unc charlotte pages · web view2016/10/06 · the authors acknowledge the helpful comments of lori...
TRANSCRIPT
Auditor Narcissism and Auditor-Client Negotiations
Bryan K. ChurchGeorgia Institute of Technology
Narisa Tianjing DaiUniversity of International Business and Economics
Xuejiao LiuUniversity of International Business and Economics
Xi (Jason) KuangGeorgia Institute of Technology
October 2016
Please do not quote, cite, or circulate without authors’ permission.
Note: The four authors contributed equally to the paper. The authors acknowledge the helpful comments of Lori Bhaskar, Jeff Cohen, Jessen Hobson, Eric Johnson, Jonathan Stanley, Karl Wang, Arnie Wright, and participants at the ABO midyear conference.
Auditor Narcissism and Auditor-Client Negotiations
Abstract: This paper reports the results of an experiment designed to examine the effect of auditor narcissism on auditor-client negotiations. We contend that narcissistic characteristics fuel auditors’ competitiveness and embolden them to stand firm in negotiations, potentially curbing clients’ propensity to report opportunistically. Our experiment includes auditor-client dyads who negotiate over a reported asset value. As predicted, we find that narcissistic auditors are more likely to be involved in negotiations that reach an impasse or take longer to resolve, controlling for client narcissism. Importantly, narcissistic auditors negotiate reported asset values that reflect less aggressive reporting choices. Consistent with our experimental results, supplemental analyses using data collected from a naturally-occurring auditor-client setting suggest that auditor narcissism is positively associated with audit delay and negatively associated with clients’ absolute and negative discretionary accruals, again controlling for client narcissism.
1. Introduction
This paper reports the results of an experiment designed to investigate the effect of
auditor narcissism on auditor-client negotiations, which ultimately impact the financial-reporting
process (Beattie et al. 2000; Gibbins et al. 2001). Narcissism is a personality trait that involves a
grandiose preoccupation with one’s superiority and self-importance (American Psychiatric
Association 2013), with people existing along a continuum from low to high (Foster and
Campbell 2007; Grijalva and Harms 2014).1 We posit that auditor narcissism can have a
pronounced effect on negotiations over accounting disagreements, helping auditors curtail
managers’ opportunistic reporting behavior.
Extant evidence suggests that corporate executives often possess narcissistic
characteristics (Chatterjee and Hambrick 2007; Nevicka et al. 2011), underlying aggressive
reporting choices (Hales et al. 2012; Ham et al. 2015; Judd et al. 2015; Majors 2015). For
example, more narcissistic CFOs are associated with less timely loss recognition, higher levels of
accruals-based and real earnings management, and a greater chance of accounting restatements
(Ham et al. 2015). Relatedly, auditors anticipate greater challenges when dealing with executives
who exhibit narcissistic characteristics (e.g., grandiosity, entitlement, and exploitativeness).
Under such conditions, auditors’ assessment of fraud risk is heightened (Johnson et al. 2013),
audit fees are increased (reflecting greater potential exposure), and the chance of auditors
resigning is elevated (Judd et al. 2015). We extend this line of study by examining auditor
narcissism as a means to bolster audit quality.
1 At the extreme, narcissism represents a personality disorder. The rate of pathological (clinical) narcissism in community samples ranges from 0 to 6.2 percent, with a mean of 1.06 percent (Dhawan et al. 2010). In this paper, we are interested in non-pathological (nonclinical) narcissism, which shares characteristics with pathological narcissism, but at lower levels. Notably, researchers provide evidence that non-pathological narcissism is on the rise (Twenge et al. 2008).
2
We use an experimental-economics approach to investigate the effect of auditor
narcissism on the financial-reporting process. We create auditor-client dyads, whereby paired
participants negotiate over reported amounts. Prior studies suggest that financial statements are a
joint product determined by auditor-client interactions (Antle and Nalebuff 1991; Beattie et al.
2000; Salterio 2012), such that negotiations are critical in establishing reported amounts (e.g.,
resolving disagreements over the application of accounting standards, estimating uncertain
amounts, or deciding whether to book or waive misstatements). We offer insight into how
auditor narcissism affects auditor-client negotiations.
Accounting researchers have shown considerable interest recently in auditors’ personality
traits (Scofield et al. 2004; Hurtt 2010; Quadackers et al. 2014; Bauer 2015; Bhaskar et al. 2015;
Church et al. 2015; Fitzgerald et al. 2015). Akers et al. (2014) find that accounting partners, on
average, have the highest narcissism score across public accountants’ professional ranks.
Furthermore, partners’ average scores are highest on certain dimensions, including authority,
entitlement, and self-sufficiency (not needing others). We put forth that partners’ narcissism can
have a marked effect on auditor-client negotiations, influencing the negotiation process and
potentially bettering audit quality.2
Partners’ narcissism is expected to impact auditor-client negotiations over accounting and
reporting matters. Narcissists are self-absorbed and unyielding in negotiations, failing to see the
merits of their counterparts’ claims (Greenhalgh and Gilkey 1997). Narcissists lack empathy
(Watson et al. 1984), making it difficult for them to understand the expectations, viewpoints,
motivations, and emotions of their counterparts (Greenhalgh and Gilkey 1997; Park et al. 2013).
2 Our study is confined to the effect of auditor narcissism on auditor-client negotiations over reported amounts. We acknowledge that auditor narcissism can have other potential effects, for better or worse. For example, an engagement partner’s narcissism might impact audit-team performance as well as auditor-client relationships (e.g., Resick et al. 2009; Golec de Zavala et al. 2013; Grijalva and Harms 2014; O’Neill and Allen 2014). Such issues are beyond the scope of the current study.
3
Hence, narcissists are less agreeable in situations involving interpersonal conflict and even
antagonistic in negotiations (Miller and Campbell 2008; Grijalva and Harms 2014). Auditors’
narcissism causes them to be dogged and uncompromising, refusing to condone clients’
aggressive reporting choices. Such behavior can lengthen the negotiation process, forestalling
agreements, and in some instances bring about an impasse. Indeed, auditor narcissism can lead to
audit delay. Though such an outcome is undesirable, auditors may have little alternative, beyond
succumbing to clients’ demands.
Narcissists are driven to come out on top, taking whatever actions are necessary to win
(Ryckman et al. 1994; Watson et al. 1997–1998; Luchner et al. 2011). They are ultracompetitive,
emboldened to stand firm in negotiations. Auditors’ narcissism propels them to hold their ground
on matters involving accounting disagreements (Ma and Jaeger 2005; Hüffmeier et al. 2014).
They are prone to resist clients’ efforts to report opportunistically. Indeed, partners’ narcissism
can serve to counter the effects of executives’ narcissism, which have been associated with
clients’ earnings management and accounting restatements (Ham et al. 2015; Judd et al. 2015).
The end result is that, when disagreements are resolved, reported amounts are less aggressive.
We are primarily interested in auditor narcissism and how its shapes reported amounts,
which reflect a joint auditor-client outcome. The effect of auditor narcissism, however, is subject
to two important contextual influences: bargaining power and client narcissism. In our
experiment, we create a setting in which the auditor and client have equal bargaining power,
precluding either party from having an advantage in negotiations. Equal bargaining power
provides an important baseline, which can be used to gauge auditor-client negotiations. For client
narcissism, we measure this variable and include it in our data analyses.
4
We conduct an experimental-economics study using undergraduate business students at a
major university in China. Prior to conducting our experiment, we administer the 40-item,
forced-choice narcissistic personality inventory (NPI), a commonly used self-report measure of
narcissism in normal populations (Raskin and Terry 1988; Miller et al. 2014). Based on
participants’ NPI scores, we assign them to auditor-client dyads and have them negotiate over a
series of periods. We examine the linkage between narcissism and the negotiation process and
outcomes. Our primary findings indicate that, as predicted, narcissistic auditors are more likely
to be involved in protracted negotiations and, further, to agree on reported amounts that are less
aggressive.
To gain further insight into the effect of auditor narcissism on the financial-reporting
process, we perform supplemental analyses using archival data on publicly listed companies in
China. For these companies, audit reports are signed by the auditor, which is noteworthy because
signature size has been used by others as a proxy for narcissism (Ham et al. 2014; Ham et al.
2015). Thus, we are able to measure auditor narcissism in a naturally-occurring, auditor-client
setting. We also obtain the CFO’s signature from the annual report, allowing us to control for
client narcissism. We find that auditor narcissism is positively associated with audit delay, which
is consistent with negotiations being lengthy (e.g., Salterio 2012). In addition, auditor narcissism
is negatively associated with clients’ signed, positive discretionary accruals and clients’ absolute
discretionary accruals, suggestive of less aggressive reporting choices. The supplemental results
are consistent with our experimental findings. Importantly, the use of multiple research methods
helps overcome the potential drawbacks inherent in each method, lending credence to our theory
and enhancing the external validity of our findings (Smith 1975; Jick 1979; Scandura and
Williams 2000). Researchers have long advocated the use of a multi-method approach, as it
5
supports triangulation and instills confidence in convergent findings (e.g., Einhorn and Hogarth
1981; Birnberg et al. 1990; Brewer and Hunter 2006; Hageman 2008).
Our findings contribute to a growing accounting literature, which suggests that auditors’
personality traits impact audit quality. For example, Quadackers et al. (2014) show that auditors
who exhibit lower levels of interpersonal trust make more skeptical judgments in high-risk
situations, potentially for the betterment of audit quality. Church et al. (2015) find that
dispositional perspective taking enhances auditors’ assessment of clients’ earnings, promoting
higher quality financial reporting. We add to this literature by showing that narcissism affects the
outcome of auditor-client negotiations, dampening clients’ attempts to report aggressively.
Our findings also have significant practical implications. Employees’ attributes (e.g.,
personality traits) have become increasingly important in accounting firms’ recruiting efforts and
promotion decisions (Campbell 2012). Our findings highlight the potential effects of auditor
narcissism on the audit process and outputs. Accounting firms may need to consider these effects
to enhance audit efficiencies and audit quality by assigning the “right” personnel to specific jobs.
For example, audit partners who are more narcissistic may be better able to cope with
contentious clients.
The remainder of this paper is organized as follows. Section 2 offers a framework for our
study and provides a basis to develop the research hypotheses. Section 3 describes our
experiment, including participants, design, and procedures. Section 4 presents the experimental
results and discusses the findings. Section 5 reports the results of additional archival analyses,
which supplement our experimental findings. Section 6 summarizes our study and offers
concluding remarks.
2. Framework
6
Auditor-client interactions occur on an ongoing basis throughout the audit process.
Disagreements give rise to conflict, which provides the basis for negotiations (e.g., Greenhalgh
1987; McCracken et al. 2008). Importantly, audited financial statements represent a negotiated
outcome; a joint statement of the auditor and manager (Antle and Nalebuff 1991). Beattie et al.
(2000) survey finance directors and audit partners and provide evidence that negotiations
frequently involve subjective accounting matters (e.g., fair value assessments in acquisitions).
Gibbins et al. (2007) survey CFOs and document that negotiations entail divergent preferences,
new and unique accounting issues, and complex accounting matters. McCracken et al. (2008)
interview CFOs and audit partners and find that accounting standards and interpretations often
are invoked in negotiations. As such, auditor-client negotiations can have a material effect on
financial statements (Beattie et al. 2000; Gibbins et al. 2001). Below, we consider how auditor
narcissism affects the negotiation process and negotiated outcomes.
2.1. Narcissism and the Length of Negotiations
Auditors’ narcissistic characteristics can impact the negotiation process, potentially
delaying agreement on reported amounts and, at the extreme, leading to a stalemate or impasse.
A narcissistic personality entails a heightened self-image, including a sense of dominance,
superiority, and grandiosity (Emmons 1984, 1987; Raskin and Terry 1988; Kubarych et al.
