Description
We conduct an experiment to investigate the effect of rankings, which are pervasive in
practice, on the honesty of managers’ budget reports, which is important for sound
decision making in organizations. Participants in our experiment are ranked in one of four
ways: (1) firm profit, (2) own compensation, (3) both firm profit and own compensation,
and (4) randomly, which serves as our baseline condition. None of the rankings affect
participants’ remuneration. Compared to our baseline (random rankings) setting, where
participants indeed exhibit honesty concerns, we find that rankings based on firm profit
significantly increase honesty and that rankings based on own compensation significantly
decrease honesty. Participants who received both rankings were significantly more honest
than participants in the own compensation rankings condition. We did not, however, find
significant differences in honesty between the both rankings and firm profit rankings
conditions. As such, participants in the both rankings condition seemed to focus more on
the firm profit metric than on the financially congruent own compensation metric. We also
find that our results are stable across periods, suggesting that the effects of rankings
neither increased nor dissipated over time
The effect of rankings on honesty in budget reporting
q
Jason L. Brown
a,?
, Joseph G. Fisher
a
, Matthew Sooy
b
, Geoffrey B. Sprinkle
a
a
Kelley School of Business, Indiana University, 1309 E. Tenth Street, Bloomington, IN 47405, United States
b
Gatton College of Business & Economics, University of Kentucky, 550 S. Limestone Street, Lexington, KY 40502, United States
a b s t r a c t
We conduct an experiment to investigate the effect of rankings, which are pervasive in
practice, on the honesty of managers’ budget reports, which is important for sound
decision making in organizations. Participants in our experiment are ranked in one of four
ways: (1) ?rm pro?t, (2) own compensation, (3) both ?rm pro?t and own compensation,
and (4) randomly, which serves as our baseline condition. None of the rankings affect
participants’ remuneration. Compared to our baseline (random rankings) setting, where
participants indeed exhibit honesty concerns, we ?nd that rankings based on ?rm pro?t
signi?cantly increase honesty and that rankings based on own compensation signi?cantly
decrease honesty. Participants who received both rankings were signi?cantly more honest
than participants in the own compensation rankings condition. We did not, however, ?nd
signi?cant differences in honesty between the both rankings and ?rm pro?t rankings
conditions. As such, participants in the both rankings condition seemed to focus more on
the ?rm pro?t metric than on the ?nancially congruent own compensation metric. We also
?nd that our results are stable across periods, suggesting that the effects of rankings
neither increased nor dissipated over time. We discuss the contributions of our study
and concomitant ?ndings to accounting research and practice.
Ó 2014 Elsevier Ltd. All rights reserved.
Introduction
Managers routinely receive feedback from the
accounting system regarding ?rm-performance variables
under their control and also, of course, their own
compensation. In turn, managers frequently are ranked
on such variables or can infer their rankings from available
information. While such rankings may be used to
determine compensation, retention, and promotion, they
also may have little bearing on managers’ remuneration
(e.g., Nordstrom, Lorenzi, & Hall, 1990; West & Mykerezi,
2011) or may even con?ict with managers’ personal ?nan-
cial incentives (Grant, 2013). Moreover, ample research
suggests that individuals’ concerns for rank affect their
behavior. Research in accounting, for example, has
documented positive (e.g., Frederickson, 1992; Tafkov,
2013) and negative (e.g., Hannan, Krishnan, & Newman,
2008; Hannan, McPhee, Newman, & Tafkov, 2013) effects
of rankings on effort and concomitant task performance.
Although the use of rankings is pervasive in practice,
prior research has not examined whether rankings affect
the honesty of managers’ reports. For a myriad of reasons,
honest reporting by ?rm participants is important for
sound decision making. Indeed, prior research in account-
ing documents that individuals not only have preferences
for honesty (Evans, Hannan, Krishnan, & Moser, 2001;
Rankin, Schwartz, & Young, 2003) but also that individuals’http://dx.doi.org/10.1016/j.aos.2014.03.001
0361-3682/Ó 2014 Elsevier Ltd. All rights reserved.
q
We thank Spencer Anderson, Jason Kuang, Victor Maas, Adam
Presslee, Mike Shields, Laura Wang, two anonymous Reviewers, and
participants at the 2013 American Accounting Association Management
Accounting Section Meeting for their very helpful comments and
suggestions. We also thank the Marketing Department at Indiana
University for the use of its laboratory.
?
Corresponding author. Tel.: +1 812 855 2381; fax: +1 812 855 4985.
E-mail addresses: [email protected] (J.L. Brown), jo?sher@
indiana.edu (J.G. Fisher), [email protected] (M. Sooy), sprinkle@
indiana.edu (G.B. Sprinkle).
Accounting, Organizations and Society 39 (2014) 237–246
Contents lists available at ScienceDirect
Accounting, Organizations and Society
j our nal homepage: www. el sevi er. com/ l ocat e/ aos
preferences for honesty are affected by fairness concerns
(Zhang, 2008), group incentives (Church, Hannan, & Kuang,
2012), and speci?c features of the managerial setting, such
as the precision of the information system (Hannan,
Rankin, & Towry, 2006) and who has budget authority
(Rankin, Schwartz, & Young, 2008).
In this paper, we examine whether rankings affect the
veracity of individuals’ reports. We employ a participative
budgeting setting similar to Evans et al. (2001) in which
managers (hereafter subordinates) receive private cost
information and then submit a budget request to a
superior. The budget request splits the available surplus be-
tween the subordinate and the superior. Subordinate com-
pensation is based on the budget request, and the
subordinate has strict ?nancial incentives to request the
highest budget amount. The superior’s residual claim (?rm
pro?t), however, is minimized by this subordinate
strategy. Consistent with Evans et al. (2001), the subordi-
nate is requested to provide an accurate (honest) budget re-
quest. Increased subordinate budget accuracy (honesty)
results in more surplus being allocated to the ?rm and less
to the subordinate.
In our setting, six subordinates, which constitute a co-
hort, are ranked in one of four ways: (1) subordinate com-
pensation; (2) superior residual claim (?rm pro?t); (3)
both subordinate compensation and ?rm pro?t; and, (4)
randomly. Rankings based on subordinate compensation
are congruent with subordinates’ ?nancial incentives to
maximize the surplus claimed but are not aligned with
honest reporting. Rankings based on ?rm pro?t are
congruent with honest reporting but at odds with subordi-
nates’ ?nancial incentives. Rankings based on both
subordinate compensation and ?rm pro?t provide subordi-
nates one metric that is congruent with subordinates’
?nancial incentives and one metric that is at odds with
subordinates’ ?nancial incentives. Random rankings are
not affected in any way by subordinates’ budget requests
and, as such, are not under subordinates’ control.
1
We ?nd that rankings signi?cantly affect managerial
reporting honesty. Compared to our baseline setting,
where subordinates indeed exhibit honesty concerns, we
?nd that rankings based on ?rm pro?t increase subordi-
nates’ honesty. This ?nding suggests that ?rms can use
rankings as a low-cost, informal control to constrain
opportunistic reporting. Incorporating rankings into the
design of control systems could allow ?rms to reduce
information asymmetry between employees and the ?rm,
which leads to better decision making and lower budget-
ary slack, both of which should increase ?rm pro?tability.
Moreover, this result suggests that, for example, the ?rm-
level rankings frequently published in the popular press
may curb self-interested behavior in organizations.
In contrast, rankings based on own compensation de-
crease subordinates’ honesty. This ?nding suggests that
providing relative performance feedback to employees
can lead to behavior that is harmful to the ?rm. As a result,
?rms should consider carefully how they rank their
employees and also be aware of how employees may use
accounting information to rank themselves. This may help
explain why some ?rms closely guard their compensation
data and are loathe to share such information with employees.
Subordinates who received both rankings were more
honest than subordinates in the own compensation condi-
tion. Moreover, we do not ?nd signi?cant differences in
honesty between the both rankings and ?rm pro?t condi-
tions. As such, subordinates in the both rankings condition
seemed to focus more on the ?rm pro?t metric than on the
?nancially congruent own compensation metric. This re-
sult is somewhat surprising as, intuitively, one might posit
that having both rankings would lead subordinates to en-
gage in more self-regarding behavior because subordinates
could internally justify their reporting decisions by focus-
ing on the compensation metric. Our ?ndings fromthis set-
ting suggest that as long as ?rm-based rankings are
provided the control loss associated with having access
to a ranking that encourages self-interested behavior may
not be as large as expected. Finally, we ?nd that our results
are stable across periods, suggesting that the effects of
rankings neither increase nor dissipate over time.
The remainder of this paper is organized into four sec-
tions. The next section develops the hypotheses, and the
third section presents the experimental design. The fourth
section reports the results, and the ?nal section provides a
summary of the study.
Background and hypotheses
Research setting
Accurate reporting by employees is important for mak-
ing sound organizational decisions. To this end, Evans et al.
(2001) employed a novel design to examine subordinates’
honesty in a managerial reporting setting. In this setting,
a subordinate, who has private information regarding the
production cost of an investment project, submits a budget
request to a superior who funds the project. Both the rev-
enue and the probabilistic distribution of the project’s cost
are known by the subordinate and the superior. Only the
subordinate, however, is aware of the actual project cost.
The subordinate’s budget request is always approved and
the subordinate receives the difference between the bud-
get request and the actual project cost. The superior re-
ceives the residual, which equals the difference between
the project’s revenue and the budget request. In this set-
ting, the subordinate is requested to accurately report the
project cost via the budget request.
Assuming subordinates are only concerned with wealth
maximization, conventional economic analysis predicts
that subordinates will report the highest cost possible
and claim the entire budget surplus. Evans et al. (2001)
?nd, however, that subordinates’ preferences for honesty
attenuate personal wealth aspirations – subordinates’ re-
ports are signi?cantly more truthful than would be pre-
dicted by conventional economic analysis. Research in
accounting has extended this result and ?nds that individ-
uals not only have preferences for honesty but also that
individuals’ preferences for honesty are affected by a
1
Random rankings serve as our baseline condition. We employ random,
rather than no, rankings in our baseline condition to ensure that partic-
ipants in all conditions have rankings information and, thus, to isolate the
honesty effects associated with how participants are ranked.
238 J.L. Brown et al. / Accounting, Organizations and Society 39 (2014) 237–246
number of factors, including fairness concerns, group
incentives, the precision of the information system, and
who has budget authority (Church et al., 2012; Hannan
et al., 2006; Rankin et al., 2008; Zhang, 2008).
