oversight on

the decision when auditor’s decision was influenced either by the financial incentive

(hypothesis 4) or by the personal relation between auditor and the client (hypothesis 5) are

not found significant. Moreover, the interaction among oversight, financial incentive and

personal relationship is significant at 14.1% level. Given the complexity of a test, a relatively

small number of observations may be responsible for inconclusive results.

       Overall, these results confirm that financial incentives impact subjects’ decisions and

that oversight can effectively unravel this effect. These results are intuitively logic and in line

with the results reported by Moore et al. (2010) who report the effect of financial incentives,

and also do not find personal relationship significant. Overall, the results confirm mitigating

effect of the oversight on bias, but do not confirm that its effect differs across conditions as

proposed in H4 and H5.



                                                20
       5. Discussion and policy implications

       Following numerous audit failures during current financial crisis, the aim of this study

is to contribute empirical evidence to the theoretical and regulatory debates how to contest

auditor’s conflict of interest. In particular, the study is intended to contribute to the debates

whether oversight can effectively mitigate auditor’s biases and whether additional measures

such as mandatory audit firm rotation are necessary. The need to question the effectiveness of

the measures arises from the new proposal draft of the European Commission on a regulation

of statutory auditors, which is for the first time about to introduce mandatory audit firm

rotation.

       To anaylse these questions we attempted to measure conscious and unconscious bias

with the conjecture that they may be induced by two different drivers. Regarding financial

incentives, our results add to the prior evidence that financial incentives adversely influence

auditor’s independence. Regarding personal relationship, despite the attempt to manipulate

this condition more efficiently than Moore et al. (2010), we also could not confirm that this

condition would be directly responsible for the auditor’s decision, although the sign of the

effect is in line with our expectation. The subjects that were in the personal relationship

condition were more inclined to submit to the volition of the client, yet the result is not

significant. Nevertheless, our more detailed analysis suggests that personal relationship is not

without importance.

       Concerning the type of bias, the results show that conscious bias in our experimental

setting largely prevails. Unconscious bias is found to be significantly explained by a personal

relationship. This is in line with our predictions based on extant research (Neuberg and Fiske,

1987; Stickel and Gamroth, 2006; Rennie et al., 2010). Moreover, we find that the oversight

decreases the overall bias and its magnitude, but that it does not affect unconscious bias. As



                                               21
unconscious bias is more likely to occur in personal relationship, represented inter alia by a

long tenure with the client and as this bias is neither perceived by individuals nor can it be

mitigated by the oversight, the results of our study suggest that only a termination of the

relationship would help eliminating auditor’s allegiance. Mandatory auditor rotation could

therefore be considered an important alternative tool for mitigation of auditor’s unconscious

bias. As biases arising from personal relationship are relatively less frequent, audit partner

rotation solves only for a less pronounced part of the problem. It does not unravel the

problem of financial incentives, as partners’ incentives jointly depend on overall firm

performance, rather than on an individual audit job. While audit partner rotation would most

likely solve biases arising from personal relationship, it is only audit firm rotation that could

completely eliminate financial incentives of partners, associated in the same audit firm.

       Public oversight that was introduced following the recent auditing regulatory reform

to scrutinize audit quality beyond professional self-regulation can be regarded as an effective

measure to increase audit quality. The only question is how far reaching it is, as the public

oversight authority cannot examine the entire market frequently enough. On the other hand,

peer oversight – the crucial element of audit firm rotation – can extend to all market

participants and could represent an effective complementary mechanism for mitigating

auditor biases.

       Our control variable (Auditing), which was used to control for social norms of the

subjects, was found to have a strong significant effect in the model in both rounds (see Table

2). It was actually more powerful than the financial incentives itself. This implies another

important measure to be undertaken to increase auditors’ independence – an intense ethical

education of the profession. This finding is consistent with Johnstone et al. (2001) who claim

that regulation alone might not effectively mitigate independence risk and suggest that

regulation combined with professional training and high ethical standards might be effective.