2004). This facet of narcissism is linked to arrogance, self-importance, self-confidence, and
assuredness. It is associated with an undue focus on self (Emmons 1987) and a need for power
(Carroll 1987), manifesting in self-aggrandizing behavior (Morf and Rhodewalt 2001). Kong
(2015) finds that negotiators’ narcissism is negatively associated with their assessments of
counterparts’ competence, indicating that counterparts are viewed as inferior. Narcissists are
inclined to disparage others who doubt or challenge them. Rhodewalt and Morf (1995) suggest
7
that, in the face of interpersonal conflict, narcissists react with cynical hostility, antagonism, and
combativeness. Simply put, narcissists are more likely to be disagreeable in negotiations than
others (Miller and Campbell 2008). Narcissists’ self-aggrandizement makes it more difficult to
reach an agreement in negotiations, thereby slowing the process.
A narcissistic personality also encompasses individuals’ sense of entitlement and a lack
of regard for others (Emmons 1984, 1987; Watson et al. 1984; Raskin and Terry 1988). A feeling
of entitlement is indicative of believing that one deserves more than others, and being willing to
exploit others for selfish reasons. Grijalva and Harms (2014) suggest that narcissists approach
negotiations with an individualistic orientation: they focus primarily on personal gain, with little
concern for how their actions affect others (Campbell et al. 2005). Such an approach can delay
agreement (De Dreu et al. 2000) and, in some instances, produce an impasse (Babcock and
Loewenstein 2005).
Lastly, narcissists lack cognitive and emotional empathy, making it hard for them to
understand others’ viewpoints and feelings (Watson et al. 1984; Watson et al. 1992). Hepper et
al. (2014) find that young male adults currently serving a prison sentence have higher levels of
narcissism, particularly entitlement, than those without a criminal history and, further, that the
relationship is mediated by low levels of cognitive and emotional empathy. Narcissists’
insensitivity to others can hinder negotiations, extending the process. Based on the preceding
discussion, we posit that auditor narcissism impacts the time needed to resolve disagreements
over reported amounts. Our first research hypothesis is formally stated as follows.
H1: Auditor narcissism is positively associated with the length of the negotiation process.
2.2. Narcissism and Negotiation Outcomes
8
Research on auditor-client negotiations suggests that disagreements generally are
resolved, resulting in an unqualified audit opinion, with the auditor being reappointed (Gibbins et
al. 2001). A common negotiation outcome is an agreement between the two parties’ original
positions, though in some instances the outcome is in line with one party’s initial stance (Gibbins
et al. 2001; Gibbins et al. 2007). Negotiated outcomes tend to be distributive, meaning that one
or neither party wins (Gibbins et al. 2001; Gibbins et al. 2005, 2007; Brown and Wright 2008;
Brown and Johnstone 2009). We maintain that auditor narcissism affects the auditor’s need to
win and, thus, has a bearing on negotiation outcomes, impacting whether reported amounts are
less aggressive (as preferred by the auditor), holding client narcissism constant.3
Narcissists possess characteristics that drive them to stand their ground in negotiations.
Narcissists are described as self-absorbed and self-serving (Greenhalgh and Gilkey 1997;
Ronningstam 2009). Such characteristics compel them to exert their superiority and dominance
in negotiations, whereby others eventually accede to their demands. Indeed, prior findings
document that narcissism is positively associated with individuals’ competitiveness (i.e., their
desire to win in interpersonal settings) and hypercompetitiveness (i.e., their need to win at all
costs in order to feel superior) (Ryckman et al. 1994; Watson et al. 1997–1998; Luchner et al.
2011). Park et al. (2013) conduct a simulated one-shot negotiation with MBA students and find
that narcissists generate greater economic gains than non-narcissists; however, narcissists also
incur greater interpersonal loss (i.e., they are judged to be less trustworthy than non-narcissists).
Relatedly, Campbell et al. (2005) conduct an experiment that mimics the tragedy of the
commons dilemma, with participants making decisions over a series of periods.4 The study’s
3 The current study focuses on auditor-client negotiations over reported amounts. Specific factors that potentially affect each party’s incentive are subsumed in players’ payoff functions. For a discussion of these factors, refer to Brown and Wright (2008), Salterio (2012), and Brown-Liburd et al. (2016). 4 In Campbell et al. (2005), participants assume that they work for a forestry company whose objective is to harvest as much timber as possible, producing short-term gains. The forest represents a common resource that is used by others, and harvested by competing companies. Though the forest replenishes itself, the resource eventually will be
9
findings indicate that narcissists are more likely to realize short-term success, benefitting self,
but at a long-term cost to others.
Narcissism also is linked to individuals’ assertiveness, aggressiveness, antagonism, and
intransigence (Baumeister et al. 2000; Hopwood et al. 2013; Küfner et al. 2013; Furnham and
Crump 2014; Jones and Neria 2015). These characteristics underscore narcissists’
competitiveness in negotiations and manifest in the use of a contending strategy (Sandy et al.
2006), whereby individuals focus on achieving a specific outcome with little regard for their
counterparts’ welfare (Wang and Tuttle 2009). A contending strategy involves demanding
concessions from one’s counterparts, while at the same time resisting offers and counteroffers.
Contending tactics include asserting positional commitments, arguing forcefully and
persuasively, making threats, and giving ultimatums (Carnevale and Pruitt 1992).
Previous research suggests that auditors use contending tactics when clients are inflexible
on a disputed accounting position (Gibbins et al. 2010; Hollindale et al. 2011). In such cases,
auditors put forth arguments to substantiate their positions and also to refute clients’ positions
(Weingart et al. 1996; Hyder et al. 2000). Auditors also can threaten to modify their audit
opinions in order to coerce concessions from clients (Bame-Aldred and Kida 2007). For our
purposes, narcissists’ use of contending tactics, such as being intransigent and unyielding, can
provide a means to achieve a desired outcome (e.g., Barry and Friedman 1998). Ma and Jaeger
(2005) conduct simulated negotiations with undergraduates and find that aggressive and
combative tactics lead to outcomes that are consistent with negotiators’ preferences. Hüffmeier
et al. (2014) perform a meta-analysis of 34 studies to examine the effect of bargaining strategy
on negotiation outcomes. The study’s findings indicate that hardline bargaining (i.e., making
depleted if the rate of harvesting is too great. The long-term cost of depletion is the foregone opportunity to harvest timber in the future.
10
extreme initial offers and minimizing own concessions) results in larger economic outcomes
(i.e., hardline bargainers do better than others in negotiations). In fact, hardline bargaining may
even be fitness-enhancing from an evolutionary perspective (Heifetz and Segev 2004).
We contend that auditor narcissism is associated with an inherent drive to prevail in
negotiations with clients. The drive to win is underscored by a focus on self, providing an avenue
to exert one’s dominance and superiority. Further, the drive to win triggers competitive
behaviors, including steadfast and uncompromising negotiating tactics. Notably, Trotman et al.
(2009) provide evidence that audit partners consider tough and conservative stances when
contemplating inventory write-downs as part of pre-negotiation planning (i.e., expectations that
partners bring to negotiations). We argue that narcissistic auditors are less inclined to move away
from such hardline positions, even though the positions are at odds with client preferences.5 In
turn, narcissist auditors are more likely to get their way in negotiations than others, demanding
that clients opt for less aggressive reporting choices. Based on the preceding discussion, our
second research hypothesis is formally stated as follows.
H2: Auditor narcissism is positively associated with negotiation outcomes that reflect less aggressive reported amounts.
As mentioned previously, in this study we focus on the effect of auditor narcissism on the
negotiation process and outcome. The way in which narcissistic characteristics influence
negotiators’ behavior, however, is not specific to the role of the negotiator. Our arguments
5 As noted, Trotman et al. (2009) do not examine the negotiation process or outcomes, rather pre-negotiation planning. Brown and Johnstone (2009) investigate auditor-client negotiations using a computer-simulated client, which is programmed to take a very aggressive position on revenue recognition. The study’s findings indicate that experienced auditors are less willing to deviate from their initial position than inexperienced auditors. While Brown and Johnstone (2009) offer valuable insights, it is important to examine negotiations when both parties are active, human participants. Prior research indicates that, in an interactive setting, behavior with another human often differs from behavior with a computerized robot (Calegari et al. 1998; Mayhew et al. 2004). Further, the client’s initial position in Brown and Johnstone (2009) is borderline acceptable, arguably making it easier for experienced auditors to hold their ground. By comparison, when auditor-client dyads are humans, initial positions are determined endogenously and variation arises, which allows us to more generally examine the effect of narcissism on the negotiation process and outcomes.
11
underlying the effect of auditor narcissism on negotiations could just as easily be applied to
client narcissism. That is, ceteris paribus, narcissistic clients could be more likely to have
protracted negotiations and obtain a negotiated outcome in their favor (e.g., more aggressive
reporting choices), as compared to non-narcissists. In our hypotheses tests, we control for client
narcissism and also report the effect of client narcissism on negotiations.
3. Research Method
3.1. Participants and Design
We design a negotiation experiment in which we manipulate the composition of auditor-
client dyads based on participants’ predisposition to engage in narcissistic behavior.
Approximately two weeks before conducting the experiment, we administer the 40-item, forced
choice narcissistic personality inventory (NPI) to 662 undergraduate business students at a major
Chinese university (Raskin and Terry 1988).6 The NPI is a standard measure of non-pathological
(nonclinical) narcissism, and it has been validated using undergraduate students in China. Kwan
et al. (2009) document that the responses of Chinese undergraduates are correlated with key
features of narcissism, self-esteem, and self-efficacy, and the correlations are similar to those
reported in relevant research using U.S. undergraduates.
The range of possible NPI scores is 0 – 40, with higher scores representing higher
degrees of narcissism. In a sample of 3,445 adults worldwide, Foster et al. (2003) report an
average NPI score of 15.2, and the average is similar (15.3) restricting the sample to respondents 6 The 40-item NPI provides a measure of respondents’ grandiose narcissism, which is the focus of our study. We note that the narcissism literature identifies two subtypes of character style: grandiose and vulnerable, also referred to as overt and covert narcissism, respectively. The grandiose subtype is characterized as self-assured, arrogant, aggressive, entitled, exploitative, and envious, with grandiosity being tied to self-presentation (Wink 1991; Dickinson and Pincus 2003). By comparison, the vulnerable subtype is characterized as insecure, self-inhibited, hypersensitive, anxious, timid, and self-entitled, harboring unconscious feelings of grandeur (Wink 1991; Miller and Campbell 2008). In this paper, we focus on the grandiose subtype because it is more likely representative of successful business professionals (e.g., audit partners and executives). More notably, grandiose narcissists tend to be extraverted, energetic, self-confident, experience-seeking, conceited, and well-suited for leadership positions (Raskin and Novacek 1989; O’Reilly et al. 2014). The NPI and experimental stimuli are administered in Chinese. We use the back-translation method to ensure that content remains the same (Brislin 1970; Hui and Triandis 1985).
12
who reside in the U.S. (n = 2,546). Akers et al. (2014) administer the NPI to practicing
accountants in the Midwest U.S. (n = 279) and find that the average score is 14.7. The average is
slightly higher (15.3) including only those in auditing (n = 155). In the current study, the average
NPI of the 662 Chinese undergraduates is 13.7. We suggest that individuals’ responses to the
NPI in China are comparable to those in the U.S.
We partition our sample into terciles based on their NPI score. We recruit from the top
and bottom third for our experiment, representing High (denoted H) and Low (denoted L)
narcissists, respectively. One hundred ninety-six students agree to participate in our negotiation
experiment. The high narcissists who take part have NPI scores ranging from 17 to 34, with an
average of 22.5 (n = 99). The low narcissists, on the other hand, have NPI scores ranging from 0
to 9, with an average of 6.0 (n = 97).