Adding to this research stream, we examine whether
rankings affect the honesty of subordinates’ budget re-
quests. We augment the Evans et al. (2001) setting by pair-
ing each subordinate with an actual superior and by
ranking subordinates within cohort groups. A cohort group
consists of six subordinates, who are ranked from highest
to lowest in one of four ways. None of the rankings affect
subordinates’ remuneration. Additionally, subordinates
were only provided with the ordinal ranks. As such, while
each subordinate certainly knows his/her actual budget re-
quest, s/he does not knowthe actual budget requests of the
other subordinates.
2
First, we examine a setting where subordinates are
ranked each period from highest to lowest based on their
own compensation.
3
In this setting, wealth and rank con-
cerns are aligned – subordinates maximize both their wealth
and their rank by submitting the highest possible budget re-
quest. Preferences for honesty are, however, at odds with
preferences for wealth and rank in this setting.
Second, we examine a setting where subordinates are
ranked each period from highest to lowest based on ?rm
pro?t.
4
In this setting, wealth and rank concerns are at odds
– subordinates maximize their wealth (rank) by submitting
the highest (lowest) possible budget request. Moreover,
preferences for honesty are congruent with preferences for
rank in this setting.
Third, we examine a setting where subordinates are
ranked each period from highest to lowest based on (a)
their own compensation and (b) ?rm pro?t. Thus, in this
setting subordinates receive two con?icting rankings, with
each ranking containing the characteristics and tensions
from our ?rst two settings. Finally, we examine a setting
where subordinates are randomly ranked each period from
highest to lowest. This setting serves as our control (base-
line) treatment condition.
Social comparison theory and prior research
Human beings are a social species, and it is widely ac-
cepted that we desire to compare ourselves to others
(Buunk & Gibbons, 2007). Festinger (1954) formalized
these thoughts and introduced social comparison theory,
which posits that individuals have a desire to compare
themselves favorably to their peers.
5
Metee and Smith
(1977) propose that social comparison theory is about our
desire to know ourselves. This desire to know ourselves is
not only satis?ed by receiving information regarding our
performance but also by comparing our performance with
others.
Rankings enable individuals to make interpersonal
comparisons which, in turn, lead to a desire to improve
one’s rank and competition. Indeed, there is considerable
research consistent with social comparisons affecting indi-
viduals’ behavior. Speci?cally, research from experiments
(Duffy & Kornienko, 2010; Fisher, Maines, Peffer, & Sprin-
kle, 2002; Frederickson, 1992; Hales, Hobson, & Resutek,
2012; Hannan et al., 2008, 2013; Tafkov, 2013) and the
?eld (Anderson, Crowell, Sponsel, Clarke, & Brence, 1982;
Azmat & Iriberi, 2009; Blanes i Vidal & Nossol, 2011; Nord-
strom et al., 1990; Wikoff, Anderson, & Crowell, 1983) both
support social comparison theory and ?nd that rankings
affect behavior in a variety of settings.
Social comparison theory (Festinger, 1954) provides a
useful framework for understanding the relationship be-
tween rankings and behavior. The effects of rankings have
been attributed to many factors. For example, rankings can
create descriptive norms of appropriate behavior, a way for
individuals to improve themselves, help foster/maintain a
positive self-image, lead to feelings of pride, or help avoid
negative feelings such as shame (see, e.g., Beach & Tesser,
1995; Goldstein & Cialdini, 2007; Lazarus, 1991; Smith,
2000; Tesser, 1988; Tesser & Campbell, 1980; Wood,
1989).
A consistent ?nding from social comparison studies
that is particularly germane to our study is that individuals
frequently exhibit an oft-referred to ‘‘upward drive’’ moti-
vation to exceed others’ performance (Festinger, 1954; Suls
& Miller, 1977). Speci?cally, prior research ?nds that indi-
viduals compare themselves to others whose performance
is better than their performance (Nosanchuk & Erickson,
1985; Wheeler, Koestner, & Driver, 1982; Wood, 1989).
For example, several studies ?nd that individuals tend to
compare themselves to individuals who are ranked above
them (Gruder, 1971; Wheeler, 1966; Wheeler et al., 1969).
Prior research, in turn, ?nds that these upward compar-
isons frequently lead individuals to increase their effort,
performance, or rank (Buunk, Kuyper, & van der Zee,
2005; Huguet, Dumas, Monteil, & Genestoux, 2001). For
example, in a study that is related to ours, Duffy and
Kornienko (2010) rank proposers in a dictator game by
the amount of money they give to recipients. Duffy and
Kornienko (2010) ?nd that ranking participants in this
fashion affects the amount of charitable giving. Moreover,
Duffy and Kornienko (2010) suggest that the difference in
giving behavior is driven, in part, by competition among
participants.
6
2
We made this design choice so as not to provide subordinates in our
baseline (random rankings) setting information that could be used to rank
themselves. We also made this design choice to examine the starkest form
of rankings.
3
As such, subordinates submitting the highest (lowest) budget request
would receive the #1 (#6) rank.
4
As such, subordinates submitting the lowest (highest) budget request
would receive the #1 (#6) rank.
5
Although it is Festinger’s (1954) classic paper that details speci?c
hypotheses regarding social comparison theory, the idea of social compar-
ison and reference groups are also highlighted in earlier research. For
example, Hyman (1942) ?nds that the assessment of one’s status regarding
?nances, intelligence, and physical attractiveness depends on one’s com-
parison group.
6
While our work shares similarities with previous social comparison
studies in accounting and other areas, there is a fundamental difference. To
obtain a higher payoff in our study, participants must misreport the true
cost, which is a feature not present in prior rank-based research. Given that
ethical systems encourage honesty (Murphy, 1993) and that honesty can
increase ef?ciencies and surplus (Arrow, 1974; Noreen, 1988), it is
important to understand how rankings affect individuals’ reporting
decisions.
J.L. Brown et al. / Accounting, Organizations and Society 39 (2014) 237–246 239
Thus, social comparison research not only suggests
that individuals will make upward comparisons but also
suggests that, as a result of these comparisons, individu-
als will take actions to improve their performance to
ultimately achieve a more favorable comparison (rank-
ing). In a similar vein, we posit that individuals will at-
tend to the rankings provided to them and exhibit an
‘‘upward drive’’ motivation to improve their rankings.
As such, we posit that individuals will change their
reporting behavior (honesty) to improve their ranking
among their peers.
Hypotheses 1 and 2
In our setting, subordinates’ ranks are not tied to their
compensation so, as a result, conventional economic the-
ory predicts that subordinates will ignore the rankings
and submit the maximum budget request.
7
Social compar-
ison theory and prior empirical research suggests, however,
that rankings will in?uence subordinates’ reporting behavior
due to their desire to improve their ranking. That is, in the
own compensation (?rm pro?t) ranking condition, this
means that subordinates will submit higher (lower) budget
requests than if subordinates are randomly ranked.
8
This
leads to our two hypotheses:
H1. Subordinates who are ranked based on their
own compensation will submit higher budget requests
(i.e., be less honest) than subordinates who are randomly
ranked.
H2. Subordinates who are ranked based on ?rm pro?t will
submit lower budget requests (i.e., be more honest) than
subordinates who are randomly ranked.
Research question – multiple rankings
In many settings, individuals are ranked on multiple
metrics. For example, popular business periodicals
frequently rank executives and ?rms on factors such
as executive compensation, ?rm pro?tability, and cor-
porate social responsibility. And, the relations among
these variables may be positive, non-existent, or
negative.
9
Likewise, companies often explicitly or implicitly rank
employees on metrics that are either not linked, or may
even con?ict with, their remuneration. For example,
companies such as Google, Southwest Airlines, and Zappos
provide employees with public relative performance infor-
mation regarding their organizational helping behaviors
(Grant, 2013). These companies do not, however, directly
compensate such behaviors, and it is unclear whether such
behaviors are ultimately linked to compensation (e.g., ‘‘no
good deed goes unpunished’’).
Additionally, many ?rms utilize ‘‘dashboards’’ that are
frequently updated to show managers’ ongoing perfor-
mance across several measures that are not necessarily
correlated or may even con?ict (e.g., average waiting/ser-
vice time versus safety/cleanliness at a fast-food restau-
rant). These dashboards, as exist at companies such as
McDonald’s and eBay, can serve as rankings by presenting
peer-manager performance information. These dashboards
also enable managers to ‘‘drill down’’ and/or query others’
performance so as to evaluate their relative performance.
Along these lines, we examine a setting where subordi-
nates are ranked each period from highest to lowest based
on (a) their own compensation and (b) ?rm pro?t. Thus, in
this setting subordinates receive two con?icting rankings.
In such settings it is unclear what rank(s), if any, to which
individuals will attend.
Some research suggests that individuals may use ‘‘mor-
al wiggle room’’ to act in their self-interest when they have
con?icting rankings (Dana, Weber, & Kuang, 2007; Haisley
& Weber, 2010). This line of research suggests that provid-
ing subordinates with both rankings may lead subordi-
nates to engage in more ?nancial self-regarding behavior
because subordinates can internally justify their reporting
decisions via the own compensation ranking. In this sce-
nario, subordinates’ budget requests (honesty) would most
closely resemble subordinates’ budget requests who re-
ceive rankings based on their own compensation.
In contrast, another research stream suggests that indi-
viduals’ behavior may be affected by drawing their atten-
tion to pro-social behavior (Cialdini, Reno, & Kallgren,
1990; Krupka & Weber, 2009).
10
For example, Krupka and
Weber (2009) examine a binary dictator game with a
pro-social allocation and a sel?sh allocation and ?nd that
7
This is also a noteworthy difference between our study and some prior
research. By not linking ranks to remuneration, we are able to examine
whether rankings per se affect behavior. This distinction is important
because, when compensation is linked to ranks, it is unclear whether
individuals are responding to the rank and/or to the reward. For example,
individuals working under a tournament contract may ramp-up their effort
if they believe they have the ability to win the prize. Analogously,
individuals working under a tournament contract may provide low effort
if they assess their chances of winning to be low. In both cases, the
compensation-based effect (related to, e.g., effort and/or skill) may swamp
the rank effect. Moreover, rankings based on, for example, corporate social
responsibility are not necessarily linked to employee remuneration in
natural settings (cf. footnote 9).
8
This means that subordinates in the own compensation (?rm pro?t)
ranking condition will be less (more) honest than subordinates who are
ranked randomly.