                                               22
       Based on our analyses and results, our study advocates that due to persistence of the

auditor bias problem, only a combination of several measures would efficiently combat it – a

public oversight, mandatory audit firm rotation and a strong emphasis on ethics as a part of

auditor’s professional training. One of the advantages of the latter two measures is that in the

times of financial distress and deteriorated public finance situation they do not increase the

costs of the public oversight of the audit profession as they are performed within profession.

       The validity of our results is to be weighed in relation to the limitations of the study.

The first limitation is attributable to the fact that the subjects were students who were asked

to assume the role of experienced auditors. Their responses might, to some degree, represent

what the subjects thought an auditor would do, not what they would do if they were actually

auditors. This limitation could be eliminated if actual certified auditors were involved in the

experiment. However, to some extent we succeeded to control for this limitation by inviting

only accounting and finance students to the experiment. These students are most familiar with

the auditing profession and regulation. Moreover, they had on average 3.5 years of work

experience and many of them have worked as audit assistants in audit firms.

       A further limitation may be in our assumption (like the one in Moore et al., 2010) that

the subjects would be able to undo their bias from the initial decision when they changed the

role. It may also be that some subjects made their initial judgments unaware that they were

giving biased answers, and then, when asked to take the role of an expert, realized that they

had failed to consider the alternative point of view previously, and responded differently.

This limitation in our approach, however, does not inflate unconscious bias, but rather

underestimates it.

       We managed to uncover the effect of personal relationship on unconscious bias, to our

knowledge previously undetected empirically in the literature. Our results indicate that when

studied simultaneously, oversight has a significant negative effect on bias resulting from



                                              23
financial incentives and personal relationship, but with our method this effect couldn’t be

disentangled. This might suggest that personal relationship involves also conscious bias.

Furthermore, the reason why oversight has an effect on personal relationship may lie in the

fact that it comprises more than only financial punishment – we conjecture that it contains the

indication of peer or a public authority dissent, which is a factor of social norms that deter

auditors from biased decisions. Overall, we consider the experiment a sufficiently powerful

tool to reveal the effects of the studied variables.

       Despite of our somewhat inconclusive results we believe that the question of

conscious and unconscious biases in auditor’s decision-making is worth pursuing further.

Future research could complement our findings by looking at the interactive effects between

financial incentives and personal relationship on conscious and unconscious bias, it could

refine the measurement of unconscious bias and it could directly address the main and

interaction effects of various measures against both types of biases.



References

    1. Ajzen, I. (1991) The Theory of Planned Behavior, Organizational behavior and

        human decision processes, 50, pp. 179-211.

    2. Auditor Rotation: Musical Chairs (2011) The Economist, Sept 3, 2011.

    3. Bamber, E. M. and Iyer, V. M. (2007) Auditors’ identification with their clients and

        its effect on auditors’ objectivity, Auditing: A Journal of Practice & Theory, 26(2),

        pp. 1-24.

    4. Bazerman, M. H. and Moore, D. (2011) Is it time for auditor independence yet?

        Accounting, Organizations and Society, 36, pp. 310–312.

    5. Blay, A. D. (2005) Independence threats, litigation risk, and the auditor's decision

        process, Contemporary Accounting Research, 22(4), pp. 759-789.



                                                 24
6. Callao, S. and Jarne, J. I. (2010) Have IFRS Affected Earnings Management in the

   European Union? Accounting in Europe, 7(2), pp. 159-189.

7. Cameran, M., Prencipe, A. and Trombetta, M. (2009) Auditor tenure and auditor

   change: does mandatory auditor rotation really improve audit quality? Paper presented

   at AAA, Annual Meeting, New York, 1-5 August 2009.

8. Carcello, J. V. and Nagy, A. (2004) Audit firm tenure and fraudulent financial

   reporting, Auditing. A journal of Practice and Theory, 23(2), pp. 57-71.

9. Carpenter, T. D. and Reimers, J. L. (2005) Unethical and Fraudulent Financial

   Reporting: Applying the Theory of Planned Behavior, Journal of Business Ethics, 60,

   pp. 115–129.

10. Chen, K. Y., Elder R. J. and Liu J. (2005) Auditor independence, audit quality and

   auditor-client negotiation outcomes: some evidence from Taiwan, Journal of

   Contemporary Accounting & Economics, 1(2), pp. 119-146.