Our experimental participants have an average age of 20.3 years and 75 percent are
female.7 In our experiment, one half of the participants are assigned the role of auditor and the
other half client.8 We create auditor-client dyads that negotiate over reported amounts. The
experiment consists of a series of period, with participants re-paired each period. Auditor-client
incentives are conflicting (described later), providing a basis for negotiations. We configure the
composition of the dyads such that both participants are H, both are L, or one is H and one L. We
have four experimental groups denoted HH, LL, HL, and LH, where the first letter represents the
auditor-participant’s narcissism and the second letter the client-participant’s narcissism.
7 We recognize that auditor-client negotiations likely comprise a smaller proportion of females than our sample. However, we do not expect gender to affect the results of our negotiation experiment. As described subsequently, the experiment involves an abstract setting, where participants negotiate over an economic outcome that unambiguously determines each party’s payoff. Participants negotiate anonymously (via networked computers), so they do not know the identity or gender of their bargaining partner. Under such conditions, the effects of gender are minimized when considering the negotiation process and outcomes (Mazei et al. 2015). In addition, we examine gender when analyzing the experimental data and find it is not statistically significant. Further, including this variable does not alter the results (not tabulated).8 In the experimental instructions, we use the terms verifier and reporter to represent the auditor and client, respectively. We use the terms auditor and client in the text for expositional convenience.
13
We systematically configure the dyads, as described above, for two reasons. First, the
narcissism of each role (auditor-client) is held constant over the course of an experimental
session. Second, although our focus is on auditor narcissism, client narcissism also can impact
the negotiation process and negotiated outcomes. For example, prior research suggests that
narcissistic clients report aggressively (Ham et al. 2015; Judd et al. 2015). Hence, we control for
client narcissism in establishing auditor-client dyads. In addition, we include client narcissism in
our data analyses and present relevant results.
3.2. Experimental Procedures
The experiment is administered in a university computer lab. Based on students’
availability, we conduct 12 experimental sessions, with three sessions per experimental group.
The number of participants (dyads) per session ranges from 12 to 20 (6 to 10). Overall, we have
25, 24, 22, and 27 dyads for the HH, LL, HL, and LH groups, respectively. Once participants
arrive, instructions are distributed and read aloud by one of the researchers. The instructions
describe participant pairings, the negotiation setting and protocol, and participants’ payoffs.
Subsequently, participants answer quiz questions covering the instructions. The quiz is included
to ensure that participants understand the instructions, and participants are not allowed to
proceed until all questions are answered correctly.
Experimental sessions consist of 18 independent, one-minute periods. Participants are not
informed beforehand of the number of periods, as a means to avoid potential end-game effects.
Each period auditor-client dyads negotiate over a reported amount, which represents asset value.
Negotiations take place over the computer and are anonymous. The negotiation setting is
programmed using the Z-tree software (Fischbacher 2007). We pair participants anonymously in
order to control for potentially confounding effects, including participants appearance, gender,
mannerisms, and prior acquaintance (i.e., whether paired participants know one another). At
14
period end, auditor-client dyads are randomly reassigned. Participants do not receive feedback on
realized asset values throughout the experiment. We suppress feedback in order to reinforce the
fact that each period represents an independent, one-shot negotiation. Furthermore, feedback on
asset value is lagged in naturally-occurring settings. Our experimental context is purposefully
stark, which circumvents possible confounds associated with participants’ beliefs about how they
are supposed to behave in auditor-client settings. We are interested in how auditor-client dyads
determine a reported amount, abstracting away from other potentially noisy factors (see also
Wang and Tuttle 2009; Wang 2010).
At the beginning of each period, neither player knows the actual asset value. We make
this choice to control participants’ bargaining power, such that one player does not have an
advantage over the other (discussed below). The players negotiate over an acceptable reported
asset value, which impacts their payoffs. The actual asset value is determined by drawing from a
uniform distribution that ranges from 50 to 1,000 in increments of 50. So the asset takes one of
20 values, each having a 5 percent chance of realization. The actual asset value is randomly
generated for each auditor-client dyad per period. The instructions include a chart that lists
potential values, the probability of occurrence, and the probability that the amount exceeds the
actual asset value (see also Wang and Tuttle 2009; Wang 2010). The chart is reproduced in Table
1.
[Insert Table 1 about here.]
Negotiations take place as follows. The client begins by proposing a reported amount.
The reported amount is communicated to the auditor, who either accepts or makes a counteroffer.
If the auditor accepts, the period ends. If the auditor makes a counteroffer, the client either
accepts or proposes another amount. The process continues back and forth until an amount is
15
accepted or time expires. As mentioned earlier, periods can last up to one minute. We deem the
allotted time sufficient because the setting is stark and straightforward. Further, the experimental
data suggest that the time is adequate because, on average, we observe more than 13 offers per
period before an agreement is reached, and agreements are achieved 76 percent of the time.
Players’ payoffs are conflicting such that the client prefers larger reported amounts and
the auditor prefers amounts that do not exceed the actual asset value. In addition, players’
payoffs are contingent on whether an agreement is reached: if an impasse occurs, both players’
payoffs are zero.9 If an agreement is reached, the client’s payoff is the reported amount. By
comparison, the auditor’s payoff depends on the relation between the reported amount and the
actual asset value. The auditor’s payoff is 400 if the reported amount is equal to or less than the
actual value. Otherwise, the auditor’s payoff is 400 minus a penalty that is computed as one half
of the difference between the reported amount and the actual asset value.10 In other words, the
auditor incurs a penalty for allowing a reported amount that is too aggressive, where the penalty
is increasing in the level of aggressiveness. The penalty encompasses litigation exposure and
regulatory pressure, which are relevant and present in naturally-occurring auditor-client settings.
The potential for a penalty means that the auditor prefers smaller reported amounts, which is
consistent with the audit ecology. Players’ payoffs are summarized in Figure 1.
[Insert Figure 1 about here.]
The negotiation setting is structured so that both player roles have equal bargaining
power (i.e., neither party has an advantage over the other). The balance of power in auditor-client
negotiations is contingent on situational factors (Gibbins et al. 2001; Brown and Wright 2008;
9 Participants are incentivized to reach an agreement, otherwise they earn nothing for the period. This design choice provides for a stringent test of the effect of narcissism on the length of the negotiation process.10 We choose 400 as a fixed amount considering the following tradeoff. We do not want the amount to be too large as compared to the client’s payoff. But at the same time, we want to avoid instances in which the auditor’s payoff is negative.
16
Salterio 2012). For example, the auditor’s relative power is fortified by a strong audit committee
and authoritative accounting standards (supporting the auditor’s position). By comparison, the
client’s relative power is strengthened by the importance of the client to the auditor and firsthand
knowledge of the negotiation issue. Our setting allows us to study the effect of auditor narcissism
on auditor-client negotiations, holding bargaining power constant and equal across the two
player roles.
After finishing 18 periods, participants complete a post-experiment questionnaire, which
collects demographics as well as other information about participants’ perceptions and
assessments of the experiment.11 Participants’ cash payouts for taking part in the experiment are
determined as follows. For each session, we rank client-participants and auditor-participants
based on their experimental earnings and pay them a fixed amount: the participant who is ranked
first in each role makes the most, the participant who is ranked second makes the next most, and
so forth. The procedure allows us to create similar incentives for participants in different player
roles (Wang and Tuttle 2009; Wang 2010). Further, it eliminates participants’ tendency to
compare own payoffs with their counterpart’s payoffs (Bolton 1991). The procedure is
appropriate because we are not concerned with auditor versus client payoffs, rather we are
interested in the effect of auditor narcissism on the negotiation process and outcome.
4. Results
Table 2 presents descriptive statistics for various measures representing auditor-client
negotiations. The data are collapsed across periods to provide simple summary measures. The
11 The post-experiment questionnaire also includes items to elicit participants’ risk preferences and ability to empathize with others. For risk preferences, we include a menu of paired-lottery choices as developed by Holt and Laury (2002). For empathy, we include seven items designed to measure cognitive perspective-taking ability as developed and validated by Davis (1980). We find that participants NPI scores are positively associated with risk preferences and negatively associated with perspective-taking ability. We perform separate analyses of the experimental data, controlling for participants’ risk preferences and perspective-taking ability, and inferences are unaffected (results not tabulated).
17
descriptive statistics are arranged to allow for visual comparisons of auditor-client pairings,
representing the four experimental groups. We tabulate the percentage of time that an agreement
is reached (AGREE), the average number of offers before agreeing (OFFERS), and the average
agreed amount (VALUE). We use data from all 18 dyads/periods to compute AGREE and data
from dyads/periods with an agreement to compute OFFERS and VALUE.
[Insert Table 2 about here.]
Two measures, AGREE and OFFERS, represent the duration of the negotiation process.
The process is protracted to the extent that AGREE decreases (i.e., impasses occur) and OFFERS
increases. Our first hypothesis suggests that high narcissist auditors (H) are more likely to be
involved in prolonged negotiations than low narcissist auditors (L). The other measure, VALUE,
represents the outcome of the negotiation process, conditioned on agreement. The auditor prefers
agreements to the extent that VALUE decreases, reflecting less aggressive reported amounts.
Our second hypothesis suggests that H auditors are more likely to agree on smaller reported
amounts in negotiations than L auditors. Specific predictions are summarized in Table 3.
[Insert Table 3 about here.]
As discussed earlier, our arguments regarding the effect of auditor narcissism on
negotiations also can apply to client narcissism, particularly when the two parties have equal
bargaining power. Below, we report the effects of auditor narcissism as well as client narcissism
on the negotiation process and outcome.
4.1. Tests of H1
To test H1 we perform analyses focusing on the duration of the negotiation process:
AGREE and OFFERS. For AGREE, the dependent variable is coded as 1 if an agreement is
reached and 0 otherwise. The independent variables include auditor-participant’s narcissism
18
(AuditorN), client-participants’ narcissism (ClientN), an interaction term (AuditorN x ClientN), and
the client’s initial offer (1st−Offer). AuditorN and ClientN are dummy variables, coded as 1 for
High (H) and 0 for Low (L). We control for 1st−Offer because it establishes a starting point and
potentially impacts the negotiation process (Chertkoff and Conley 1967; Galinsky and
Mussweiler 2001). Because AGREE is a categorical variable, we perform a logistic regression.
Recall that in our experiment, players are re-paired to form a different negotiating dyad
each period. Unless otherwise specified, our main analyses reported in this subsection treat each
negotiating dyad as an independent data point. For analyses where the data can be broken down
to the individual player level, we repeat the test using each player as a data point, controlling for
repeated measurement by including the player as a cluster variable, and statistical inferences are
unchanged.
The logistic regression results are summarized in Panel A of Table 4. We find that the
coefficient of AuditorN is negative (β = −0.841) and statistically significant at p < 0.001.12 The
negative coefficient indicates that H auditors are less likely to reach an agreement (more likely to
reach an impasse) than L auditors, which is consistent with H1. We also find that the coefficient
of ClientN is negative (β = −0.479) and statistically significant at p = 0.007, paralleling the effect
of AuditorN. Notably, the interaction term (AuditorN x ClientN) is not significant (p > 0.40).
Hence, our results suggest that an agreement is less likely when at least one of the negotiating
parties (auditor or client) is a high narcissist. The cell means, summarized in Table 2, indicate
that the frequency of agreement is highest when both negotiating parties are low narcissists, with
the percentages decreasing when the auditor and/or client is a high narcissist. Finally, the
coefficient of 1st−Offer is negative (β = −0.002) and statistically significant at p < 0.001,
12 The p-values for AuditorN are one-tailed because we have directional predictions (refer to Table 3), whereas the p-values for the other variables are two-tailed.
19
indicating that the failure to reach an agreement is associated with higher initial offers. Recall
that clients prefer higher values, while auditors prefer lower values. Our findings provide
evidence that more extreme starting points (i.e., initial offers that promote clients’ self-interests)
can result in an impasse.
[Insert Table 4 about here.]