9
Research generally ?nds that, despite the strong predictions by
economic theory, compensation is often independent of performance
(Baker, Jensen, & Murphy, 1988). For example, a meta-analytic review of
the literature regarding the determinants of CEO pay ?nds that size
accounts for 40% of the variance in CEO pay and ?rm performance accounts
for less than 5% (Tosi, Wermer, Katz, & Gomez-Mejia, 2000). A number of
researchers suggest that CEO pay seems to re?ect managerial power and
rent-seeking by CEOs rather than the provision of ef?cient incentive
contracting (Bebchuk & Fried, 2003; Blanchard, Lopez-de-Silanes, & Shle-
ifer, 1994). Due to the other forces that determine CEO pay, some studies
even document a negative relation between compensation and perfor-
mance (Core & Larker, 1999; Bick, Palmon, & Wald, 2006). While there are
not many studies that examine the relation between CEO pay and corporate
social responsibility (CSR), at least one study ?nds a negative relation
between CEO pay and CSR (Cai, Jo, & Pan, 2011).
10
This is consistent with social comparison theory, which suggests that
rankings can provide descriptive norms of appropriate behavior.
240 J.L. Brown et al. / Accounting, Organizations and Society 39 (2014) 237–246
having participants focus on what they believe other people
should do in such a decision context resulted in more partic-
ipants choosing the pro-social allocation.
Additionally, research relating to groups (Cookson,
2000; Kramer & Brewer, 1984; Rowe, 2004; Towry, 2003)
suggests that the presence of a ranking that encourages
the consideration of others may induce subordinates to
act less self-interested. For example, Rowe (2004) ?nds
that aligning the accounting report structure and the team
structure in such a way to create a ‘‘group frame’’ helps
mitigate the free-rider problem.
11
Moreover, in this sce-
nario, subordinates’ budget requests (honesty) would most
closely resemble subordinates’ budget requests who receive
rankings based on ?rm pro?t.
Finally, the two opposing ranking measures may simply
‘‘cancel’’ each other out resulting in little or no impact. In
this scenario, subordinates’ budget requests would most
closely resemble the budget requests of subordinates
who receive random rankings. Given the dif?culty in
developing a directional prediction for our two rankings
condition, we examine the following research question:
RQ: Will subordinates who are ranked based on both
their own compensation and ?rm pro?t submit
budget requests that are different than the
budget requests submitted by subordinates who
are ranked on either their own compensation,
?rm pro?t, or randomly?
Experimental method
Task
We employ a budgeting setting similar to Evans et al.
(2001). In our setting, each participant plays the role of
either subordinate or superior.
12
The subordinate submits
a budget request to the superior to fund the cost of an
investment project. Both the revenue ($4.00) and the proba-
bilistic distribution of the project’s cost (uniformly distrib-
uted, in increments of $0.01, between $2.00 and $4.00) are
known by the subordinate and superior but only the subor-
dinate knows the actual project cost. The subordinate’s bud-
get request (maximum request = $4.00) is always approved,
and the subordinate receives as compensation the difference
between the budget request and the actual project cost. The
superior receives the residual pro?t, which equals the differ-
ence between the project’s revenue and the budget report.
Thus, if the project cost equals $3.00 and the subordinate
submits a budget request of $3.60, then the subordinate
earns $0.60 ($3.60–$3.00) and the superior earns $0.40
($4.00–$3.60). In this setting, subordinates who wish to
maximize their own compensation should, for every cost
realization, submit a budget request of $4.00.
Each subordinate/superior dyad is paired for the entire
experiment, which consisted of two practice periods and ten
compensated periods. Each experimental session comprised
12participants, or sixsubordinate/superior dyads. Ineachper-
iod, all six subordinates received identical project cost draws.
Manipulations
We manipulate, between-participants, the rankings
provided to subordinates. We examine four ranking condi-
tions. None of the rankings, per se, affected participants’
remuneration.
In the Firm Pro?t Condition, subordinates are ranked
each period from highest (?rst) to lowest (sixth) based
on the pro?t they generated for the superior. In the Own
Compensation Condition, subordinates are ranked each per-
iod from highest to lowest based on their individual com-
pensation. In the Both Condition, participants received
rankings based on ?rm pro?t and on their own compensa-
tion. Finally, in the Random Condition, subordinates are
ranked each period from highest to lowest based on a ran-
domly generated number.
13
Participants and procedures
We conducted a computer-based laboratory experiment
with 180 undergraduate students using z-tree software (Fis-
chbacher, 2007). We ran 15 sessions – three sessions for the
?rm pro?t, own compensation, and random conditions, and
six sessions for the both condition (three sessions where
the ?rmpro?t ranking was displayed ?rst, and three sessions
wheretheowncompensationrankingwas displayed?rst). To
ensure the privacy of participants, each subordinate was as-
signed a color identi?er (blue, green, orange, purple, red, or
yellow) throughout the experiment. The color identi?ers
wereusedwhendisplayingsubordinaterankings.
14
Bothsub-
ordinates and superiors received a $5.00 one-time participa-
tion fee and a salary of $0.50 per period ($5.00 in total).
At the beginning of each period, subordinates received
information about the project’s cost.
15
Each subordinate
then submitted a budget request to his/her superior. The
budget request could range from $0.50 below the project’s
11
Rowe (2004) manipulated the accounting report structure by showing
the individual manager’s payoffs or both the individual manager’s payoffs
and the payoffs for other managers.
12
We chose to have a real participant play the role of the superior
because this design choice most closely mirrors settings in the natural
environment where one’s actions invariably affect the welfare of others.
Additionally, accounting research examining preferences for honesty has
frequently employed real participants for both subordinate and superior
roles (see, e.g., Hannan et al., 2006; Rankin et al. 2003; Rankin et al. 2008;
Schatzberg & Stevens, 2008).
13
Notably, subordinates receiving ?rm pro?t rankings also have available
own compensation rankings, and subordinates receiving own compensa-
tion rankings also have available ?rm pro?t rankings. This obtains because,
for example, in the ?rm pro?t condition, the subordinate with highest ?rm
pro?t rank (#1) will have the lowest own compensation rank (#6), and so
on. Thus, by inverting the ?rm pro?t ranks, subordinates can obtain the
own compensation ranks. In a like fashion, subordinates in the own
compensation ranking condition can construct the ?rm pro?t ranks. As
such, our manipulations make salient certain aspects of the payoff
distribution as, informationally, subordinates in all ranking conditions
(other than the random condition) have available both ?rm pro?t and own
compensation ranks.
14
Each subordinate knew his/her own color but never knew which colors
corresponded to the other subordinates. The color was visible to superiors
and other subordinates when rankings were displayed.
15
Participants were informed that cost draws were determined in
advance and that the project cost in each period was the same for all
subordinates. To facilitate comparisons across treatment conditions, the
cost sequences in all sessions were identical.
J.L. Brown et al. / Accounting, Organizations and Society 39 (2014) 237–246 241
cost to $4.00. As in Evans et al. (2001), we allow underre-
porting. As subordinates received $0.50 in salary each peri-
od, we limited underreporting to $0.50 to ensure that
subordinates’ earnings were not negative in any period.
Subordinates’ budget requests were automatically ap-
proved. After all subordinates submitted their budget re-
quests, subordinates were publicly ranked from ?rst to
sixth as per our manipulations.
16
When subordinates’ bud-
get requests were identical, they received the same rank and
their order was randomly determined. Rankings were visible
to all subordinates and superiors and remained visible until
all subordinates indicated that they were ready to proceed
to the ensuing period. After the tenth period, subordinates
received a ?nal ranking based on their cumulative perfor-
mance over the experiment. Participants then completed a
post-experimental questionnaire and were paid in cash for
their participation. Average earnings were $15.00, and each
experimental session lasted approximately 45 min.
Results
Hypotheses 1 and 2
There is a ?xed amount of resources available each per-
iod to be split between the superior and the subordinate.
Panel B: Percentage of Surplus Retained by Subordinates by Ranking Condition and Period
Period
Ranking Condition n 1 2 3 4 5 6 7 8 9 10 Overall
Own Compensation 18 80.9% 81.0% 84.6% 81.0% 86.0% 88.2% 88.6% 90.7% 78.1% 80.5% 84.0%
Random 18 73.8% 67.1% 64.5% 66.0% 68.9% 63.0% 71.3% 65.9% 70.2% 68.9% 68.0%
Both Rankings 36 51.7% 65.0% 49.6% 33.9% 68.9% 50.7% 58.6% 59.1% 56.1% 62.7% 55.6%
Firm Profit 18 45.2% 46.5% 33.9% 44.1% 57.5% 44.7% 49.0% 58.5% 32.3% 38.6% 45.0%
The percentage of surplus kept = Surplus Claimed = (Budget Request – Project Cost)
Surplus Available (Revenue – Project Cost)
When subordinates reported less than the actual cost, the percentage of surplus retained is negative. To ensure that
(the few) negative reports were computed out of a possible 100%, we calculated the percentage of surplus retained
relative to the fixed wage. Specifically, we calculated the percentage of surplus retained for negative reports as:
(Budget Request – Project Cost) /Fixed Wage per Period. Measuring underreporting relative to the fixed salary is a
more appropriate benchmark because it avoids denominator distortions caused by cost draws that are dramatically
larger or smaller than the $0.50 salary constraint. For example, in a period where the actual project cost was $2.50, a
subordinate could report a cost as low as $2.00. As such, we calculate the misreporting as $0.50/$0.50 = 100%
(Maximum allowed).
Panel A: Percentage of Surplus Retained by Subordinates by Ranking Condition and Period
Fig. 1. Percentage of surplus retained by subordinates.
16
As mentioned earlier, only ordinal rankings were displayed – partic-
ipants did not receive any cardinal information. That is, the subordinate
only knew his/her own budget request and not the budget requests of the
other subordinates.
242 J.L. Brown et al. / Accounting, Organizations and Society 39 (2014) 237–246
We calculate the total surplus available as the resources in
excess of the actual cost. The subordinate’s budget request
determines how the surplus is split between the superior
and subordinate. As such, the surplus retained by subordi-
nates is our primary dependent variable.
Collectively, H1 and H2 predict that the surplus re-
tained by subordinates will be greatest when subordinates
are ranked on their own compensation and least when
subordinates are ranked on ?rm pro?t. We posit that the
surplus retained by subordinates who are ranked ran-
domly will be between the two. Fig. 1 presents the average
percentage of surplus retained by subordinates for each
period by ranking condition.
17
Consistent with H1 and H2,
the average percentage of surplus retained by subordinates
in the own compensation ranking condition (84%) is higher
than the average percentage of surplus retained by subordi-
nates in the random ranking condition (68%) which, in turn,
is higher than the average percentage of surplus retained by
subordinates in the ?rm pro?t ranking condition (45%). The
average percentage of surplus retained by subordinates in
the both rankings condition (55.6%) lies between the
random ranking and ?rm pro?t ranking conditions.