11. Cloyd, C. B. and Spilker B. C. (1999) The influence of client preferences on tax

   professionals'   search   for   judical        precedents,   subsequent   judgments   and

   recommendations, The Accounting Review, 74(3), pp. 299-322.

12. Cohen, D. A., Dey, A. and Lys, T. (2008) Real and accrual-based earnings

   management in the pre- and post- Sarbanes Oxley periods, The Accounting Review,

   83(3), pp. 757-787.

13. Dart E. (2011) UK investors’ perceptions of auditor independence, The British

   Accounting Review, 43(3), pp. 173-185.

14. DeAngelo (1981) Auditor independence, "low balling", and disclosure regulation,

   Journal of Accounting and Economics, 3(2), pp. 113-127.

15. Deis D. R. and Giroux G. A. (1992) Determinants of audit quality in the public sector,

   The Accounting Review, 67, pp. 462-479.



                                             25
16. Directive 2006/43/EC of the European Parliament and of the Council of 17 May 2006

   on statutory audits of annual accounts and consolidated accounts. http://eur-

   lex.europa.eu/LexUriServ/LexUriServ.do?uri=OJ:L:2006:157:0087:0087:EN:PDF

17. EC (European Commission) (2010) Green Paper - Audit Policy: Lessons from the

   Crisis, October (Brussels: European Commission).

18. EC (European Commission) (2011) Proposal for a Directive of the European

   Parliament and of the Council amending Directive 2006/43/EC on statutory audits of

   annual accounts and consolidated accounts (Brussels: EC).

19. EC (European Commission) (2011a) Proposal for a Regulation of the European

   Parliament and of the Council on specific requirements regarding statutory audit of

   public interest entities (Brussels: EC).

20. Evans L. and Nobes C. (1998) Harmonization relating to auditor independence: the

   Eighth Directive, the UK and Germany, The European Accounting Review, 7(3), pp.

   493-516.

21. Farmer, T. A., Rittenberg, L. E. and Trompeter, G. M. (1987) An investigation of the

   impact of economic and organizational factors on auditor independence, Auditing: A

   Journal of Practice & Theory, 7(1), pp. 1-14.

22. Gates, S. K., Lowe, D. J., Reckers, P. M. J. (2007) Restoring public confidence in

   capital markets through auditor rotation, Managerial Auditing Journal, 22(1), pp. 5-

   17.

23. Geiger, M.A. and Raghunandan, K. (2002) Auditor tenure and auditor reporting

   failure, Auditing. A journal of Practice and Theory, 21(1), pp. 12-67.

24. Gavious, I. (2007) Alternative perspectives to deal with auditors’ agency problem,

   Critical Perspectives on Accounting, 18(4), pp. 451-467.




                                              26
25. Hackenbrack K. and Nelson M. W. (1996) Auditors’ incentives and their application

   of financial accounting standards, The Accounting Review, 71(1), pp. 43-59.

26. Humphrey, C., Kausar, A., Loft, A. and Woods, M. (2011) Regulating audit beyond

   the crisis: A critical discussion of the EU Green Paper, European Accounting Review,

   20(3), pp. 431-457.

27. Jackson, A. B., Moldrich, M. and Roebuck, P. (2008) Mandatory audit rotation and

   audit quality, Managerial Auditing Journal, 23(5), pp. 420-437.

28. Johnson, V. E., Khurana I. K. and Reynolds, J. K. (2002) Audit firm tenure and the

   quality of financial reports, Contemporary Accounting Research, 19(4), pp. 637-660.

29. Johnstone, K. M., Sutton, M. H. and Warfield T. D. (2001) Antecedents and

   consequences of independence risk: framework for analysis, Accounting Horizons,

   15(1), pp. 1-18.

30. Kadous K., Kennedy S. J. and Peecher M. E. (2003) The effect of quality assessment

   and directional goal commitment on auditors’ acceptance of client-preferred

   accounting methods, The Accounting Review, 78(3), pp. 759-778.

31. Kadous K., Magro, A. M. and Spilker B. C. (2008) Do effects of client preference on

   accounting professionals’ information search and subsequent judgments persist with

   high practice risk? The Accounting Review, 83(1), pp. 133-156.