Next, we perform an analysis using OFFERS as the dependent variable, representing the
total number of offers preceding an agreement. The independent variables are the same as those
described above, although the observations are restricted to agreements. We perform an ordinary
least squares (OLS) regression, with the results summarized in Panel A of Table 4. The
coefficient of AuditorN is positive (β = 3.975) and statistically significant at p < 0.01, which
indicates more offers with H auditors than L auditors. We maintain that increasing the number of
offers extends the negotiation process. As might be expected, we document a statistically
significant correlation between OFFERS and the time taken to reach an agreement (ρ = 0.50, p =
0.0001).13 Our findings suggest that negotiations are more likely to be drawn out with H auditors
as compared to L auditors, providing further support for H1.
The OLS findings also indicate that the coefficient of ClientN is positive (β = 2.658) and
statistically significant at p = 0.008, again paralleling the effect of AuditorN. Further, 1st−Offer
and the interaction term are both statistically significant at p < 0.025. For 1st–Offer, the
coefficient is positive (β = 0.013), indicating that higher first offers prolong the negotiation
process (i.e., more offers before reaching an agreement). To interpret the statistically significant
interaction term, we partition the data by client narcissism and examine the simple effects.
13 We also perform an analysis in which the dependent variable is the time taken per period: that is, the number of seconds to reach an agreement, where the value is 60 seconds if an agreement is not reached. Again, the results mirror those reported in Table 4 (using OFFERS as the dependent measure).
20
The simple effects results are summarized in Panel B of Table 4. In both analyses, the
coefficient of AuditorN is positive and statistically significant (p < 0.05), though the effect is
more pronounced with L clients than H clients. Scrutinizing the data by experimental condition,
we note that the difference in OFFERS is greater when comparisons involve the LL dyad (i.e.,
when both parties are low narcissists). With the LL dyad, the number of offers preceding an
agreement is noticeably lower than in other dyads.
All in all, our findings support H1.14 Auditor narcissism is associated with the likelihood
of reaching an impasse in negotiations. In addition, auditor narcissism is associated with the
length of the negotiation process, and this relationship is more discernible when auditors
negotiate with clients who are low, rather than high, on the narcissistic personality dimension.
4.2. Tests of H2
To test H2, we examine the agreed amount, which represents the negotiated outcome. We
perform an OLS regression using VALUE as the dependent measure. The independent variables
are the same as those described earlier. The OLS findings, summarized in Panel A of Table 4
(rightmost column), indicate that the coefficient of AuditorN is negative (β = −28.837) and
statistically significant at p < 0.01. The negative coefficient indicates that the agreed amount is
smaller (less aggressive) with H auditors than L auditors, supporting H2. We also find that the
coefficient of ClientN is positive (β = 42.978) and statistically significant (p < 0.01). In this case,
the agreed amount is larger (more aggressive) with H clients than L clients.
An inspection of the cell means, summarized in Table 2, sheds further light on the effects
of auditor and client narcissism on the negotiated outcome. When the dyad includes a high and
low narcissist (HL and LH dyads), the agreed amount is more in line with the high narcissist’s 14 We repeat our analysis of H1 using the average per participant (collapsed over the 18 periods), and the results are similar to those presented in Table 4, with one exception. The simple effects results change slightly (refer to Panel B of Table 4). Specifically, H auditors are associated with a higher number of offers when dealing with L clients (p < 0.01), but not H clients (p > 0.10).
21
preferences: that is, the agreed amount is lower with H auditors and higher with H clients. When
the dyad includes two high narcissists (HH dyad), the agreed amount falls between that of the
HL and LH dyads. In this case, players’ narcissistic behavior appears to offset one another. The
auditor’s narcissism works to dampen the reported amount, and the client’s narcissism works to
increase the amount. Lastly, when the dyad includes two low narcissists (LL dyad), the agreed
amount is similar to that of the HH dyad. Both dyads (LL and HH) end up agreeing on a reported
amount that seems to reflect a middle ground; however, the negotiation process differs markedly.
More specifically, the LL dyad reaches more agreements and makes fewer offers than the HH
dyad (refer to Table 2).
The OLS results also indicate that the coefficient of 1st−Offer is positive (β = 0.268) and
statistically significant (p < 0.01). Higher initial offers are associated with higher agreed
amounts, which is consistent with prior findings in the negotiation literature (Chertkoff and
Conley 1967; Galinsky and Mussweiler 2001). As discussed earlier though, higher initial offers
can lead to protracted negotiations: that is, fewer agreements and more offers before reaching an
agreement.
Overall, our findings are consistent with H2.15 Auditors who score high, rather than low,
on the narcissistic personality inventory are more likely to agree on an amount that is smaller
(less aggressive) and, thus, more in line with their preferences. In our experimental setting,
auditors’ prefer a smaller reported amount because it reduces the chance of incurring a penalty
(e.g., costs associated with legal liability and dealing with regulators). We also document that
client narcissism is associated with the agreed amount, but in the opposite direction. That is,
clients who score high, rather than low, on the narcissistic personality inventory are more likely
15 We repeat our analysis of H2 using the average per participant (collapsed over the 18 periods) and inferences are unaffected.
22
to agree on an amount that is higher (more aggressive), which is more in line with their
preferences.16 In sum, participants’ narcissism has a very marked effect on the negotiation
outcome.
5. Supplementary Data and Results
To ensure internal validity, we use a stylized negotiation setting in our experiment. The
setting abstracts away from professional/ethical values (e.g., conforming to a code of
conduct/ethics), social costs (auditor-client relationships), and institutional knowledge (e.g.,
experience/expertise surrounding the application of GAAP). The setting also does not include
certain features common to negotiation settings, such as persuasive communication, threats,
opportunities for reputation building, and ongoing relationships. We believe that our design
choices are necessary to ensure a clean and parsimonious test of our hypotheses. To shed light on
the applicability of our findings to naturally-occurring auditor-client settings, we collect archival
data and perform additional analyses.
We collect data related to issues that correspond to our hypotheses. Our first hypothesis
suggests that auditor narcissism is associated with protracted negotiations. For this hypothesis we
focus on audit delay, defined as the length of time from fiscal year end to the audit report date
(Ashton et al. 1987). Salterio (2012) points out that auditor-client negotiations likely play a role
in audit delay. For example, disagreements over the application of GAAP, particularly late in the
process, hamper the completion of an engagement.17 We use audit delay as a proxy for the time
associated with auditor-client negotiations.
16 We repeat all analyses using participants’ raw NPI scores to proxy for narcissism. The results are similar to those presented in Panel A of Table 4 with one exception: the interaction between auditor-client narcissism on OFFERS loses it statistical significance. This difference does not affect the results of our hypotheses tests. Importantly, the main effect of auditor narcissism is robust. 17 Relatedly, McCracken et al. (2008) find that some CFOs do not involve the auditor in difficult accounting choices and, further, that such CFOs tend to be strongly committed to their positions. Under these circumstances, contentious accounting issues usually are not dealt with until after year-end, potentially extending completion of the audit.
23
Our second hypothesis suggests that auditor narcissism results in reported amounts that
are less aggressive. We use absolute and signed discretionary accruals as a proxy for negotiated
outcomes. We assume that the auditor prefers smaller amounts, as it constrains the client’s
latitude in financial reporting. Prior findings suggest that clients’ discretionary accruals are
associated with auditors’ litigation risk: that is, higher discretionary accruals increase the
likelihood of auditor liability (e.g., Boone et al. 2011).
5.1. Auditor Narcissism
We collect archival data on Chinese public companies, which provides a natural basis of
comparison with our experimental data, also collected in China. Importantly, the auditor’s report
in China is signed by two auditors, a review auditor and an engagement auditor. The review
auditor typically is the lead auditor, having ultimate authority and responsibility (Chinese
Ministry of Finance 2001; Liu and Chi 2014). The engagement auditor, on the other hand,
directs the audit (Wang et al. 2015). The review auditor usually is more senior and experienced
than the engagement auditor (Gul et al. 2013; Lennox et al. 2014). The signature of the review
auditor is placed above that of the engagement auditor, which allows us to identify the two roles
(Lennox et al. 2014; Wang et al. 2015).18 We measure the size of each signature following prior
research (Zweigenhaft 1977; Ham et al. 2014; Ham et al. 2015). Specifically, we construct a
rectangle around each signature, with each side touching the most extreme point of the signature.
Then, we standardize the area, dividing by the number of characters in the signature. This
approach yields the area-per-letter, which helps to control for differences in name lengths.
Prior research provides evidence of a positive association between signature size and
various facets of personality linked to narcissism, including awareness of status (Zweigenhaft
18 We interviewed audit partners at several major accounting firms in China and confirmed that the review auditor has more responsibility than the engagement auditor and that the review auditor’s signature appears above that of the engagement auditor.
24
1970; Zweigenhaft and Marlow 1973), self-esteem (Zweigenhaft 1977), and dominance
(Jorgenson 1977). Previous accounting research suggests a positive relation between signature
size and narcissism: that is, narcissists have larger signatures than others (Ham et al. 2014; Ham
et al. 2015).
We take additional steps to validate signature size as a proxy for narcissism. We recruit
52 undergraduate business students at a major Chinese university and administer the 40-item NPI
(Raskin and Terry 1988). We also collect students’ signatures. The standardized signature size is
positively correlated with students’ NPI score (ρ = 0.27, p = 0.053, two-tailed). The correlation
is similar to that reported in Ham et al. (2015), and offers evidence of a linkage between
signature size and narcissism.
5.2. Sample
Our sample includes 429 Chinese A-share listed firm-year observations from 2007 –
2013, representing 281 firms in 53 industries.19 The shares are listed on the Shanghai and
Shenzhen Stock Exchanges, the two major exchanges in China. We begin our sample period in
2007, which corresponds with a significant change in China’s financial reporting regime. In
2006, the Ministry of Finance of China issued new accounting standards for business enterprises,
effective January 1, 2007.
5.3. Audit Delay
First, we estimate the following OLS regression model to examine the effect of auditor
narcissism on audit delay.
19 We begin with an initial sample of 535 firm-year observations with available signatures. We require both auditors signatures (review and engagement auditors) as well as the CFO’s signature (taken from the annual report). We include the CFO’s signature to control for client narcissism. We exclude 71 observations with signatures that are illegible, and 35 observations that have financial information missing from the China Stock Market & Accounting Research database (as explained below financial information is needed for control variables). In addition, companies issuing only B-shares are excluded from our sample. The B-share market differs from that of the A-share market in terms of pricing, liquidity, and accounting and auditing requirements (Chen et al. 2010; Firth et al. 2012).
25
LnLag¿=α 0+α1 SIZEAUD 1 j+α2 SIZEAUD 2 j+α 3 SIZECFO j+α4 BIG¿+α 5 ¿¿¿+α 6GROWTHR¿+α7 LEV ¿+α 8QUICK ¿+α 9 INVREC¿+α 10 LOSS¿+α11 FOREIGN ¿+α 12 FIRMAGE¿+α 13 BOARDIND¿+α14 BOARDMEET ¿+α15TOPSH ¿+α16−21YearIndicators+ IndustryIndicators+Error .¿
The dependent variable, LnLagit, is the natural log of the number of days between a company’s
fiscal year-end and the audit report date. The natural log transformation typically is applied when
the variable is positively skewed, as with audit delay. For comparative purposes, we also perform
the analysis using SLagit, defined as the number of days between fiscal year-end and the audit
report date scaled by 360 days. We point out that all publicly listed companies in China have a
December 31st fiscal year-end, and are required to disclose an annual report no later than April
30th.
The primary independent variables of interest are SIZEAUD1 and SIZEAUD2. These
variables proxy for auditor narcissism based on the signature size of the review auditor and
engagement auditor, respectively. We include SIZECFO to proxy for the client’s narcissism
based on the CFO’s signature size. The CFO usually takes part in auditor-client negotiations and,
thus, can impact the duration of the process. Gibbins et al. (2001) report that CFOs are involved
in 96 percent of the negotiation examples identified by audit partners. In another study, Gibbins
et al. (2007) report that CFOs are involved 81 percent of the time. In addition, we interviewed
several audit partners in China who confirmed that the CFO typically is involved in auditor-
client negotiations.