Panel A of Table 1 reports the results of an ANOVA with
the percentage of surplus retained by each subordinate per
period as the dependent variable, condition as the be-
tween-subjects factor, and period as the within-subjects
(repeated-measures) factor. The results of this ANOVA re-
veal signi?cant main effects of ranking condition
(p = 0.003) and period (p = 0.053), indicating that the per-
centage of surplus retained by subordinates differs signi?-
cantly across conditions and periods.
Panel B of Table 1 presents the pairwise comparisons
between all ranking conditions.
18
These comparisons pro-
vide support for H1 and H2. Panel B of Table 1 shows that
subordinates who were ranked on their own compensation
retained signi?cantly more surplus than subordinates who
were ranked randomly (p = 0.024). Panel B of Table 1 also
shows that subordinates who were ranked based on ?rm
pro?t retained signi?cantly less surplus than subordinates
who were ranked randomly (p = 0.051), providing support
for H2.
19
Collectively, our results indicate that, even though the
rankings did not affect remuneration in any way, subordi-
nates’ budget requests are signi?cantly in?uenced by the
ranking metrics. We ?nd that the percentage of surplus
claimed by subordinates almost doubles when rankings
are based on own compensation (84%) compared to the
condition where rankings are based on ?rm pro?t (45%).
Our ?ndings suggest that ranking metrics are powerful
motivators that ?rms should carefully consider when
designing their control systems. In particular, ?rms may
bene?t from ranking on ?rm-based metrics (e.g., pro?tabil-
ity-based metrics), but they may exacerbate control prob-
lems when ranking on individual-compensation metrics.
Research question – multiple rankings
Panels A and B of Fig. 1 show that the percentage of sur-
plus retained by subordinates receiving both rankings
(55.6%) is closer to the percentage of surplus retained by
subordinates receiving the ?rm pro?t ranking (45%) than
Table 1
ANOVA on percentage of surplus retained by subordinates.
Source df M.S. F-statistic p-Value
Panel A: Overall ANOVA
Between subjects
Condition 3 5.32 5.03 0.003
Participant 86 1.06
Within subjects
Period 9 0.17 1.87 0.053
Period  Condition 27 0.09 1.02 0.433
Residual 774 0.09
Total 899 0.20
Ranking conditions F-statistic p-Value
Panel B: Follow-up ANOVA’s – pairwise comparisons
Own compensation: Random 4.20 0.024
*
Firm pro?t: Random 2.81 0.051
*
Firm pro?t: Own compensation 8.82 0.003
*
Own compensation: Both rankings 14.85 0.001
Firm pro?t: Both rankings 0.97 0.330
Random: Both rankings 2.47 0.122
Panel A presents the repeated-measures ANOVA for the percentage of surplus retained by subordinates. In Panel B, we present the pairwise comparisons
between ranking conditions – for these tests, we replicate the repeated-measures ANOVA using the two conditions of interest.
*
One-tailed p-value.
17
An ANOVA revealed no signi?cant order effects in the both rankings
condition. That is, the percentage of surplus retained by subordinates did
not differ depending on whether subordinates received the own compen-
sation rank or the ?rm pro?t rank ?rst or second. As such, we do not
distinguish between these two treatment conditions.
18
In Panel B, we replicate the repeated-measures ANOVA in Panel A using
the two conditions of interest. None of the period-by-condition pairwise
interaction effects are signi?cant (smallest p > 0.19).
19
Panel B of Table 1 also reports that the surplus retained by subordinates
who were ranked on ?rm pro?t is less than (p = 0.003) the surplus retained
by subordinates who were ranked based on their own compensation.
J.L. Brown et al. / Accounting, Organizations and Society 39 (2014) 237–246 243
to the percentage of surplus retained by subordinates
receiving the own compensation ranking (84%). Pairwise
comparisons bear this observation out. As documented in
Panel B of Table 1, subordinates in the both rankings con-
dition submitted signi?cantly lower budget requests than
subordinates in the own compensation ranking condition
(p = 0.001). The budget requests of subordinates in the
both rankings condition, however, did not differ signi?-
cantly from the budget requests of subordinates in the ?rm
pro?t ranking condition (p = 0.330). This ?nding is some-
what unexpected as one might posit that having both rank-
ings would lead subordinates to anchor on the own
compensation ranking and internally ‘‘justify’’ their report-
ing decisions via this ranking metric.
Subordinates’ budget requests throughout the experiment
In our tests of H1 and H2, we ?nd that the period effect is
signi?cant in the overall ANOVA. This is attributable primar-
ily to the period-by-period reporting variance resulting from
the actual cost differing across periods rather than to report-
ing trends over time. To more closely examine the period ef-
fect, we compared the reporting behavior of subordinates in
the ?rst half of the experiment to the reporting behavior of
subordinates in the second half of the experiment. For each
of the ranking conditions, the surplus retained by subordi-
nates in the ?rst half of the experiment did not differ signi?-
cantly from the surplus retained by subordinates in the
second half of the experiment (the largest difference is less
than 4%, and the smallest p > 0.15). Thus, the effects of the
rankings neither increased nor dissipated over time.
Additional analysis – subordinates’ strategies
Finally, we examine subordinates’ reporting strategies
across ranking conditions. We classify subordinates who
claim between: (1) 67% and 100% of the surplus as primar-
ily focused on maximizing their own compensation; (2)
33% and 67% of the surplus as primarily focused on a more
equal (equitable) split; and (3) 0% and 33% of the surplus as
primarily focused on ?rm pro?t. Using these three catego-
ries, Fig. 2 shows subordinates’ strategies in each of the
four ranking conditions.
Fig. 2 reveals that the strategies chosen by subordinates
across the four ranking conditions vary markedly. In the
random ranking condition, 56% of the subordinates are fo-
cused on their own compensation, 39% are focused on a
more equal split, and only 6% are focused on ?rm pro?t.
When rankings are based on own compensation, the per-
centage of subordinates in these categories is 83%, 17%,
and 0%, respectively. A striking result in the own compen-
sation ranking condition is that no subordinates are de-
?ned as ?rm-focused.
In the ?rm pro?t ranking condition, subordinates were
split evenly across the three categories. This condition also
had the highest amount of subordinates who were classi-
?ed as ?rm-focused. In the both rankings condition, 33%
of the subordinates are focused on their own compensa-
tion, 56% are focused on a more equal split of the surplus,
and 11% are focused on ?rm pro?t. Collectively, these re-
sults show how subordinates’ reporting decisions were af-
fected by our ranking conditions. Moreover, these results
reveal why we observe differences in the levels of honesty
among our ranking conditions.
Conclusion
In this study, we examine whether rankings affect the
honesty of subordinates’ budget requests. In our experi-
ment, subordinates are requested to truthfully report a
project’s actual cost. Subordinates, however, have ?nancial
We classify subordinates’ strategies using three categories based on the percentage of surplus subordinates retained.
Specifically, we classify subordinates who claim between: (1) 67% and 100% of the surplus as primarily focused on
maximizing their own compensation; (2) 33% and 67% of the surplus as primarily focused on a more equal
(equitable) split; and (3) 0% and 33% of the surplus as primarily focused on firm profit.
Fig. 2. Subordinates’ strategies by ranking condition.
244 J.L. Brown et al. / Accounting, Organizations and Society 39 (2014) 237–246
incentives to maximally misrepresent the project’s true
cost and, as such, claim 100% of the project’s surplus. We
rank subordinates in one of four ways: based on (1) own
compensation, (2) ?rm pro?t, (3) both own compensation
and ?rm pro?t, and (4) randomly. None of the rankings af-
fect subordinates’ remuneration.
We ?nd that subordinates’ budget requests (honesty)
are signi?cantly affected by the rankings. Subordinates
ranked on ?rm pro?t claim the least surplus (45%), fol-
lowed by subordinates ranked on both individual compen-
sation and ?rm pro?t (55.6%), subordinates ranked
randomly (68%) and, ?nally, subordinates ranked on own
compensation (84%). Moreover, the surplus claimed by
subordinates in the both rankings condition is signi?cantly
lower than the surplus claimed by subordinates in the own
compensation condition but not signi?cantly different
from the surplus claimed by subordinates in the ?rm pro?t
condition. We also ?nd that our results are stable across
periods, suggesting that the ranking effects we observe
neither increase nor dissipate over time.
Our results have several important implications. First,
complementing extant research that documents the effects
of rankings on effort and performance (e.g., Frederickson,
1992; Hales et al., 2012; Hannan et al., 2008, 2013; Tafkov,
2013), we show how rankings signi?cantly affect another
variable vital to organizational success, managerial hon-
esty. Collectively, our ?ndings suggest that organizations
can use ?rm-based rankings as a low-cost, informal control
to constrain opportunistic reporting. Our results also
suggest, however, that compensation-, or even perqui-
site-, based rankings could lead employees to engage in
dishonest reporting and, as a result, ?rms should consider
carefully how they rank their employees and also be aware
of how employees may use accounting information to rank
themselves.
Second, we examine a setting in which subordinates re-
ceive two con?icting rankings. We ?nd that subordinates
in this condition respond more to the ?rm pro?t ranking
than to the own compensation ranking. This result is some-
what counter-intuitive, as one might reasonably posit that
having both rankings would lead subordinates to engage in
more self-regarding behavior (because subordinates could
internally justify their reporting decisions by focusing on
the own compensation ranking). This particular ?nding
suggests that, as long as ?rm-based rankings are provided,
the control loss associated with having access to a ranking
that encourages self-interested behavior may not be as
large as expected.
Certain limitations of our study provide opportunities
for further inquiry. For example, we chose to have a real
participant play the role of the superior because account-
ing research examining preferences for honesty has used
real participants for both subordinate and superior roles
and because this setting most closely mirrors the natural
environment where one’s actions invariably affect the wel-
fare of others. However, as noted by an anonymous Re-
viewer, this design choice may have prodded participants
receiving both the ?rm pro?t and own compensation rank-
ings to, for other-regarding reasons, focus more on the ?rm
pro?t ranking than on the own compensation ranking. Fu-
ture research could examine the robustness of our results
to this speci?c design choice, either by having a hypothet-
ical superior and/or by examining a setting where the ac-
tions of the subordinate have a more tenuous/indirect
relation to the superior’s remuneration.
As suggested by another anonymous Reviewer, future
research also could examine how cardinal ranking infor-
mation affects the veracity of employees’ reports. In our
experiment, subordinates were only provided with ordinal
ranks. We made this design choice so as not to provide
subordinates in our baseline (random rankings) setting
information that could be used to rank themselves. We
also made this design choice to examine the starkest form
of rankings. That said, employees frequently know the le-
vel of a performance variable (e.g., the amount of ?rm prof-
it) and, as such, could use this information to rank
themselves. Moreover, employees often are provided with
a formal rank coupled with the level of performance. Re-
search that extends our work to settings with different
types of ranking information and feedback could prove
quite fruitful.