32. Kahneman, D. and Tversky, A. (1979) Prospect Theory: An Analysis of Decision

   under Risk, Econometrica, 47(2), pp. 263-291.

33. Krishnan, J. and Krishnan, J. (1996) The role of economic trade-offs in the audit

   opinion decision: an empirical Analysis, Journal of Accounting, Auditing and

   Finance, 11(4), pp. 565-586.

34. Kunda, Z. (1990) The case for motivated reasoning, Psychological Bulletin, 108(3),

   pp. 480-498.



                                          27
35. Lee, C. J. and Gu, Z. (1998) Low balling, legal liability and auditor independence,

   The Accounting Review, 73(4), pp. 533-555.

36. Lobo, G. and Zhou, J. (2006) Did conservatism in financial reporting increase after

   the Sarbanes Oxley Act? Initial evidence, Accounting Horizons, 20(1), pp. 57-73.

37. Lord, C. G., Lepper, M. R. and Preston, E. (1984) Considering the opposite: A

   corrective strategy for social judgment, Journal of Personality and Social Psychology,

   47, pp. 1231-1243.

38. Lord, A. T. (1992) Pressure: A methodological consideration for behavioral research

   in auditing, Auditing: A Journal of Practice & Theory, 11(2), pp. 89-108.

39. Mautz, R. K. and Sharaf, H. A. (1961) The philosophy of auditing (Sarasota:

   American Accounting Association).

40. Milgram, S. (1974) Obedience to authority (NY: Harper and Row).

41. Moore, D. A., Tanlu L. and Bazerman M. H. (2010) Conflict of interest and the

   intrusion of bias, Judgment and Decision Making, 5(1), pp. 37-53.

42. Moore, D. A., Tetlock, P. E., Tanlu, L. and Bazerman, M. H. (2006) Conflicts of

   interest and the case of auditor independence: Moral seduction and strategic issue

   cycling, Academy of Management Review, 31(1), pp. 10-29.

43. Morck, R. (2008) Behavioral finance in corporate governance: economics and ethics

   of the devil’s advocate, Journal of Management and Governance, 12, pp. 179–200.

44. Myers, J., Myers, L. A. and Omer, T. C. (2003) Exploring the term of the auditor-

   client relationship and the quality of earnings: A case for mandatory auditor rotation?

   The Accounting Review, 78(3), pp. 779-800.

45. Nelson, M. W. (2005) A review of experimental and archival conflicts-of-interest

   research in auditing. Conflicts of interest: challenges and solutions in business law,




                                          28
   medicine and public policy, pp. 41-69 (Cambridge and New York: Cambridge

   University Press).

46. Nelson, M. W. (2006) Response: Ameliorating conflicts of interest in auditing: effects

   of recent reforms on auditors and their clients, Academy of Management Review,

   31(1), pp. 30-42.

47. Nelson, M. W. (2009) A model and literature review of professional scepticism in

   auditing, Auditing: A Journal of Practice & Theory, 28(2), pp. 1-34.

48. Neuberg, S. L. and Fiske, S. T. (1987) Motivational influences on impression

   formation: outcome dependency, accuracy-driven attention, and individuating

   processes, Journal of Personality and Social Psychology, 53(3), pp. 431-444.

49. Rennie, M. D., Kopp, L. S. and Lemon W. M. (2010) Exploring trust and the auditor-

   client relationship: factors influencing the auditor’s trust of a client representative,

   Auditing: A Journal of Practice & Theory, 29(1), pp. 279-293.

50. Ruiz-Barbadillo, E., Gomez-Aguilar, N. and Carrera, N. (2009) Does mandatory audit

   firm rotation enhance auditor independence? Evidence from Spain, Auditing: A

   Journal of Practice & Theory, 28(1), pp. 113-135.

51. Sarbanes-Oxley Act of 2002. http://www.soxlaw.com/

52. Sikka, P. (2009) Financial crisis and the silence of the auditors, Accounting,

   Organizations and Society, 34, pp. 868-873.

53. Stickel, D. and Gamroth, C. (2006) Put your money in trust, CA Magazine, 139(7),

   pp. 47-55.