We include numerous control variables in our analysis based on prior research (e.g.,
Ashton et al. 1989; Ettredge et al. 2006; Whitworth and Lambert 2014). We expect auditor and
client characteristics to impact the negotiation process and, in turn, audit delay. The variables are
defined in Table 5 and summarized below. For the auditor, we include an indicator variable for
Big 4 auditor (BIG). For client characteristics, we include the natural log of total assets (SIZE),
growth in revenues (GROWTHR), the leverage ratio (LEV), the quick ratio (QUICK), inventory
26
and receivables divided by total assets (INVREC), negative earnings (LOSS), foreign operations
(FOREIGN), the natural log of firm age (FIRMAGE), the percentage of independent board
directors (BOARDIND), the number of board meetings (BOARDMEET), and the percentage of
shares held by the top ten shareholders (TOPSH).20 In addition, we control for year
(YearIndicators) and industry (IndustryIndicators) fixed effects, and we adjust standard errors
for clustering at the firm level. All continuous variables also are winsorized at the top and bottom
one percent to avoid the impact of outliers.
[Insert Table 5 about here.]
Table 6 reports descriptive statistics for audit delay (LnLag and SLag), auditor and client
narcissism (SIZEAUD1, SIZEAUD2, and SIZECFO), and other control variables. We observe
that audit delay is, on average, 91.4 days. Inspection of Table 6 indicates that SIZEAUD1 is
larger than SIZEAUD2, suggesting that the review auditor’s narcissism exceeds that of the
engagement auditor. Moreover, SIZEAUD1 and SIZEAUD2 are larger than SIZECFO. This
difference, however, is unlikely to be related to narcissism because, in general, more space is
allotted for signatures in the auditor’s report than in the annual report.
[Insert Table 6 about here.]
The left-hand side of Table 7 presents the results of two separate models to examine audit
delay: one with LnLag as the dependent measure and the other with SLag. The results of the two
models are similar. The coefficient of SIZEAUD1 is positive and statistically significant at p <
0.05 using LnLag as the dependent measure and p < 0.10 using SLag as the dependent measure.
By comparison, SIZEAUD2 is insignificant (p > 0.10). The findings suggest that audit delay is
positively associated with the review auditor’s narcissism, but not that of the engagement 20 Ettredge et al. (2006) document that companies with material weaknesses have longer audit delays than other companies, using the auditor’s opinion on internal control to identify companies with material weaknesses. We do not include this variable in our analyses because, in China, audits of internal control were not mandated until 2012 for state-owned entities and 2013 for non-state-owned entities (Lei, Wu, and Sun 2013).
27
auditor. This result is not surprising because the review auditor is more senior than the
engagement auditor, and, further, becomes more involved at a later stage of the audit. Thus,
substantial issues raised by the review auditor have the potential to delay the audit’s completion.
Our findings for the review auditor are consistent with H1.21
[Insert Table 7 about here.]
The OLS results also indicate that the coefficient of SIZECFO is positive and statistically
significant (p < 0.01), whereas the interaction terms are not significant. The positive coefficient
of SIZECFO indicates that the CFO’s narcissism contributes to audit delay. In addition, we find
that audit delay is negatively associated with having a Big 4 auditor (p < 0.01) and the
percentage of shares held by large shareholders (p < 0.10), and positively associated with client
size (p < 0.01) and clients’ growth in revenues (p < 0.05).
5.4. Discretionary Accruals
Next, we estimate the following OLS regression model to examine the effect of auditors’
narcissism on the outcome of auditor-client negotiations.
DA¿=α 0+α 1 SIZEAUD1 j+α 2 SIZEAUD2 j+α 3 SIZECFO j+α4 BIG¿+α5 ¿¿¿+α 6GROWTHR¿+α7 LEV ¿+α 8QUICK ¿+α 9 INVREC ¿+α 10 LOSS¿+α11 FOREIGN ¿+α 12 FIRMAGE¿+α 13 BOARDIND¿+α14 BOARDMEET ¿+α15TOPSH ¿+α 16−21 YearIndicators+ IndustryIndicators+Error .¿
The left-hand-side variable, DA, is the discretionary accruals estimated using the modified Jones
model with firm profitability included for each year-industry group, conditional on having at
least five firms in the group (Kothari et al. 2005). We run the model three times, using three
different dependent measures: absolute discretionary accruals (|DA|), positive discretionary
accruals (PosDA), and negative discretionary accruals (NegDA). The primary independent
variables of interest, again, proxy for auditor narcissism, SIZEAUD1 and SIZEAUD2. As before,
21 We re-perform the OLS regressions including indicator variables for whether the client’s audit opinion was modified and whether an auditor change occurred. We also include a variable to control for client importance, defined as the client’s audit fees scaled by the aggregate audit fees charged by the audit firm. None of the added variables are statistically significant, and the results presented in Table 7 are unaffected.
28
we include SIZECFO to control for the client’s narcissism and other control variables following
prior literature (e.g., Ashbaugh-Skaife et al. 2008). All variables are defined previously in the
model for audit delay.
The right-hand side of Table 7 presents the results of three separate OLS regression
models. Looking at the results using |DA| as the dependent measure, we find that SIZEAUD1 is
negative and statistically significant at p < 0.01, whereas SIZEAUD2 is insignificant (p > 0.10).
The absolute value of clients’ discretionary accruals is negatively associated with the review
auditor’s narcissism, but not the engagement auditor’s narcissism. Again, the review auditor’s
narcissism plays a significant role. We contend that smaller absolute discretionary accruals
improve financial reporting quality and, in turn, decrease the auditor’s potential exposure. The
findings for the review auditor are consistent with H2.
We also document that the coefficient of SIZECFO is positive and statistically significant
(p < 0.10), whereas the interaction terms are not significant. The positive coefficient of
SIZECFO provides evidence that the CFO’s narcissism leads to larger absolute discretionary
accruals. Further, absolute discretionary accruals are negatively associated with having a Big N
auditor (p < 0.01) and clients’ size (p < 0.05), and positively associated with the percentage of
shares held by large shareholders (p < 0.01) and clients’ age (p < 0.10).
The results using PosDA as the dependent measure are similar to those discussed above,
though significance levels are reduced. We find that SIZEAUD1 is negative and statistically
significant at p < 0.05. The review auditor’s narcissism is negatively associated with clients’
positive discretionary accruals. So the review auditor’s narcissism leads to less aggressive
financial reporting, consistent with H2. The results using NegDA as the dependent measure
indicate that SIZEAUD1 is positive and statistically significant at p < 0.10. In this case, the
29
review auditor’s narcissism is positively associated with clients’ negative discretionary accruals,
which leads to more conservative financial reporting.
Finally, we repeat the analyses using alternative measures of clients’ discretionary
accruals. Specifically, we re-compute DA to allow for asymmetry in the timing of gain and loss
recognition (see also Ball and Shivakumar 2006; Carver et al. 2011; Stanley et al. 2015). We also
compute DA following Francis et al. (2005). The results (not tabulated), using the alternative
measures, are consistent with those presented in Table 7. In every case, SIZEAUD1 is significant
at p < 0.05, and the findings suggest that the review auditor’s narcissism is associated with less
aggressive financial reporting.
6. Concluding Remarks
This paper reports the results of an experiment designed to examine the effect of auditor
narcissism on auditor-client negotiations. The benefits of using an experimental approach are
that we can observe the negotiation process and control for the influence of potentially
confounding, extraneous factors. Our experimental findings indicate that auditor narcissism is
associated with more impasses and longer negotiations. Notably, when agreements are reached,
auditor narcissism leads to outcomes that are indicative of less aggressive reporting choices. To
shed light on the generalizability of our findings to a naturally-occurring setting, we collect
archival data and perform supplemental analyses. We provide evidence that auditor narcissism is
positively associated with audit delay, taking longer to issue audited financial statements. We
also find that auditor narcissism is associated with clients’ making less aggressive reporting
choices. We document that auditor narcissism is linked to clients having smaller absolute and
positive discretionary accruals and larger negative discretionary accruals. The supplemental
30
results are consistent with our experimental findings, underscoring the role of auditor narcissism
in negotiations.
Our experimental findings suggest that narcissistic auditors-participants are more likely
to hold their ground in negotiations, allowing them to reach agreements that reflect less
aggressive reported amounts. This result suggests that auditor narcissism is potentially
beneficial, as a means to promote audit quality, though it might undermine auditor-client
relations. Accounting firms make employee selection decisions, including promotions, based on
employees’ attributes (Campbell 2012). Such attributes oftentimes reflect personality dimensions
(e.g., empathy, confidence, propensity for risk, etc.). Indeed, workplace personality testing is
estimated to be a $500 million-a-year business, with annual growth rates of 10 to 15 percent
(Weber and Dwoskin 2014). We suggest that auditor narcissism is an important attribute that
merits attention: accounting firms may need to carefully consider its benefits and costs when
making recruiting and promotion decisions in order to maximize overall audit effectiveness.
Our study is subject to certain limitations, which provide potential avenues for future
research. First, our negotiation setting is one in which the auditor and client have equal
bargaining power. We acknowledge that situational factors may shift the balance of power
toward one party (Gibbins et al. 2001; Brown and Wright 2008; Salterio 2012). Yet, we do not
have any obvious reason to believe that our results for auditor narcissism would not apply to
settings in which bargaining power is unequal. The issue, however, remains an empirical
question. Second, our negotiation setting excludes interpersonal interactions, which are crucial in
an audit context. To the extent that narcissistic auditors hold firm in negotiations, such behavior
potentially may create ill will with the client. Future interactions may be tainted, and auditor-
client relations may deteriorate over time. The auditor’s demeanor and tactics may not sit well
31
with the client and, eventually, the auditor-client relationship may be severed. We encourage
future study to explore this important issue.
Relatedly, in our experimental setting auditor-client dyads interact in a series of one-shot
negotiations, where the dyads are re-paired each period. Thus, negotiation outcomes in one
period do not affect outcomes, including participants’ payoffs, in future periods. Yet, in a
naturally-occurring setting, the auditor and client are involved in a multi-period relationship.
Along these lines, McCracken et al. (2008) observe that, in auditor-client interactions, the audit
partner assumes the role of relationship manager, responsible for keeping the client happy. The
researchers collect field data and provide evidence that accounting firms manage the assignment
of audit partners based on client (CFO) preferences. Poor relationships, irrespective of the
underlying reason, provide a basis to remove partners from engagements. For our purposes, the
client might apply pressure to have a narcissistic audit partner taken off the engagement. We
leave this issue for future study.
We use Chinese participants in our experiment. Chinese and Westerners could approach
negotiations differently, primarily because the former emphasizes collectivism whereas the latter
emphasizes individualism (e.g., Triandis 1986; Lee 2003). Lee et al. (2012) suggest that Chinese
are generally concerned with relational harmony, mutual trust, and reciprocity in negotiations.
However, quite apart from culture, we expect high narcissists to behave differently than low
narcissists. That is, we expect the effects of narcissism on the negotiation process and outcomes
to hold across cultures.
Lastly, the applicability of our experimental findings to naturally-occurring settings rely
on the assumption that our parameters capture critical features of the audit ecology. Though our
setting is abstract, we exercise care in determining participants’ relative payoffs, contingent on
32
their actions. Further, we take comfort that the archival findings mirror those obtained in the
laboratory, providing evidence of external validity. Despite the limitations, our findings provide
important insights into the effect of auditor narcissism on auditor-client interactions.
33
Figure 1
Players’ Payoffs per Period
Panel A: Client Payoff
Panel B: Auditor Payoff
Notes: The figure depicts participants’ payoffs in our experiment, contingent on negotiated outcomes.
Does the period end with an agreement?