References
Anderson, D., Crowell, C., Sponsel, S., Clarke, M., & Brence, J. (1982).
Behavior management in the public accommodations industry: A
three-project demonstration. Journal of Organizational Behavior
Management, 4, 33–66.
Arrow, K. J. (1974). Gifts and exchanges. Philosophy and Public Affairs I, 4,
343–362.
Azmat, G., & Iriberi, N. (2009). The provision of relative feedback
performance information: An experimental analysis of performance
and happiness. Working paper, Universitat Pompeu Fabra and
Barcelona GSE.
Baker, G. P., Jensen, M. C., & Murphy, K. J. (1988). Compensation and
incentives: Practice vs. theory. The Journal of Finance, 43, 593–616.
Beach, S., & Tesser, A. (1995). Self-esteem and the extended self-
evaluation maintenance model: The self in social context. In M.
Kernis (Ed.), Ef?cacy, agency, and self-esteem (pp. 145–170). New York:
NY Plenum Press.
Bebchuk, L. A., & Fried, J. M. (2003). Executive compensation as an agency
problem. Journal of Economic Perspectives, 17, 71–92.
Bick, I., Palmon, O., & Wald, J. (2006). CEO compensation, director
compensation, and ?rm performance: Evidence of cronyism? Journal
of Corporate Finance, 12, 403–423.
Blanchard, O. J., Lopez-de-Silanes, F., & Shleifer, A. (1994). What do ?rms
do with cash windfalls? Journal of Financial Economics, 36, 337–360.
Blanes i Vidal, J., & Nossol, M. (2011). Tournaments without prizes:
Evidence from personnel records. Management Science, 5, 1721–1736.
Buunk, A. P., & Gibbons, F. X. (2007). Social comparison: The end of a
theory and the emergence of a ?eld. Organizational Behavior and
Human Decision Processes, 102, 3–21.
Buunk, B. P., Kuyper, H., & van der Zee, Y. G. (2005). Affective response to
social comparison in the classroom. Basic and Applied Social
Psychology, 27, 229–237.
Cai, T., Jo, H., & Pan, C. (2011). Vice or virtue? The impact of corporate
social responsibility on executive compensation. Journal of Business
Ethics, 104, 159–173.
Church, B. K., Hannan, R. L., & Kuang, X. (2012). Shared interest and
honesty in budget reporting. Accounting, Organizations and Society, 37,
156–167.
Cialdini, R., Reno, R., & Kallgren, C. (1990). A focus theory of normative
conduct: Recycling the concept of norms to reduce littering in public
places. Journal of Personality and Social Psychology, 58, 1015–1026.
Cookson, R. (2000). Framing effects in public goods experiments.
Experimental Economics, 3, 55–79.
Core, J. R., & Larker, D. (1999). Corporate governance, chief executive
of?cer compensation and ?rm performance. Journal of Financial
Economics, 51, 371–406.
Dana, J., Weber, R. A., & Kuang, J. X. (2007). Exploiting moral wiggle room:
Experiments demonstrating an illusory preference for fairness.
Economic Theory, 33, 67–80.
Duffy, J., & Kornienko, T. (2010). Does competition affect giving? Journal of
Economic Behavior & Organization, 74, 82–103.
J.L. Brown et al. / Accounting, Organizations and Society 39 (2014) 237–246 245
Evans, J. H., III, Hannan, R. L., Krishnan, R., & Moser, D. V. (2001). Honesty
in managerial reporting. The Accounting Review, 76, 537–559.
Festinger, L. (1954). A theory of social comparison process. Human
Relations, 7, 117–140.
Fischbacher, U. (2007). Z-Tree: Zurich toolbox for ready-made economic
experiments. Experimental Economics, 10, 171–178.
Fisher, J. G., Maines, L. A., Peffer, S. A., & Sprinkle, G. B. (2002). Using
budgets for performance evaluation: Effects of resource allocation
and horizontal information asymmetry on budget proposals, budget
slack, and performance. The Accounting Review, 77, 847–865.
Frederickson, J. R. (1992). Relative performance information: The effects
of common uncertainty and contract type on subordinate effort. The
Accounting Review, 67, 647–669.
Goldstein, N. J., & Cialdini, R. B. (2007). The spyglass self: A model of
vicarious self-perception. Journal of Personality and Social Psychology,
92, 402–417.
Grant, A. (2013). Givers take all: The hidden dimension of corporate
culture. McKinsey Quarterly (April).
Gruder, C. L. (1971). Determinants of social comparison choices. Journal of
Experimental Social Psychology, 7, 473–489.
Haisley, E. C., & Weber, R. A. (2010). Self-serving interpretations of
ambiguity in other-regarding behavior. Games and Economic Behavior,
68, 614–625.
Hales, J., Hobson, J. H., & Resutek, R. J. (2012). The dark side of socially
mediated rewards: How narcissism and social status affect
managerial reporting. Working paper, University of Illinois at
Urbana-Champaign.
Hannan, L. R., Krishnan, R., & Newman, A. H. (2008). The effects of
disseminating relative performance feedback in tournament and
individual performance compensation plans. The Accounting Review,
83, 893–913.
Hannan, L. R., McPhee, G. P., Newman, A. H., & Tafkov, I. D. (2013). The
effect of relative performance information on effort allocation and
performance in a multi-task environment. The Accounting Review, 88,
553–575.
Hannan, L. R., Rankin, F. W., & Towry, K. L. (2006). The effect of
information systems on honesty in managerial reporting: A
behavioral perspective. Contemporary Accounting Research, 23,
885–918.
Huguet, P., Dumas, F., Monteil, J. M., & Genestoux, N. (2001). Social
comparison choices in the classroom: Further evidence for
students’ upward comparison tendency and its bene?cial impact
on performance. European Journal of Social Psychology, 31, 557–
578.
Hyman, H. (1942). The psychology of subjective status. Psychological
Bulletin, 39, 473–474.
Kramer, R. M., & Brewer, M. B. (1984). Effects of group identity on
resource use in a simulated commons dilemma. Journal of Personality
and Social Psychology, 46, 1044–1057.
Krupka, E., & Weber, R. A. (2009). The focusing and informational effects
of norms on pro-social behavior. Journal of Economic Psychology, 30,
307–320.
Lazarus, R. (1991). Emotion and adaption. NewYork, NY: Oxford University
Press.
Metee, D. R., & Smith, G. (1977). Social comparison and interpersonal
attraction: The case for dissimilarity. In J. M. Sulls & R. L. Miller (Eds.),
Social comparison processes: Theoretical and empirical perspectives
(pp. 69–101). Washington: Hemisphere.
Murphy, K. R. (1993). Honesty in the workplace. Pac?c Grove, CA: Brooks/
Cole.
Nordstrom, R., Lorenzi, P., & Hall, R. (1990). A review of public posting
of performance feedback. Journal of Organizational Behavior
Management, 11, 101–123.
Noreen, E. (1988). The economics of ethics: A new perspective on agency
theory. Accounting, Organizations and Society, 13, 359–369.
Nosanchuk, T. A., & Erickson, B. H. (1985). How high is up? Calibrating
social comparison in the real world. Journal of Personality and Social
Psychology, 48, 624–634.
Rankin, F. W., Schwartz, S. T., & Young, R. A. (2003). Management control
using nonbinding budgetary announcements. Journal of Management
Accounting Research, 15, 75–93.
Rankin, F. W., Schwartz, S. T., & Young, R. A. (2008). The effect of honesty
and superior authority on budget proposals. The Accounting Review,
83, 1083–1099.
Rowe, C. (2004). The effect of accounting report structure and team
structure on performance in cross-functional teams. The Accounting
Review, 79, 1153–1180.
Schatzberg, J., & Stevens, D. E. (2008). Public and private forms of
opportunism with the organization: A joint examination of budget
and effort behavior. Journal of Management Accounting Research, 20,
59–81.
Smith, R. H. (2000). Assimilative and contrastive emotional reactions to
upward and downward social comparisons. In J. Suls & L. Wheelers
(Eds.), Handbook of social comparison: Theory and research
(pp. 173–200). New York: Plenum.
Suls, J. M., & Miller, R. L. (1977). Social comparison processes: Theoretical
and empirical perspectives. Washington, DC: Hemisphere.
Tafkov, I. D. (2013). Private and public relative performance information
under different compensation contracts. The Accounting Review, 88,
327–350.
Tesser, A. (1988). Toward a self-evaluation maintenance model of social
behavior. Advances in Experimental Social Psychology, 21, 181–228.
Tesser, A., & Campbell, J. (1980). Self-de?nition: The impact of the relative
performance and similarity of others. Social Psychology Quarterly, 43,
341–347.
Tosi, H. L., Wermer, S., Katz, J. P., & Gomez-Mejia, L. R. (2000). How much
does performance matter? A meta-analysis of CEO pay studies.
Journal of Management, 26, 301–339.
Towry, K. L. (2003). Control in a teamwork environment – The impact of
social ties on the effectiveness of mutual monitoring contracts. The
Accounting Review, 78, 1069–1095.
West, K. L., & Mykerezi, E. (2011). Teachers’ unions and compensation:
The impact of collective bargaining on salary schedules and
performance pay schemes. Economics of Education Review, 30, 99–108.
Wheeler, L. (1966). Motivation as a determinant of upward comparison.
Journal of Experimental Social Psychology, 1, 27–31.
Wheeler, L., Koestner, R., & Driver, R. E. (1982). Related attributes in the
choice of comparison others: It’s there, but it isn’t all there is. Journal
of Experimental Social Psychology, 18, 489–500.
Wheeler, L., Shaver, K. G., Jones, R. A., Goethals, G. R., Cooper, J., Robinson,
J. E., et al. (1969). Factors determining choice of a comparison other.
Journal of Experimental Social Psychology, 5, 219–232.
Wikoff, M., Anderson, D., & Crowell, C. (1982). Behavior management in a
factory setting: Increasing work ef?ciency. Journal of Organizational
Behavior Management, 4, 97–127.
Wood, J. V. (1989). Theory and research concerning social comparisons of
personal attributes. Psychological Bulletin, 106, 231–248.
Zhang, M. (2008). The effects of perceived fairness and communication on
honesty and collusion in a multi-agent setting. The Accounting Review,
83, 1125–1146.