                                          29
APPENDIX

Variable abbreviations and coding

    I. Manipulated (independent) variables

Financial incentive (Incentive): value 0: fixed fee, value 1: future business (dependent on the

choice)

Personal relationship (Relation): value 0: anonymous, value 1: personal relationship

Oversight: mimicked by random number in the computer only in the second round of the experiment;

value 0: no oversight, value 1: oversight (at 50% of probability)



    II. Dependent variables:

Decision in the both rounds: Decision1 and Decision 2: value 0: low value of the investment, value

1: high value of the investment (in favour of the principal)

Bias (the difference between the investment value judged as auditor and as expert) in both rounds:

value -3: As auditor "80 mio", as expert "50 mio ", value -2: As auditor "80 mio ", as expert "60 mio ",

value -1: As auditor "60 mio ", as expert "50 mio ", value 0: no difference

Unconscious bias (Unconsc): value 0: conscious bias, value 1: unconscious bias



    III. Control variable

Auditing: value 0: the subject has not concluded the course of auditing, value 1: the subject has

concluded the course of auditing




                                                   30
FIGURES AND TABLES

             Figure 1: Financial incentive by Personal relationship plot for Decision 1, Round 1.




Figure 2: Financial incentive by Oversight plot for Decision in rounds 1 and 2 (Oversight present in round 2).




                                                     31
 Figure 3: Personal relationship by Oversight plot for Decision in rounds 1 and 2 (Oversight present in round
                                                       2).




                   Table 1: Test of Between-Subjects Effects; Dependent variable: Decision1

Dependent Variable: Decision1
Source                Type III Sum of        df       Mean Square      F          Sig.
                          Squares
                                        a
Corrected Model                 2.511             3            .837    3.650        .015
Intercept                       16.492            1          16.492   71.921        .000
Incentive                        1.000            1           1.000    4.363        .039
Relation                          .115            1            .115     .501        .481
Auditing                         1.358            1           1.358    5.920        .017
Error                           24.766         108             .229
Total                           47.000         112
 Corrected Total                27.277          111
a. R Squared = ,092 (Adjusted R Squared = ,067)




                                                        32
             Table 2: Test of Between-Subjects Effects. Dependent variable Decision in both rounds

Measure: Oversight
Source             Type III             df     Mean Square                 F              Sig.           Partial Eta            Observed
                   Sum of                                                                                    Squared                 Power
                  Squares                                                                                                             (r)
Intercept             24.462            1              24.462          82.047             .000                .432                   1.000
Incentive              1.829            1                 1.829           6.134           .015                .054                   .690
Relation                .452            1                  .452           1.518           .221                .014                   .231
Auditing               3.103            1                 3.103        10.406             .002                .088                   .892
Error                 32.200        108                    .298




             Table 3: Difference in the decision taken in the role of auditor and expert in both rounds

                                                                         Difference in decision,         Difference in decision,
                                                                                round 1                         round 2
                                                                         Frequency     Percent           Frequency     Percent
Conscious bias:                                                                60              53,6            48             42,8

  As auditor "80", as expert "50"                                              17              15,2            11             9,8

  As auditor "80", as expert "60"                                              22              19,6            10             8,9

  As auditor "60", as expert "50"                                              21              18,8            27             24,1

Same answer as auditor and expert:                                             52              46,4            64             57,1

  No difference: As auditor "60", as expert "60"                               44              39,3            56              50

  Unconscious bias: As auditor "80", as expert "80"                            8               7,1              8             7,1

Total                                                                          112             100             112            100




                        Table 4: Logistic regression. Dependent variable: Unconscious bias 1

                                    B           S.E.              Wald               df               Sig.           Exp(B)
         a
Step 1        Incentive(1)          -.012          .796             .000                   1            .988           1.012
              Relation(1)           2.036          .939            4.703                   1            .030            .131
              Auditing(1)         -1.281           .964            1.768                   1            .184           3.602
               Constant             2.093          .810          6.670                     1            .010            .123
a. Variable(s) entered on step 1: Incentive, Relation, Auditing.