(i.e., is there an accepted reported
value?)
Is the reported value higher than the actual value?
Auditor’s payoff =
400 – 50% × (reported value – actual value)
Auditor’s payoff = 400
Auditor’s payoff = Zero
Yes
Yes
No
No
Does the period end with an agreement?
(i.e., is there an accepted value?)
Client’s payoff = Accepted value
Client’s payoff = Zero
Yes
No
34
Table 1
Relation between Potential and Actual Asset Value
Value
Probability of being the actual value
Probability of being higher than the actual value
1000 5% 95%950 5% 90%900 5% 85%850 5% 80%800 5% 75%750 5% 70%700 5% 65%650 5% 60%600 5% 55%550 5% 50%500 5% 45%450 5% 40%400 5% 35%350 5% 30%300 5% 25%250 5% 20%200 5% 15%150 5% 10%100 5% 5%50 5% 0%
Notes: In our experiment, the client chooses a reported amount. The first column includes potential amounts that can be reported. The second column indicates that the probability that each potential amount corresponds with the actual asset value is 5 percent. The third column shows the probability that the reported amount is greater than the actual asset value.
35
Table 2
Descriptive Statistics
Auditor Narcissism
Client NarcissismLow High
LowAGREE 0.86 (n=432)OFFERS 10.99 (n=370)VALUE 472.70 (n=370)
AGREE 0.77 (n=486)OFFERS 14.34 (n=374)VALUE 529.95 (n=374)
HighAGREE 0.74 (n=396)OFFERS 14.14 (n=294)VALUE 426.70 (n=294)
AGREE 0.67 (n=450)OFFERS 14.38 (n=301)VALUE 464.45 (n=301)
Notes: Auditor- and client-participants are categorized as high (H) or low (L) based on their responses to the narcissistic personality inventory. We establish auditor-client dyads such that narcissism pairings are HH, HL, LH, and LL, where the first letter represents the auditor and the second represents the client. The cell entries include averages for various measures, with the number of observations shown parenthetically. The measures are defined as follows: AGREE = the percentage of time that auditor-client dyads reach an agreement; OFFERS = the average number of offers before an agreement is reached; andVALUE = the average reported amount that is agreed upon.
36
Table 3
Experimental Predictions
Hypothesis Measure Auditor Narcissism
Predicted Sign
H1: Auditor narcissism is positively associated with the length of the negotiation process.
AGREEt H < L −OFFERSt H > L +
H2: Auditor narcissism is positively associated with negotiation outcomes that are consistent with auditors’ preferences.
VALUEt H < L −
Notes: The variables of interest are defined as follows: AGREE = the percentage of time that auditor-client dyads reach an agreement; OFFERS = the average number of offers before an agreement is reached; andVALUE = the average reported amount that is agreed upon.For auditor narcissism, H refers to auditor-participants who have High NPI scores and L refers to those who have Low NPI scores. The predicted sign represents the predicted difference between H and L auditors.
37
Table 4. Experimental Results
Panel A: Analyses of the Negotiation Process and Outcome
Independent Variables Dependent VariableAGREEt OFFERSt VALUEt
AuditorN −0.841***
(−4.60) 3.975***
(4.80)−28.837***
(−3.17)
ClientN −0.479**
(−2.71) 2.658***
(3.06) 42.978***
(4.75)
AuditorN x ClientN 0.183 (0.78)
−2.794**
(−2.26)−13.131(−1.00)
1st−Offer −0.002***
(−5.15) 0.013***
(6.51) 0.268***
(11.62)
Constant 3.188***
(10.13) 1.534 (1.00)
277.189***
(15.93)Observations 1,764 1,339 1,339
Panel B: Simple Effects (DV=OFFERSt)
Independent Variables Clients’ NarcissismLow High
AuditorN 3.634***
(4.41) 1.556**
(1.70)
1st−Offer 0.008***
(3.06) 0.017***
(5.86)
Constant 5.410***
(2.88) 0.842 (0.38)
Observations 664 675
Notes: The cell entries include estimated coefficients (test statistics). For AGREE, we perform a logistic regression, controlling for period, so the values shown parenthetically are z-statistics. For the other dependent variables, we perform an ordinary least squares regression, so the values shown parenthetically are t-statistics. One, two, and three asterisks (*, **, and ***) indicate statistical significance at the 10 percent, 5 percent, and 1 percent levels, respectively. The p-values are one-tailed for AuditorN because we have directional predictions and two-tailed for the other variables.
The independent variables are defined as follows. AuditorN and ClientN are auditors’ and clients’ narcissism, respectively, coded as 1 for High and 0 for Low; and 1st−Offer is the client’s first offer. For H1, we test whether auditor narcissism affects the negotiation process using two dependent variables, AGREEt and OFFERSt. AGREEt is defined as 1 if an agreement is reached and 0 otherwise; and OFFERSt represents the number of offers preceding an agreement. In Panel B, we partition the data based on clients’ narcissism (Low and High) and present the simple effects of the significant interaction term revealed in Panel A (for OFFERSt). For H2, we test whether auditor narcissism affects negotiated outcome using VALUEt as the dependent variable. VALUEt is defined as the agreed upon amount.
38
Table 5. Variables in the Supplemental Analyses
Variable Definition
LnLagThe natural log of the number of days between fiscal year-end and the audit report date;
SLagThe number of days between fiscal year-end and the audit report date scaled by 360 days;
DADiscretionary accruals estimated based on modified Jones model with firm profitability included (Kothari et al. 2005) for each year-industry group, conditional on having at least five firms in the group;
SIZEAUD1 Standardized signature size of the reviewing auditor;
SIZEAUD2 Standardized signature size of the engagement auditor;
SIZECFO Standardized signature size of chief financial officer (CFO);
BIG 1 if a firm is audited by a Big 4 audit firm, 0 otherwise;
SIZE The natural log of the firm’s total assets;
GROWTHR Revenues in year t minus revenues in year t-1 divided by revenues in year t-1;
LEV Total debt divided by total assets;
QUICK Current assets minus inventory divided by current liabilities;
INVREC Inventory plus accounts receivable divided by total assets;
LOSS 1 if a firm reports negative net earnings and 0 otherwise;
FOREIGN 1 if a firm is involved in foreign operation and 0 otherwise;
FIRMAGE The natural log of the firm’s age.
BOARDIND Percentage of independent directors at the end of year t;
BOARDMEET Number of board meetings in year t;
TOPSH Percentage of shares held by the top ten shareholders at the end of year t;
Notes: The first two variables, LNLAG and SLAG are dependent measures used to examine audit delay (H1). The third variable, DA, is used to compute dependent measures to examine clients’ discretionary accruals. For DA, we examine the absolute value of discretionary accruals (|DA|) as well as positive and negative discretionary accruals (PosDA and NegDA). The remainder are independent variables in the supplemental analyses.
39
Table 6. Descriptive Statistics: Supplemental DataVariable Mean Std. Dev. 25th Percentile 50th Percentile 75th Percentile
LnLag 4.483 0.257 4.382 4.466 4.682
SLag
|DA| 0.050 0.047 0.016 0.037 0.067
PosDA 0.051 0.049 0.015 0.036 0.068
NegDA
SIZEAUD1 0.279 0.168 .0178 0.232 0.319
SIZEAUD2 0.234 .0115 0.152 0.207 0.289
SIZECFO 0.085 0.050 0.051 0.073 0.108
SIZE 22.334 1.260 21.499 22.210 23.162
GROWTHR 0.148 0.499 -0.012 0.082 0.206
LEV 0.550 0.190 0.434 0.563 0.684
QUICK 0.985 0.846 0.492 0.756 1.207
INVREC 0.284 0.207 0.113 0.241 0.426
FIRMAGE 2.624 0.299 2.398 2.639 2.890
BOARDIND 0.365 0.050 0.333 0.333 0.385
BOARDMEET 2.180 0.358 1.946 2.197 2.398
TOPSH 0.524 0.163 0.398 0.522 0.650
Categorical Variables
BIG 0.061
LOSS 0.089
FOREIGN 0.289
Observations 429
Notes: The variables are defined in Table 5.
40
Table 7. Results: Supplemental ResultsIndependentVariables
Audit Delay Clients’ Discretionary AccrualsLnLag SLag |DA| PosDA NegDA
SIZEAUD1 0.137** 0.023* -0.044*** -0.074** 0.034*(2.03) (1.48) (-3.00) (-1.71) (1.49)
SIZEAUD2 0.048 0.017 0.005 0.037 0.001(0.52) (0.77) (0.20) (0.99) (0.02)
SIZECFO 0.762*** 0.181*** 0.138* 0.153 -0.171(3.15) (3.15) (1.83) (1.32) (-1.51)
SIZEAUD1*SIZECFO -1.807 -0.385 0.343 0.273 -0.31(-0.88) (-0.82) (0.78) (0.24) (-0.42)
SIZEAUD2*SIZECFO -2.574 -0.61 -0.066 2.628* 1.577 (-0.83) (-0.83) (-0.09) (1.79) (1.54)BIG -0.115** -0.034*** -0.022*** -0.031* 0.008
(-2.21) (-3.28) (-2.74) (-1.85) (0.57)SIZE 0.044*** 0.009*** -0.005** -0.007 0.010**
(3.68) (3.51) (-2.10) (-1.24) (2.27)GROWTH 0.047** 0.011** 0.007 0.011 0.004
(2.30) (2.29) (1.13) (1.33) (0.69)LEV -0.035 -0.002 0.005 -0.035 -0.033
(-0.28) (-0.08) (0.24) (-0.94) (-1.00)QUICK 0.007 0.001 0.004 0.009 0.002
(0.29) (0.25) (0.68) (0.92) (0.28)INVREC 0.009 -0.002 0.007 0.005 -0.005
(0.10) (-0.10) (0.38) (0.19) (-0.18)LOSS 0.055 0.015 0.014 0.011 -0.001
(1.29) (1.62) (1.60) (0.60) (-0.10)FOREIGN 0.043 0.009 0.001 0.007 -0.002
(1.50) (1.50) (0.24) (0.55) (-0.28)FIRMAGE -0.006 0.000 0.020* 0.022 -0.024
(-0.13) (0.01) (1.83) (1.08) (-1.60)BOARDIND -0.304 -0.073 0.046 0.088 -0.150**
(-1.19) (-1.39) (0.96) (1.00) (-2.36)BOARDMEET -0.059 -0.012 0.002 0.000 -0.01
(-1.58) (-1.56) (0.26) (-0.01) (-0.98)TOPSH -0.209** -0.039* 0.054*** 0.066** -0.043
(-2.14) (-1.96) (2.74) (2.02) (-1.47)Constant 3.736*** 0.078 0.084 0.015 -0.153*
(13.72) (1.29) (1.33) (0.12) (-1.69)Adjusted R-squared 0.232 0.182 0.103 0.172 0.084Observations 429 429 429 189 240
Notes: The left-hand side of the table (columns 1 and 2) reports the OLS regression results using audit delay (LnLag and SLag) as the dependent variable. The right-hand side of the table (columns 3 – 5) reports the OLS regression results using discretionary accruals (|DA|, PosDA, and NegDA) as the dependent variable. The variables are defined in Table 5. Though not shown, we also include year and industry fixed effects. In addition, standard errors are adjusted for clustering at the firm level. The table entries represent the estimated coefficient (t-statistic). One, two, and three asterisks *, **, and ***) indicate statistical significance at the 10 percent, 5 percent, and 1 percent levels, respectively. The p-value are two-tailed, excepting SIZEAUD1 and SIZEAUD2. The p-values are one-tailed for the auditor narcissism variables as we have directional predictions.
41
References
Akers, M.D., D.E. Giacomino, and J. Weber. 2014. Narcissism in public accounting firms. Accounting and Finance Research 3(3): 170–178.
American Psychiatric Association. 2013. Diagnostic and Statistical Manual of Mental Disorders, 5th edition. Arlington, VA: American Psychiatric Association.