246 J.L. Brown et al. / Accounting, Organizations and Society 39 (2014) 237–246
doc_922341880.pdf
We conduct an experiment to investigate the effect of rankings, which are pervasive in
practice, on the honesty of managers’ budget reports, which is important for sound
decision making in organizations. Participants in our experiment are ranked in one of four
ways: (1) firm profit, (2) own compensation, (3) both firm profit and own compensation,
and (4) randomly, which serves as our baseline condition. None of the rankings affect
participants’ remuneration. Compared to our baseline (random rankings) setting, where
participants indeed exhibit honesty concerns, we find that rankings based on firm profit
significantly increase honesty and that rankings based on own compensation significantly
decrease honesty. Participants who received both rankings were significantly more honest
than participants in the own compensation rankings condition. We did not, however, find
significant differences in honesty between the both rankings and firm profit rankings
conditions. As such, participants in the both rankings condition seemed to focus more on
the firm profit metric than on the financially congruent own compensation metric. We also
find that our results are stable across periods, suggesting that the effects of rankings
neither increased nor dissipated over time
The effect of rankings on honesty in budget reporting
q
Jason L. Brown
a,?
, Joseph G. Fisher
a
, Matthew Sooy
b
, Geoffrey B. Sprinkle
a
a
Kelley School of Business, Indiana University, 1309 E. Tenth Street, Bloomington, IN 47405, United States
b
Gatton College of Business & Economics, University of Kentucky, 550 S. Limestone Street, Lexington, KY 40502, United States
a b s t r a c t
We conduct an experiment to investigate the effect of rankings, which are pervasive in
practice, on the honesty of managers’ budget reports, which is important for sound
decision making in organizations. Participants in our experiment are ranked in one of four
ways: (1) ?rm pro?t, (2) own compensation, (3) both ?rm pro?t and own compensation,
and (4) randomly, which serves as our baseline condition. None of the rankings affect
participants’ remuneration. Compared to our baseline (random rankings) setting, where
participants indeed exhibit honesty concerns, we ?nd that rankings based on ?rm pro?t
signi?cantly increase honesty and that rankings based on own compensation signi?cantly
decrease honesty. Participants who received both rankings were signi?cantly more honest
than participants in the own compensation rankings condition. We did not, however, ?nd
signi?cant differences in honesty between the both rankings and ?rm pro?t rankings
conditions. As such, participants in the both rankings condition seemed to focus more on
the ?rm pro?t metric than on the ?nancially congruent own compensation metric. We also
?nd that our results are stable across periods, suggesting that the effects of rankings
neither increased nor dissipated over time. We discuss the contributions of our study
and concomitant ?ndings to accounting research and practice.
Ó 2014 Elsevier Ltd. All rights reserved.
Introduction
Managers routinely receive feedback from the
accounting system regarding ?rm-performance variables
under their control and also, of course, their own
compensation. In turn, managers frequently are ranked
on such variables or can infer their rankings from available
information. While such rankings may be used to
determine compensation, retention, and promotion, they
also may have little bearing on managers’ remuneration
(e.g., Nordstrom, Lorenzi, & Hall, 1990; West & Mykerezi,
2011) or may even con?ict with managers’ personal ?nan-
cial incentives (Grant, 2013). Moreover, ample research
suggests that individuals’ concerns for rank affect their
behavior. Research in accounting, for example, has
documented positive (e.g., Frederickson, 1992; Tafkov,
2013) and negative (e.g., Hannan, Krishnan, & Newman,
2008; Hannan, McPhee, Newman, & Tafkov, 2013) effects
of rankings on effort and concomitant task performance.
Although the use of rankings is pervasive in practice,
prior research has not examined whether rankings affect
the honesty of managers’ reports. For a myriad of reasons,
honest reporting by ?rm participants is important for
sound decision making. Indeed, prior research in account-
ing documents that individuals not only have preferences
for honesty (Evans, Hannan, Krishnan, & Moser, 2001;
Rankin, Schwartz, & Young, 2003) but also that individuals’http://dx.doi.org/10.1016/j.aos.2014.03.001
0361-3682/Ó 2014 Elsevier Ltd. All rights reserved.
q
We thank Spencer Anderson, Jason Kuang, Victor Maas, Adam
Presslee, Mike Shields, Laura Wang, two anonymous Reviewers, and
participants at the 2013 American Accounting Association Management
Accounting Section Meeting for their very helpful comments and
suggestions. We also thank the Marketing Department at Indiana
University for the use of its laboratory.
?
Corresponding author. Tel.: +1 812 855 2381; fax: +1 812 855 4985.
E-mail addresses: [email protected] (J.L. Brown), jo?sher@
indiana.edu (J.G. Fisher), [email protected] (M. Sooy), sprinkle@
indiana.edu (G.B. Sprinkle).
Accounting, Organizations and Society 39 (2014) 237–246
Contents lists available at ScienceDirect
Accounting, Organizations and Society
j our nal homepage: www. el sevi er. com/ l ocat e/ aos
preferences for honesty are affected by fairness concerns
(Zhang, 2008), group incentives (Church, Hannan, & Kuang,
2012), and speci?c features of the managerial setting, such
as the precision of the information system (Hannan,
Rankin, & Towry, 2006) and who has budget authority
(Rankin, Schwartz, & Young, 2008).
In this paper, we examine whether rankings affect the
veracity of individuals’ reports. We employ a participative
budgeting setting similar to Evans et al. (2001) in which
managers (hereafter subordinates) receive private cost
information and then submit a budget request to a
superior. The budget request splits the available surplus be-
tween the subordinate and the superior. Subordinate com-
pensation is based on the budget request, and the
subordinate has strict ?nancial incentives to request the
highest budget amount. The superior’s residual claim (?rm
pro?t), however, is minimized by this subordinate
strategy. Consistent with Evans et al. (2001), the subordi-
nate is requested to provide an accurate (honest) budget re-
quest. Increased subordinate budget accuracy (honesty)
results in more surplus being allocated to the ?rm and less
to the subordinate.
In our setting, six subordinates, which constitute a co-
hort, are ranked in one of four ways: (1) subordinate com-
pensation; (2) superior residual claim (?rm pro?t); (3)
both subordinate compensation and ?rm pro?t; and, (4)
randomly. Rankings based on subordinate compensation
are congruent with subordinates’ ?nancial incentives to
maximize the surplus claimed but are not aligned with
honest reporting. Rankings based on ?rm pro?t are
congruent with honest reporting but at odds with subordi-
nates’ ?nancial incentives. Rankings based on both
subordinate compensation and ?rm pro?t provide subordi-
nates one metric that is congruent with subordinates’
?nancial incentives and one metric that is at odds with
subordinates’ ?nancial incentives. Random rankings are
not affected in any way by subordinates’ budget requests
and, as such, are not under subordinates’ control.
1
We ?nd that rankings signi?cantly affect managerial
reporting honesty. Compared to our baseline setting,
where subordinates indeed exhibit honesty concerns, we
?nd that rankings based on ?rm pro?t increase subordi-
nates’ honesty. This ?nding suggests that ?rms can use
rankings as a low-cost, informal control to constrain
opportunistic reporting. Incorporating rankings into the
design of control systems could allow ?rms to reduce
information asymmetry between employees and the ?rm,
which leads to better decision making and lower budget-
ary slack, both of which should increase ?rm pro?tability.
Moreover, this result suggests that, for example, the ?rm-
level rankings frequently published in the popular press
may curb self-interested behavior in organizations.
In contrast, rankings based on own compensation de-
crease subordinates’ honesty. This ?nding suggests that
providing relative performance feedback to employees
can lead to behavior that is harmful to the ?rm. As a result,
?rms should consider carefully how they rank their
employees and also be aware of how employees may use
accounting information to rank themselves. This may help
explain why some ?rms closely guard their compensation
data and are loathe to share such information with employees.
Subordinates who received both rankings were more
honest than subordinates in the own compensation condi-
tion. Moreover, we do not ?nd signi?cant differences in
honesty between the both rankings and ?rm pro?t condi-
tions. As such, subordinates in the both rankings condition
seemed to focus more on the ?rm pro?t metric than on the
?nancially congruent own compensation metric. This re-
sult is somewhat surprising as, intuitively, one might posit
that having both rankings would lead subordinates to en-
gage in more self-regarding behavior because subordinates
could internally justify their reporting decisions by focus-
ing on the compensation metric. Our ?ndings fromthis set-
ting suggest that as long as ?rm-based rankings are
provided the control loss associated with having access
to a ranking that encourages self-interested behavior may
not be as large as expected. Finally, we ?nd that our results
are stable across periods, suggesting that the effects of
rankings neither increase nor dissipate over time.
The remainder of this paper is organized into four sec-
tions. The next section develops the hypotheses, and the
third section presents the experimental design. The fourth
section reports the results, and the ?nal section provides a
summary of the study.
Background and hypotheses
Research setting
Accurate reporting by employees is important for mak-
ing sound organizational decisions. To this end, Evans et al.
(2001) employed a novel design to examine subordinates’
honesty in a managerial reporting setting. In this setting,
a subordinate, who has private information regarding the
production cost of an investment project, submits a budget
request to a superior who funds the project. Both the rev-
enue and the probabilistic distribution of the project’s cost
are known by the subordinate and the superior. Only the
subordinate, however, is aware of the actual project cost.
The subordinate’s budget request is always approved and
the subordinate receives the difference between the bud-
get request and the actual project cost. The superior re-
ceives the residual, which equals the difference between
the project’s revenue and the budget request. In this set-
ting, the subordinate is requested to accurately report the
project cost via the budget request.
Assuming subordinates are only concerned with wealth
maximization, conventional economic analysis predicts
that subordinates will report the highest cost possible
and claim the entire budget surplus. Evans et al. (2001)
?nd, however, that subordinates’ preferences for honesty
attenuate personal wealth aspirations – subordinates’ re-
ports are signi?cantly more truthful than would be pre-
dicted by conventional economic analysis. Research in
accounting has extended this result and ?nds that individ-
uals not only have preferences for honesty but also that
individuals’ preferences for honesty are affected by a
1
Random rankings serve as our baseline condition. We employ random,
rather than no, rankings in our baseline condition to ensure that partic-
ipants in all conditions have rankings information and, thus, to isolate the
honesty effects associated with how participants are ranked.
238 J.L. Brown et al. / Accounting, Organizations and Society 39 (2014) 237–246
number of factors, including fairness concerns, group
incentives, the precision of the information system, and
who has budget authority (Church et al., 2012; Hannan
et al., 2006; Rankin et al., 2008; Zhang, 2008).
Adding to this research stream, we examine whether
rankings affect the honesty of subordinates’ budget re-
quests. We augment the Evans et al. (2001) setting by pair-
ing each subordinate with an actual superior and by
ranking subordinates within cohort groups. A cohort group
consists of six subordinates, who are ranked from highest
to lowest in one of four ways. None of the rankings affect
subordinates’ remuneration. Additionally, subordinates
were only provided with the ordinal ranks. As such, while
each subordinate certainly knows his/her actual budget re-
quest, s/he does not knowthe actual budget requests of the
other subordinates.