                                                                  33
              Table 5: Test of With-in Subjects Effect. Dependent variable Decision in both rounds

Measure: Oversight
Source                                                 Type III      df      Mean             F        Sig.      Partial       Observed
                                                                                                                                        a
                                                       Sum of                Square                                Eta         Power
                                                      Squares                                                    Squared          (r)
Oversight               Sphericity Assumed              .621         1         .621         5.891      .017             .052     .672
                                                                     1
Oversight *             Sphericity Assumed              .012                   .012          .110      .741             .001     .062
Auditing
                                                                     1
Oversight *             Sphericity Assumed              .003                   .003          .026      .873             .000     .053
Incentive
                                                                     1
Oversight *             Sphericity Assumed              .029                   .029          .272      .603             .003     .081
Relation
                                                                     1
Oversight *             Sphericity Assumed              .232                   .232         2.198      .141             .020     .312
Incentive *
Relation
Error(Oversight)         Sphericity Assumed            11.280     107        .105
    a. Computed using alpha = .05
    b. As in our design there are only two conditions of treatment (with and without oversight), the sphericity, i.e.
        equality of variances across conditions, is met.




                                                            34
EXPERIMENTAL MATERIAL

Part 1

Background

You work as an auditor for an established audit firm. Currently, you are auditing financial statements of a major
public company A. As it seems, the company will report a modest operating profit of 30 million EUR. Some
years ago Company A bought shares of Company B, which operates in the food industry. At the time Company
A paid 100 million EUR for this significant stake in Company B that provided Company A with the significant
influence over Company B. Due to the financial crisis and the fact that cash flows of Company B are not
meeting the expectations, Company A is faced with the issue of investment impairment due to the decrease in
the fair value of its stake in Company B. The current market value of the mentioned financial investment is 50
million EUR. The Chief Financial Officer (CFO) of Company A believes that it is inappropriate to value their
investments in Company B by using the market value, because of the extremely low liquidity in the market.
Therefore, he proposed to value the investment with a valuation model. As the auditor you are currently
reviewing the valuation model the CFO has proposed. In the model the CFO assumed a 5% growth of cash
flows without any apparent capital expenditure to support this growth. The valuation based on this assumption
gives the value of the investment into Company B of 80 million EUR.

Historical patterns show that the growth in demand for products of Company B is closely related to the growth
of the purchasing power of population. Generally, the purchasing power is reflected in the growth of the gross
domestic product. According to forecasts GDP is expected to grow in the next years by 1% only. At the
assumption of a 1% growth rate in the valuation model, the estimated value of the investment is 60 million
EUR.

If the investment value proposed by the CFO will be used in the financial statements, an impairment of 20
million EUR will have to be accounted for. If, on the other hand, the lower investment value will be recognized,
an impairment of 40 million EUR will be needed. In this latter case the valuation of the investment will cause a
net loss of Company A, while the valuation according to the assumption proposed by the CFO, would not.

Summary
Purchase price of the investment in Company B: 100 million, projected growth of GDP: 1%.
  Valuation          Mark-to-        Assumption     Impairment         Operating   Net profit / loss of
    model          model value        regarding         of the          profit of   Company A after
                                    growth of cash   investment       Company A       impairment
                                        flows
    CFO’s        80 million              5%          20 million        30 million       10 million
 Alternative     60 million              1%          40 million        30 million      -10 million

As the auditor you are aware of the importance of your professional reputation and the fact that the auditor is
primarily committed to serve public interest - namely, that the creditors, shareholders and other users of
financial information get the reliable information about fair presentation of the financial position and financial
performance of Company A.

Your audit firm was inspected by the Agency for public oversight just last year. Therefore, no oversight will
take place in your audit firm this year.

Either:
FINANCIAL INCENTIVE (FIXED): Your payment for auditing services is fixed, regardless which of the two
valuation models you confirm. Your payment is EUR 2.

Or:
FINANCIAL INCENTIVE (FUTURE BUSINESS): Your payment for auditing service is fixed and amounts to
2 EUR. If you will require the recognition of the lower value of the investment, you can with certainty expect
that next year you will not get the audit job in Company A. If you will approve the higher value of the
investment, you may with certainty expect that your contract in the Company A will be renewed in the next
years. Present value of future business with Company A amounts to 1 EUR. Your total compensation will hence
amount to 3 EUR (2 EUR of fixed fee + 1 EUR of future business).