Antle, R., and B. Nalebuff. 1991. Conservatism and auditor-client negotiations. Journal of Accounting Research 29(Supplement): 31–54.
Ashbaugh-Skaife, H., D.W. Collins, W.R. Kinney, and R. LaFond. 2008. The effect of SOX internal control deficiencies on their remediation on audit quality. The Accounting Review 83(1): 217–250.
Ashton, R.H., J.J. Willingham, and R.K. Elliott. 1987. An empirical analysis of audit delay. Journal of Accounting Research 25(2): 275-–292.
Ashton, R.H., P.R. Graul, and J.D. Newton. 1989. Audit delay the timeliness of corporate reporting. Contemporary Accounting Research 5(2): 657—673.
Babcock, L.H., and G. Loewenstein. 2005. Explaining bargaining impasse: The role of self-serving bias. In M.H. Bazerman (Ed.) Negotiation, Decision Making and Conflict Management, Vol. 1 – 3 (231–248). Northampton, MA: Edward Elgar Publishing.
Ball, R., and L. Shivakumar. 2006. The role of accruals in asymmetrically timely gain and loss recognition. Journal of Accounting Research 44(2): 207–242.
Bame-Aldred, C.W., and T. Kida. 2007. A comparison of auditor and client initial positions and tactics. Accounting, Organizations and Society 32: 497–511.
Barry, B., and R.A. Friedman. 1998. Bargainer characteristics in distributive and integrative negotiation. Journal of Personality and Social Psychology 74(2): 345–359.
Bauer, T.D. 2015. The effects of client identity strength and professional identity salience on auditor judgments. The Accounting Review 90(1): 95–114.
Baumeister, R.F., B.J. Bushman, and W.K. Campbell. 2000. Self-esteem, narcissism, and aggression: Does violence result from low self-esteem or from threatened egotism? Current Directions in Psychological Science 9(1): 26–29.
Beattie, V., S. Fearnley, and R. Brandt. 2000. Behind the audit report: A descriptive study of discussions and negotiations between auditors and directors. International Journal of Auditing 4(2): 177–202.
42
Bhaskar, L., T. Majors, and A.M. Vitalis. 2015. Are “good” auditors impacted more by ego depletion? Threats to valued auditor attributes. Working paper, Indiana University (December).
Birnberg, J.G., M.D. Shields, and S.M. Young. 1990. The case for multiple methods in empirical management accounting research (with an illustration from budget setting). Journal of Management Accounting Research 2: 33–66.
Bolton, G. 1991. A comparative model of bargaining: Theory and evidence. The American Economic Review 81(5): 1096–1136.
Boone, J.P., I.K. Khurana, and K.K. Raman. 2011. Litigation risk and abnormal accruals. Auditing: A Journal of Practice & Theory 30(2): 231–256.
Brewer, J., and A. Hunter. 2006. Foundations of Multimethod Research: Synthesizing Styles. Thousand Oaks, CA: Sage Publications.
Brislin, R.W. 1970. Back-Translation for cross-cultural research. Journal of Cross-Cultural Psychology 1(3): 185–216.
Brown, H.L., and A.M. Wright. 2008. Negotiation research in auditing. Accounting Horizons 22(1): 91–109.
Brown, H.L., and K.M. Johnstone. 2009. Resolving disputed financial reporting issues: Effects of auditor negotiation experience and engagement risk on negotiation process and outcome. Auditing: A Journal of Practice & Theory 28(2): 65–92.
Brown-Liburd, H.L., A.M. Wright, and T. Zamora. 2016. Managers’ strategic reporting judgments in audit negotiations. Auditing: A Journal of Practice & Theory 35(1): 47–64.
Calegari, M.J., J.W. Schatzberg, and G.R. Sevcik. 1998. Experimental evidence of differential auditor pricing and reporting strategies. The Accounting Review 73(2): 255–275.
Campbell, D. 2012. Employee selection as a control system. Journal of Accounting Research 50(4): 931–966.
Campbell, W.K., C.P. Bush, A.B. Brunell, and J. Shelton. 2005. Understanding the social cost of narcissism: The case of the tragedy of the commons. Personality and Social Psychology Bulletin 31(10): 1358–1368.
Carnevale, P.J., and D.G. Pruitt. 1992. Negotiation and mediation. Annual Review of Psychology 43: 531–582.
Carroll, L. 1987. A study of narcissism, affiliation, intimacy, and power motives among students in business administration. Psychological Reports 61(2): 355–358.
43
Carver, B.T., C.W. Hollingsworth, and J.D. Stanley. 2011. Recent auditor downgrade activity and changes in clients’ discretionary accruals. Auditing: A Journal of Practice & Theory 30(3): 33–58.
Chatterjee, A., D.C. Hambrick. 2007. It’s all about me: Narcissistic chief executive officers and their effects on company strategy and performance. Administrative Science Quarterly 52(3): 351–386.
Chen, S., S. Y. Sun, and D. Wu. 2010. Client importance, institutional improvements, and audit quality in China: An office and individual auditor level analysis. The Accounting Review 85(1): 127–158.
Chertkoff, J.M., and M. Conley. 1967. Opening offer and frequency of concessions as bargaining strategies. Journal of Personality and Social Psychology 7(2): 181–185.
Chinese Ministry of Finance. 2001. Notice of issues regarding CPAs’ signatures and stamps on audit reports. Chinese Ministry of Finance, File No. [2001] 1035, July 2nd.
Church, B.K., M. Peytcheva, W. Yu, and O.-A. Singtokul. 2015. Perspective taking in auditor-manager interactions: An experimental investigation of auditor behavior. Accounting, Organizations and Society 45: 40–51.
Davis, M.H. 1980. A multidimensional approach to individual differences in empathy. JSAS Catalog of Selected Documents in Psychology 10: 85–104.
De Dreu, C.K.W., L.R. Weingart, and S. Kwon. 2000. Influence of social motives on integrative negotiation: A meta-analytic review and test of two theories. Journal of Personality and Social Psychology 78(5): 889–905.
Dhawan, N., M.E. Kunik, J. Oldham, and J. Coverdale. 2010. Prevalence and treatment of narcissistic personality disorder in the community: A systematic review. Comprehensive Psychiatry 51(4): 333–339.
Dickinson, K.A., and A.L. Pincus. 2003. Interpersonal analysis of grandiose and vulnerable narcissism. Journal of Personality Disorders 17(3): 188–207.
Einhorn, H., and R. Hogarth. 1981. Behavioral decision theory: Processes of judgment and choice. Annual Review of Psychology 32: 53–88.
Emmons, R.A. 1984. Factor analysis and construct validity of the Narcissistic Personality Inventory. Journal of Personality Assessment 48(3): 291–300.
Emmons, R.A. 1987. Narcissism: Theory and measurement. Journal of Personality and Social Psychology 52(1): 11–17.
44
Ettredge, M.L., C. Li, and L. Sun. 2006. The impact of SOX Section 404 internal control quality assessment on audit delay in the SOX era. Auditing: A Journal of Practice & Theory 25(2): 1–23.
Fischbacher, U. 2007. Z-tree: Zurich toolbox for ready-made economic experiments. Experimental Economics 10(2): 171–178.
Fitzgerald, B.C., C.J. Wolfe, and K.W. Smith. 2015. Management’s preference: Can auditors stop it from biasing accounting estimates? Working paper, Northeastern University and Texas A&M University.
Foster, J.D., and W.K. Campbell. 2007. Are there such things as ‘‘narcissists’’ in social psychology? A taxometric analysis of the narcissistic personality inventory. Personality and Individual Differences 439(6): 1321–1332.
Foster, J.D., W.K. Campbell, and J.M. Twenge. 2003. Individual differences in narcissism: Inflated self-views across the lifespan and around the world. Journal of Research in Personality 37(6): 469–486.
Firth, M., P. H. Malatesta, Q. Xin, and L. Xu. 2012. Corporate investment, government control, and financing channels: Evidence from China’s listed companies. Journal of Corporate Finance 18(3): 433–450.
Francis, J., R. LaFond, P. Olsson, and K. Schipper. 2005. The market pricing of accruals quality. Journal of Accounting and Economics 39(2): 295–327. Furnham, A., and J. Crump. 2014. A big-five facet analysis of sub-clinical narcissism: Understanding boldness in terms of well-known personality traits. Personality and Mental Health 8(3): 209–217.
Galinsky, A.D., and T. Musweiler. 2001. First offers as anchors: The role of perspective-taking and negotiator focus. Journal of Personality and Social Psychology 81(4): 657–669.
Gibbins, M., S.A. McCracken, and S.E. Salterio. 2005. Negotiations over accounting issues: The congruency of audit partner and chief financial officer. Auditing: A Journal of Practice & Theory 24(Supplement): 171–193.
Gibbins, M., S.A. McCracken, and S.E. Salterio. 2007. The chief financial officer’s perspective on auditor-client negotiations. Contemporary Accounting Research 24(2): 387–422.
Gibbins, M., S.A. McCracken, and S.E. Salterio. 2010. The auditor’s strategy selection for negotiation with management: flexibility of initial accounting position and nature of the relationship. Accounting, Organizations and Society 35(6): 579–595.
Gibbins, M., S. Salterio, and A. Webb. 2001. Evidence about auditor-client management negotiation concerning client’s financial reporting. Journal of Accounting Research 31(3): 535–563.
45
Golec de Zavala, A., A. Cichocka, and I. Iskra-Golec. 2013. Collective narcissism moderates the effect of in-group image threat on intergroup hostility. Journal of Personality and Social Psychology 104(6): 1019–1039.
Greenhalgh, L. 1987. Relationships in negotiation. Negotiation Journal 3(3): 235–243.
Greenhalgh, L., and R.W. Gilkey. 1997. Clinical assessment methods in negotiation research: The study of narcissism and negotiator effectiveness. Groups Decision and Negotiation 6(4): 289–316.
Grijalva, E., and P.D. Harms. 2014. Narcissism: An integrative synthesis and dominance complementarity model. Academy of Management Perspectives 28(2): 108–127.
Gul, F.A., D. Wu, and Z. Yang. 2013. Do individual auditors affect audit quality? Evidence from archival data. The Accounting Review 88(6): 1993–2023.
Hageman, A.M. 2008. A review of the strengths and weaknesses of archival, behavioral, and qualitative research methods: recognizing the potential benefits of triangulation. Advances in Accounting Behavioral Research 11: 1–30.
Hales, J., J.L. Hobson, and R. Resuteck. 2012. The dark side of socially mediated rewards: How narcissism and social status affect managerial reporting. Working paper, Georgia Tech (March).
Ham, C., N. Seybert, and S. Wang. 2014. Narcissism is a bad sign: CEO signature size, investment, and performance. Working paper, University of Maryland (June).
Ham, C., M. Lang, N. Seybert, and S. Wang. 2015. CFO narcissism and financial reporting quality. Working paper, University of Maryland (April).
Heifetz, A., and E. Segev. 2004. The evolutionary role of toughness in bargaining. Games and Economic Behavior 49(1): 117–134.
Hepper, E.G., C.M. Hart, R. Meek, S. Sicek, and C. Sedikides. 2014. Narcissism and empathy in young offenders and non-offenders. European Journal of Personality 28(2): 201–210.
Hollindale, J., P. Kent, and R. McNamara. 2011. Auditor tactics in negotiations: A research note. International Journal of Auditing 15: 288–300.
Holt, C.A, and S.K. Laury. 2002. Risk aversion and incentive effects. American Economic Review 92(5): 1644–1655.
Hopwood, C.W., M.B. Donnellan, R.A. Ackerman, K.M. Thomas, L.C. Morey, and A.E. Skodol. 2013. The validity of the personality diagnostic questionnaire-4 narcissistic personality disorder scale for assessing pathological grandiosity. Journal of Personality Assessment 95(3): 274–283.