2
First, we examine a setting where subordinates are
ranked each period from highest to lowest based on their
own compensation.
3
In this setting, wealth and rank con-
cerns are aligned – subordinates maximize both their wealth
and their rank by submitting the highest possible budget re-
quest. Preferences for honesty are, however, at odds with
preferences for wealth and rank in this setting.
Second, we examine a setting where subordinates are
ranked each period from highest to lowest based on ?rm
pro?t.
4
In this setting, wealth and rank concerns are at odds
– subordinates maximize their wealth (rank) by submitting
the highest (lowest) possible budget request. Moreover,
preferences for honesty are congruent with preferences for
rank in this setting.
Third, we examine a setting where subordinates are
ranked each period from highest to lowest based on (a)
their own compensation and (b) ?rm pro?t. Thus, in this
setting subordinates receive two con?icting rankings, with
each ranking containing the characteristics and tensions
from our ?rst two settings. Finally, we examine a setting
where subordinates are randomly ranked each period from
highest to lowest. This setting serves as our control (base-
line) treatment condition.
Social comparison theory and prior research
Human beings are a social species, and it is widely ac-
cepted that we desire to compare ourselves to others
(Buunk & Gibbons, 2007). Festinger (1954) formalized
these thoughts and introduced social comparison theory,
which posits that individuals have a desire to compare
themselves favorably to their peers.
5
Metee and Smith
(1977) propose that social comparison theory is about our
desire to know ourselves. This desire to know ourselves is
not only satis?ed by receiving information regarding our
performance but also by comparing our performance with
others.
Rankings enable individuals to make interpersonal
comparisons which, in turn, lead to a desire to improve
one’s rank and competition. Indeed, there is considerable
research consistent with social comparisons affecting indi-
viduals’ behavior. Speci?cally, research from experiments
(Duffy & Kornienko, 2010; Fisher, Maines, Peffer, & Sprin-
kle, 2002; Frederickson, 1992; Hales, Hobson, & Resutek,
2012; Hannan et al., 2008, 2013; Tafkov, 2013) and the
?eld (Anderson, Crowell, Sponsel, Clarke, & Brence, 1982;
Azmat & Iriberi, 2009; Blanes i Vidal & Nossol, 2011; Nord-
strom et al., 1990; Wikoff, Anderson, & Crowell, 1983) both
support social comparison theory and ?nd that rankings
affect behavior in a variety of settings.
Social comparison theory (Festinger, 1954) provides a
useful framework for understanding the relationship be-
tween rankings and behavior. The effects of rankings have
been attributed to many factors. For example, rankings can
create descriptive norms of appropriate behavior, a way for
individuals to improve themselves, help foster/maintain a
positive self-image, lead to feelings of pride, or help avoid
negative feelings such as shame (see, e.g., Beach & Tesser,
1995; Goldstein & Cialdini, 2007; Lazarus, 1991; Smith,
2000; Tesser, 1988; Tesser & Campbell, 1980; Wood,
1989).
A consistent ?nding from social comparison studies
that is particularly germane to our study is that individuals
frequently exhibit an oft-referred to ‘‘upward drive’’ moti-
vation to exceed others’ performance (Festinger, 1954; Suls
& Miller, 1977). Speci?cally, prior research ?nds that indi-
viduals compare themselves to others whose performance
is better than their performance (Nosanchuk & Erickson,
1985; Wheeler, Koestner, & Driver, 1982; Wood, 1989).
For example, several studies ?nd that individuals tend to
compare themselves to individuals who are ranked above
them (Gruder, 1971; Wheeler, 1966; Wheeler et al., 1969).
Prior research, in turn, ?nds that these upward compar-
isons frequently lead individuals to increase their effort,
performance, or rank (Buunk, Kuyper, & van der Zee,
2005; Huguet, Dumas, Monteil, & Genestoux, 2001). For
example, in a study that is related to ours, Duffy and
Kornienko (2010) rank proposers in a dictator game by
the amount of money they give to recipients. Duffy and
Kornienko (2010) ?nd that ranking participants in this
fashion affects the amount of charitable giving. Moreover,
Duffy and Kornienko (2010) suggest that the difference in
giving behavior is driven, in part, by competition among
participants.
6
2
We made this design choice so as not to provide subordinates in our
baseline (random rankings) setting information that could be used to rank
themselves. We also made this design choice to examine the starkest form
of rankings.
3
As such, subordinates submitting the highest (lowest) budget request
would receive the #1 (#6) rank.
4
As such, subordinates submitting the lowest (highest) budget request
would receive the #1 (#6) rank.
5
Although it is Festinger’s (1954) classic paper that details speci?c
hypotheses regarding social comparison theory, the idea of social compar-
ison and reference groups are also highlighted in earlier research. For
example, Hyman (1942) ?nds that the assessment of one’s status regarding
?nances, intelligence, and physical attractiveness depends on one’s com-
parison group.
6
While our work shares similarities with previous social comparison
studies in accounting and other areas, there is a fundamental difference. To
obtain a higher payoff in our study, participants must misreport the true
cost, which is a feature not present in prior rank-based research. Given that
ethical systems encourage honesty (Murphy, 1993) and that honesty can
increase ef?ciencies and surplus (Arrow, 1974; Noreen, 1988), it is
important to understand how rankings affect individuals’ reporting
decisions.
J.L. Brown et al. / Accounting, Organizations and Society 39 (2014) 237–246 239
Thus, social comparison research not only suggests
that individuals will make upward comparisons but also
suggests that, as a result of these comparisons, individu-
als will take actions to improve their performance to
ultimately achieve a more favorable comparison (rank-
ing). In a similar vein, we posit that individuals will at-
tend to the rankings provided to them and exhibit an
‘‘upward drive’’ motivation to improve their rankings.
As such, we posit that individuals will change their
reporting behavior (honesty) to improve their ranking
among their peers.
Hypotheses 1 and 2
In our setting, subordinates’ ranks are not tied to their
compensation so, as a result, conventional economic the-
ory predicts that subordinates will ignore the rankings
and submit the maximum budget request.
7
Social compar-
ison theory and prior empirical research suggests, however,
that rankings will in?uence subordinates’ reporting behavior
due to their desire to improve their ranking. That is, in the
own compensation (?rm pro?t) ranking condition, this
means that subordinates will submit higher (lower) budget
requests than if subordinates are randomly ranked.
8
This
leads to our two hypotheses:
H1. Subordinates who are ranked based on their
own compensation will submit higher budget requests
(i.e., be less honest) than subordinates who are randomly
ranked.
H2. Subordinates who are ranked based on ?rm pro?t will
submit lower budget requests (i.e., be more honest) than
subordinates who are randomly ranked.
Research question – multiple rankings
In many settings, individuals are ranked on multiple
metrics. For example, popular business periodicals
frequently rank executives and ?rms on factors such
as executive compensation, ?rm pro?tability, and cor-
porate social responsibility. And, the relations among
these variables may be positive, non-existent, or
negative.
9
Likewise, companies often explicitly or implicitly rank
employees on metrics that are either not linked, or may
even con?ict with, their remuneration. For example,
companies such as Google, Southwest Airlines, and Zappos
provide employees with public relative performance infor-
mation regarding their organizational helping behaviors
(Grant, 2013). These companies do not, however, directly
compensate such behaviors, and it is unclear whether such
behaviors are ultimately linked to compensation (e.g., ‘‘no
good deed goes unpunished’’).
Additionally, many ?rms utilize ‘‘dashboards’’ that are
frequently updated to show managers’ ongoing perfor-
mance across several measures that are not necessarily
correlated or may even con?ict (e.g., average waiting/ser-
vice time versus safety/cleanliness at a fast-food restau-
rant). These dashboards, as exist at companies such as
McDonald’s and eBay, can serve as rankings by presenting
peer-manager performance information. These dashboards
also enable managers to ‘‘drill down’’ and/or query others’
performance so as to evaluate their relative performance.
Along these lines, we examine a setting where subordi-
nates are ranked each period from highest to lowest based
on (a) their own compensation and (b) ?rm pro?t. Thus, in
this setting subordinates receive two con?icting rankings.
In such settings it is unclear what rank(s), if any, to which
individuals will attend.
Some research suggests that individuals may use ‘‘mor-
al wiggle room’’ to act in their self-interest when they have
con?icting rankings (Dana, Weber, & Kuang, 2007; Haisley
& Weber, 2010). This line of research suggests that provid-
ing subordinates with both rankings may lead subordi-
nates to engage in more ?nancial self-regarding behavior
because subordinates can internally justify their reporting
decisions via the own compensation ranking. In this sce-
nario, subordinates’ budget requests (honesty) would most
closely resemble subordinates’ budget requests who re-
ceive rankings based on their own compensation.
In contrast, another research stream suggests that indi-
viduals’ behavior may be affected by drawing their atten-
tion to pro-social behavior (Cialdini, Reno, & Kallgren,
1990; Krupka & Weber, 2009).
10
For example, Krupka and
Weber (2009) examine a binary dictator game with a
pro-social allocation and a sel?sh allocation and ?nd that
7
This is also a noteworthy difference between our study and some prior
research. By not linking ranks to remuneration, we are able to examine
whether rankings per se affect behavior. This distinction is important
because, when compensation is linked to ranks, it is unclear whether
individuals are responding to the rank and/or to the reward. For example,
individuals working under a tournament contract may ramp-up their effort
if they believe they have the ability to win the prize. Analogously,
individuals working under a tournament contract may provide low effort
if they assess their chances of winning to be low. In both cases, the
compensation-based effect (related to, e.g., effort and/or skill) may swamp
the rank effect. Moreover, rankings based on, for example, corporate social
responsibility are not necessarily linked to employee remuneration in
natural settings (cf. footnote 9).
8
This means that subordinates in the own compensation (?rm pro?t)
ranking condition will be less (more) honest than subordinates who are
ranked randomly.
9
Research generally ?nds that, despite the strong predictions by
economic theory, compensation is often independent of performance
(Baker, Jensen, & Murphy, 1988). For example, a meta-analytic review of
the literature regarding the determinants of CEO pay ?nds that size
accounts for 40% of the variance in CEO pay and ?rm performance accounts
for less than 5% (Tosi, Wermer, Katz, & Gomez-Mejia, 2000). A number of
researchers suggest that CEO pay seems to re?ect managerial power and
rent-seeking by CEOs rather than the provision of ef?cient incentive
contracting (Bebchuk & Fried, 2003; Blanchard, Lopez-de-Silanes, & Shle-
ifer, 1994). Due to the other forces that determine CEO pay, some studies
even document a negative relation between compensation and perfor-
mance (Core & Larker, 1999; Bick, Palmon, & Wald, 2006). While there are
not many studies that examine the relation between CEO pay and corporate
social responsibility (CSR), at least one study ?nds a negative relation
between CEO pay and CSR (Cai, Jo, & Pan, 2011).