                                                       35
Your decision I

Please specify which valuation model you will approve in the financial statements audit (tick the box).

 □ the valuation model that takes into account the lower growth rate: You request that the value of the
   investment is set at 60 million EUR and that the impairment of the investment amounts to 40 million
   EUR.

 □ the valuation model that takes into account the higher growth rate: You approve the model that was
   prepared by the CFO with the value of the investment at 80 million EUR. You agree that the impairment
   of the financial investment amounts to 20 million EUR.


When you are ready to continue please raise your hand. The administrator will take this page and then you will
proceed to the next part.


Part two

You are making the same decision as before, but circumstances have changed. The Act on Auditing has been
amended and it is possible that your audit will be subject to oversight.

Either:
FINANCIAL INCENTIVE (FIXED): As before, your payment is fixed (2 EUR).

Or:
FINANCIAL INCENTIVE (FUTURE BUSINESS): Your payment for auditing service is fixed and amounts to
2 EUR. If you will require the recognition of the lower value of the investment, you can with certainty expect
that next year you will not get the audit job in Company A. If you will approve the higher value of the
investment, you may with certainty expect that your contract in the Company A will be renewed in the next
years. CFO has increased the offer of future business. Present value of future business with Company A
amounts to 4 EUR. Your total compensation will hence amount to 6 EUR (2 EUR of fixed fee + 4 EUR of
future business).

If you approve the valuation model with a lower growth rate, it is unlikely that you get disapproval of your audit
by the oversight.

If you approve the valuation model with a higher growth rate, there is a 50 % probability that oversight will
disagree with your decision. You will get fined and reprimanded. In such case your payment will be 0 EUR.

Your decision II

Considering the new information (regarding the possibility of the oversight), please specify which valuation
model will you approve (tick the box):

 □    the valuation model that takes into account the lower growth rate: You request that the value of the
      investment is set at 60 million EUR and that the impairment of the investment amounts to 40 million
      EUR.

 □    the valuation model that takes into account the higher growth rate: You agree with the model that
      was prepared by the CFO and with the value of the investment at 80 million EUR. You agree that the
      impairment of the investment amounts to 20 million EUR.

When you are ready to continue please raise your hand. The administrator will take this page and then you will
proceed to the next part.


EXIT QUESTIONS



                                                       36
    1. How well do you know the person who represented the CFO who prepared the valuation model?

 I don't know who         We do not know          We know each              We are friends.       We are close
this person is, (s)he      each other.               other.                                        friends.
was not introduced
    to me in the
    experiment.
          1                      2                      3                         4                       5

    2. When deciding between the two alternative options, to what extent did you take into account the
       following factors (please, circle):

                                                                   Not at              Somewhat               A lot
                                                                    all
a. Financial incentives                                              1            2           3     4          5
b. Arguments presented to me by the CFO                              1            2           3     4          5
c. My trust in the CFO                                               1            2           3     4          5
d. Probability of the oversight.                                     1            2           3     4          5

    3. Now suppose that in this experiment you would not be in the role of the auditor, but that you would be
       a financial expert and would give advise to an investor who is interested in the acquisition of Company
       A. He/she would you hire you to give him/her your opinion regarding the value of the investment in
       Company B. Which option would you choose (please, tick the box).

          □     I would tell him/her that the value of the investment in Company B is 50 million EUR (its
                current market value).

          □     I would tell him/her that the value of the investment in Company B is 60 million EUR
                (according to the model with the lower growth rate that follows the growth of the gross
                domestic product).

          □     I would tell him/her that the value of the investment in Company B is 80 million EUR
                (following the model with the higher growth rate, prepared by the CFO).

    4.   Sex (please circle):        male   female

    5.   Your age: ______

    6.   Years of work experience you have (including student work): ______ years

    7.   Your study programme (please circle):       undergraduate    postgraduate

    8.   Have you been enrolled in the course of Auditing?   Yes     No




                                                      37

								
To top