46
Hüffmeier, J., P.E. Freund, A. Zerres, K. Backhaus, and G. Hertel. 2014. Being tough or being nice? A meta-analysis on the impact of hard- and softline strategies in distributive negotiations. Journal of Management 40(3): 866–892.
Hui, C.H., and H.C. Triandis. 1985. Measurement in cross-cultural psychology. Journal of Cross-Cultural Psychology 16(2): 131–152.
Hurtt, R.K. 2010. Development of a scale to measure professional skepticism. Auditing: A Journal of Practice & Theory 29(1): 149–171.
Hyder, E.B., M.J. Prietula, and L.R. Weingart. 2000. Getting to best: Efficiency versus optimality in negotiation. Cognitive Science 24(2): 169–204.
Jick, T.D. 1979. Mixing qualitative and quantitative methods: Triangulation in action. Administrative Science Quarterly 24: 602–611.
Johnson, E.N., J.R. Kuhn, Jr., B.A. Apostolou, and J.M. Hassell. 2013. Auditor perceptions of client narcissism as a fraud attitude risk factor. Auditing: A Journal of Practice & Theory 32(1): 203–219.
Jones, D.N., and A.L. Neria. 2015. The dark triad and dispositional aggression. Personality and Individual Differences 86: 60–64.
Jorgenson, D.O. 1977. Signature size and dominance: A brief note. Journal of Psychology 97(2): 269–270.
Judd, J.S., K.J. Olsen, and J.M. Stekelberg. 2015. CEO narcissism, accounting quality, and external audit fees. Working paper, University of Arizona (May).
Kong, D.T. 2015. Narcissists’ negative perceptions of their counterpart’s competence and benevolence and their own reduced trust in a negotiation context. Personality and Individual Differences 74(February): 196–201.
Kothari, S. P., A. J. Leone, and C. E. Wasley. 2005. Performance matched discretionary accrual measures. Journal of Accounting and Economics 39(1): 163–197.
Kubarych. T.S, I.J. Deary, and E.J. Austin. 2004. The narcissistic personality inventory: Factor structure in a non-clinical sample. Personality and Individual Differences 36(4): 857–872.
Küfner, A.C.P., S. Nestler, and M.D. Back. 2013. The two pathways to being an (un-)popular narcissist. Journal of Personality 81(2): 184–195.
Kwan, V.S.Y., L.L. Kuang, and N.H.H. Hui. 2009. Identifying the sources of self-esteem: The mixed medley of benevolence, merit, and bias. Self and Identity 8(2-3): 176–195.
47
Lee, C. 2003. Cowboys and Dragons: Shattering Cultural Myths to Advance Chinese-American Business. Chicago, IL: Dearborn Trade Publishing.
Lee, S., J. Brett, and J. H. Park. 2012. East Asians’ social heterogeneity: Differences in norms among Chinese, Japanese, and Koreans negotiators. Negotiation Journal 28(4): 429–452.
Lei, Y., J. Wu, and H. Sun. 2013. Impact of internal control audit on earnings quality: Evidence from listed companies in Shanghai Stock Exchange. Accounting Research 11: 75–81 (in Chinese).
Lennox, C.S., X. Wu, and T. Zhang. 2014. Does mandatory rotation of audit partners improve audit quality? The Accounting Review 89(5): 1775–1803.
Liu, X., and W. Chi. 2014. Does disclosure of CPAs’ names improve audit quality? Evidence from clients’ accounting restatements. The Chinese Certified Public Accountant 3: 56-64 (in Chinese).
Luchner, A.F., J.M. Houston, C. Walker, and M.A. Houston. 2011. Exploring the relationship between two forms of narcissism and competitiveness. Personality and Individual Differences 51(6): 779–782.
Ma, Z., and A. Jaeger. 2005. Getting to yes in China: Exploring personality effects in Chinese negotiation style. Group Decision and Negotiation 14(5): 415–437.
Majors, T. 2015. The interaction of communicating uncertainty and the dark triad on managers’ reporting decisions. Working paper, University of Illinois at Urbana-Champaign (April).
Mayhew, B.W., J.W. Schatzberg, and G.R. Sevcik. 2004. Examining the role of auditor quality and retained ownership in IPO markets: Experimental Evidence. Contemporary Accounting Research 21(1): 89–130.
Mazei, J., J. Hüffmeier, P.A. Freund, A.F. Stuhlmacher, L. Bilke, and G. Hertel. 2015. A meta-analysis on gender differences in negotiation outcomes and their moderators. Psychological Bulletin 141(1): 85–104.
McCracken, S.A., S.E. Salterio, and M. Gibbins. 2008. Auditor-client management relationships and roles in negotiating financial reporting. Accounting, Organizations and Society 33(4/5): 362–383.
McCracken, S.A., S.E. Salterio, and M. Gibbins. 2011. Do managers intend to use the same negotiation strategies as partners? Behavioral Research in Accounting 23(1): 131–160.
Miller, J.D., and W.K. Campbell. 2008. Comparing clinical and social-personality conceptualizations of narcissism. Journal of Personality 76(3): 449–476.
48
Miller, J.D., J. McCain, D.R. Lynam, L.R. Few, B. Gentile, J. MacKillop, and W.K. Campbell. 2014. A comparison of the criterion validity of popular measures of narcissism and narcissistic personality disorder via the use of expert ratings. Psychological Assessments 26(3): 958–969.
Morf, C.C., and F. Rhodewalt. 2001. Unraveling the paradoxes of narcissism: A dynamic self-regulatory processing model. Psychological Inquiry 12(4): 176–191.
Nevicka, B., A.H.B. De Hoogh, A.E.M. Van Vianen, B. Beersma, and D. McIlwain. 2011. All I need is a stage to shine: Narcissists’ leader emergence and performance. Leadership Quarterly 22(5): 910–925.
O’Neill, T.A., and N.J. Allen. 2014. Team task conflict resolution: An examination of its linkage to team personality composition and team effectiveness outcome. Group Dynamics: Theory, Research, and Practice 18(2): 159–173.
O’Reilly, C.A., B. Doer, D.F. Caldwell, and J.A. Chatman. 2014. Narcissistic CEOs and executive compensation. Leadership Quarterly 25(2): 218–234.
Park, S.W., J. Ferrero, C.R. Colvin, and D.R. Carney. 2013. Narcissism and negotiation: Economic gain and interpersonal loss. Basic and Applied Social Psychology 35(6): 569–574.
Quadackers, L., T. Groot, and A. Wright. 2014. Auditors’ professional skepticism: Neutrality versus presumptive doubt. Contemporary Accounting Research 31(3): 639–657.
Raskin, R., and J. Novacek. 1989. An MMPI description of the narcissistic personality. Journal of Personality Assessment 53(1): 66–80.
Raskin, R., and H. Terry. 1988. A principal-components analysis of the narcissistic personality inventory and further evidence of its construct validity. Journal of Personality and Social Psychology 54(5): 890–902.
Resick, C.J., D.S. Whitman, S.M. Weingarden, and N.J. Hiller. 2009. The bright-side and the dark-side of CEO personality: Examining core self-evaluations, narcissism, transformational leadership, and strategic influence. Journal of Applied Psychology 94(6): 1365–1381.
Rhodewalt, F., and C.C. Morf. 1995. Self and interpersonal correlates with of the narcissistic personality inventory: A review and new findings. Journal of Personality Research 29(1): 1–23.
Ronningstam, E. 2009. Narcissistic personality disorder: Facing DSM-V. Psychiatric Annals 39(3): 111–121.
49
Ryckman, R.M., B. Thornton, and J.C. Butler. 1994. Personality correlates of the hypercompetitive attitude scale: Validity tests of Horney’s theory of neurosis. Journal of Personality Assessment 62(1): 84–94.
Salterio, S.E. 2012. Fifteen years in the trenches. Auditor-client negotiations exposed and explored. Accounting and Finance 52(Supplement): 233–286.
Sandy, S.V., S.K. Boardman, and M. Deutsch. 2006. Personality and conflict. In M. Deutsch, P.T. Coleman, and E.C. Marcus (Eds.) The Handbook of Conflict Resolution, Second Edition (331–355). Hoboken, NJ: Jossey-Bass.
Scandura, T.A., and E.A. Williams. 2000. Research methodology in management: Current practices, trends, and implications for future research. Academy of Management Journal 43(6): 1248–1264.
Scofield, S.B., T.J. Phillips, Tand C.D. Bailey. 2004. An empirical reanalysis of the selection-socialization hypothesis: A research note. Accounting, Organizations and Society 29(5/6): 543–563.
Smith, H.W. 1975. Strategies of social research: The methodological imagination. Englewood Cliffs, NJ: Prentice-Hall.
Stanley, J.D., D.M. Brandon, and J.J. McMillan. 2015. Does lowballing impair audit quality? Evidence from client accruals surrounding analyst forecasts. Journal of Accounting and Public Policy 34(6): 625–645.
Triandis, H. 1986. Collectivism vs. individualism: A reconceptualization of a basic concept in cross-cultural psychology. In C. Bagley and G. Verma (Eds.) Personality, Cognition, and Values: Cross-Cultural Perspectives of Childhood and Adolescence (223–249). London: Macmillan Publishers.
Trotman, K.T., A.M. Wright, and S. Wright. 2009. An examination of the effects of auditor rank on pre-negotiation judgments. Auditing: A Journal of Practice & Theory 28(1): 191–203.
Twenge, J. M., S. Konrath, J.D. Foster, W.K. Campbell, and B.J. Bushman. 2008. Egos inflating over time: A cross-temporal meta-analysis of the Narcissistic Personality Inventory. Journal of Personality 76(4): 875–901.
Wang, K.J. 2010. Negotiating a fair value under accounting uncertainty: A laboratory experiment. Behavioral Research in Accounting 22(1): 109–134.
Wang, K.J., and B.M. Tuttle. 2009. The impact of auditor rotation on auditor-client negotiation. Accounting, Organizations and Society 34(2): 222–243.
50
Wang, Y., L. Yu, and Y. Zhao. 2015. The association between audit-partner quality and engagement quality: Evidence from financial report misstatements. Auditing: A Journal of Practice & Theory 34(3): 81–112.
Watson, P.J., S.O. Grisham, M.V. Trotter, and M.D. Biderman. 1984. Narcissism and empathy: Validity evidence for the narcissistic personality inventory. Journal of Personality Assessment 48(3): 301–305.
Watson, P.J., T. Little, S. Sawrie, and M.D. Biderman. 1992. Measures of the narcissistic personality: Complexity of relationships with self-esteem and empathy. Journal of Personality Disorder 6(4): 434–449.
Watson, P.J., R.J. Morris, and L. Miller. 1997–1998. Narcissism and the self as continuum: Correlations with assertiveness and hypercompetitiveness. Imagination, Cognition, and Personality 17(3): 249–259.
Weber, L., and E. Dwoskin. 2014. Are workplace personality tests fair? Wall Street Journal, September 29, available for download at http://www.wsj.com/articles/are-workplace-personality-tests-fair-1412044257.
Weingart, L.R., E.B. Hyder, and M.J. Prietula. 1996. Knowledge matters: the effect of tactical descriptions on negotiation behavior and outcome. Journal of Personality and Social Psychology 70(6): 1205–1217.
Whitworth, J.D., and T.A. Lambert. 2104. Office-level characteristics of the Big 4 and audit report timeliness. Auditing: A Journal of Practice & Theory 33(3): 129–152.
Wink, P. 1991. Two faces of narcissism. Journal of Personality and Social Psychology 61(4): 590–597.
Zweigenhaft, R. L. 1970. Signature size: A key to status awareness. Journal of Social Psychology 81(1): 49–54.
Zweigenhaft, R. L. 1977. The empirical study of signature size. Social Behavior and Personality 5(1): 177–185.
Zweigenhaft, R.L., and D. Marlowe. 1973. Signature size: Studies in expressive movement. Journal of Consulting and Clinical Psychology 40(3): 469–473.