10
This is consistent with social comparison theory, which suggests that
rankings can provide descriptive norms of appropriate behavior.
240 J.L. Brown et al. / Accounting, Organizations and Society 39 (2014) 237–246
having participants focus on what they believe other people
should do in such a decision context resulted in more partic-
ipants choosing the pro-social allocation.
Additionally, research relating to groups (Cookson,
2000; Kramer & Brewer, 1984; Rowe, 2004; Towry, 2003)
suggests that the presence of a ranking that encourages
the consideration of others may induce subordinates to
act less self-interested. For example, Rowe (2004) ?nds
that aligning the accounting report structure and the team
structure in such a way to create a ‘‘group frame’’ helps
mitigate the free-rider problem.
11
Moreover, in this sce-
nario, subordinates’ budget requests (honesty) would most
closely resemble subordinates’ budget requests who receive
rankings based on ?rm pro?t.
Finally, the two opposing ranking measures may simply
‘‘cancel’’ each other out resulting in little or no impact. In
this scenario, subordinates’ budget requests would most
closely resemble the budget requests of subordinates
who receive random rankings. Given the dif?culty in
developing a directional prediction for our two rankings
condition, we examine the following research question:
RQ: Will subordinates who are ranked based on both
their own compensation and ?rm pro?t submit
budget requests that are different than the
budget requests submitted by subordinates who
are ranked on either their own compensation,
?rm pro?t, or randomly?
Experimental method
Task
We employ a budgeting setting similar to Evans et al.
(2001). In our setting, each participant plays the role of
either subordinate or superior.
12
The subordinate submits
a budget request to the superior to fund the cost of an
investment project. Both the revenue ($4.00) and the proba-
bilistic distribution of the project’s cost (uniformly distrib-
uted, in increments of $0.01, between $2.00 and $4.00) are
known by the subordinate and superior but only the subor-
dinate knows the actual project cost. The subordinate’s bud-
get request (maximum request = $4.00) is always approved,
and the subordinate receives as compensation the difference
between the budget request and the actual project cost. The
superior receives the residual pro?t, which equals the differ-
ence between the project’s revenue and the budget report.
Thus, if the project cost equals $3.00 and the subordinate
submits a budget request of $3.60, then the subordinate
earns $0.60 ($3.60–$3.00) and the superior earns $0.40
($4.00–$3.60). In this setting, subordinates who wish to
maximize their own compensation should, for every cost
realization, submit a budget request of $4.00.
Each subordinate/superior dyad is paired for the entire
experiment, which consisted of two practice periods and ten
compensated periods. Each experimental session comprised
12participants, or sixsubordinate/superior dyads. Ineachper-
iod, all six subordinates received identical project cost draws.
Manipulations
We manipulate, between-participants, the rankings
provided to subordinates. We examine four ranking condi-
tions. None of the rankings, per se, affected participants’
remuneration.
In the Firm Pro?t Condition, subordinates are ranked
each period from highest (?rst) to lowest (sixth) based
on the pro?t they generated for the superior. In the Own
Compensation Condition, subordinates are ranked each per-
iod from highest to lowest based on their individual com-
pensation. In the Both Condition, participants received
rankings based on ?rm pro?t and on their own compensa-
tion. Finally, in the Random Condition, subordinates are
ranked each period from highest to lowest based on a ran-
domly generated number.
13
Participants and procedures
We conducted a computer-based laboratory experiment
with 180 undergraduate students using z-tree software (Fis-
chbacher, 2007). We ran 15 sessions – three sessions for the
?rm pro?t, own compensation, and random conditions, and
six sessions for the both condition (three sessions where
the ?rmpro?t ranking was displayed ?rst, and three sessions
wheretheowncompensationrankingwas displayed?rst). To
ensure the privacy of participants, each subordinate was as-
signed a color identi?er (blue, green, orange, purple, red, or
yellow) throughout the experiment. The color identi?ers
wereusedwhendisplayingsubordinaterankings.
14
Bothsub-
ordinates and superiors received a $5.00 one-time participa-
tion fee and a salary of $0.50 per period ($5.00 in total).
At the beginning of each period, subordinates received
information about the project’s cost.
15
Each subordinate
then submitted a budget request to his/her superior. The
budget request could range from $0.50 below the project’s
11
Rowe (2004) manipulated the accounting report structure by showing
the individual manager’s payoffs or both the individual manager’s payoffs
and the payoffs for other managers.
12
We chose to have a real participant play the role of the superior
because this design choice most closely mirrors settings in the natural
environment where one’s actions invariably affect the welfare of others.
Additionally, accounting research examining preferences for honesty has
frequently employed real participants for both subordinate and superior
roles (see, e.g., Hannan et al., 2006; Rankin et al. 2003; Rankin et al. 2008;
Schatzberg & Stevens, 2008).
13
Notably, subordinates receiving ?rm pro?t rankings also have available
own compensation rankings, and subordinates receiving own compensa-
tion rankings also have available ?rm pro?t rankings. This obtains because,
for example, in the ?rm pro?t condition, the subordinate with highest ?rm
pro?t rank (#1) will have the lowest own compensation rank (#6), and so
on. Thus, by inverting the ?rm pro?t ranks, subordinates can obtain the
own compensation ranks. In a like fashion, subordinates in the own
compensation ranking condition can construct the ?rm pro?t ranks. As
such, our manipulations make salient certain aspects of the payoff
distribution as, informationally, subordinates in all ranking conditions
(other than the random condition) have available both ?rm pro?t and own
compensation ranks.
14
Each subordinate knew his/her own color but never knew which colors
corresponded to the other subordinates. The color was visible to superiors
and other subordinates when rankings were displayed.
15
Participants were informed that cost draws were determined in
advance and that the project cost in each period was the same for all
subordinates. To facilitate comparisons across treatment conditions, the
cost sequences in all sessions were identical.
J.L. Brown et al. / Accounting, Organizations and Society 39 (2014) 237–246 241
cost to $4.00. As in Evans et al. (2001), we allow underre-
porting. As subordinates received $0.50 in salary each peri-
od, we limited underreporting to $0.50 to ensure that
subordinates’ earnings were not negative in any period.
Subordinates’ budget requests were automatically ap-
proved. After all subordinates submitted their budget re-
quests, subordinates were publicly ranked from ?rst to
sixth as per our manipulations.
16
When subordinates’ bud-
get requests were identical, they received the same rank and
their order was randomly determined. Rankings were visible
to all subordinates and superiors and remained visible until
all subordinates indicated that they were ready to proceed
to the ensuing period. After the tenth period, subordinates
received a ?nal ranking based on their cumulative perfor-
mance over the experiment. Participants then completed a
post-experimental questionnaire and were paid in cash for
their participation. Average earnings were $15.00, and each
experimental session lasted approximately 45 min.
Results
Hypotheses 1 and 2
There is a ?xed amount of resources available each per-
iod to be split between the superior and the subordinate.
Panel B: Percentage of Surplus Retained by Subordinates by Ranking Condition and Period
Period
Ranking Condition n 1 2 3 4 5 6 7 8 9 10 Overall
Own Compensation 18 80.9% 81.0% 84.6% 81.0% 86.0% 88.2% 88.6% 90.7% 78.1% 80.5% 84.0%
Random 18 73.8% 67.1% 64.5% 66.0% 68.9% 63.0% 71.3% 65.9% 70.2% 68.9% 68.0%
Both Rankings 36 51.7% 65.0% 49.6% 33.9% 68.9% 50.7% 58.6% 59.1% 56.1% 62.7% 55.6%
Firm Profit 18 45.2% 46.5% 33.9% 44.1% 57.5% 44.7% 49.0% 58.5% 32.3% 38.6% 45.0%
The percentage of surplus kept = Surplus Claimed = (Budget Request – Project Cost)
Surplus Available (Revenue – Project Cost)
When subordinates reported less than the actual cost, the percentage of surplus retained is negative. To ensure that
(the few) negative reports were computed out of a possible 100%, we calculated the percentage of surplus retained
relative to the fixed wage. Specifically, we calculated the percentage of surplus retained for negative reports as:
(Budget Request – Project Cost) /Fixed Wage per Period. Measuring underreporting relative to the fixed salary is a
more appropriate benchmark because it avoids denominator distortions caused by cost draws that are dramatically
larger or smaller than the $0.50 salary constraint. For example, in a period where the actual project cost was $2.50, a
subordinate could report a cost as low as $2.00. As such, we calculate the misreporting as $0.50/$0.50 = 100%
(Maximum allowed).
Panel A: Percentage of Surplus Retained by Subordinates by Ranking Condition and Period
Fig. 1. Percentage of surplus retained by subordinates.
16
As mentioned earlier, only ordinal rankings were displayed – partic-
ipants did not receive any cardinal information. That is, the subordinate
only knew his/her own budget request and not the budget requests of the
other subordinates.
242 J.L. Brown et al. / Accounting, Organizations and Society 39 (2014) 237–246
We calculate the total surplus available as the resources in
excess of the actual cost. The subordinate’s budget request
determines how the surplus is split between the superior
and subordinate. As such, the surplus retained by subordi-
nates is our primary dependent variable.
Collectively, H1 and H2 predict that the surplus re-
tained by subordinates will be greatest when subordinates
are ranked on their own compensation and least when
subordinates are ranked on ?rm pro?t. We posit that the
surplus retained by subordinates who are ranked ran-
domly will be between the two. Fig. 1 presents the average
percentage of surplus retained by subordinates for each
period by ranking condition.
17
Consistent with H1 and H2,
the average percentage of surplus retained by subordinates
in the own compensation ranking condition (84%) is higher
than the average percentage of surplus retained by subordi-
nates in the random ranking condition (68%) which, in turn,
is higher than the average percentage of surplus retained by
subordinates in the ?rm pro?t ranking condition (45%). The
average percentage of surplus retained by subordinates in
the both rankings condition (55.6%) lies between the
random ranking and ?rm pro?t ranking conditions.
Panel A of Table 1 reports the results of an ANOVA with
the percentage of surplus retained by each subordinate per
period as the dependent variable, condition as the be-
tween-subjects factor, and period as the within-subjects
(repeated-measures) factor. The results of this ANOVA re-
veal signi?cant main effects of ranking condition
(p = 0.003) and period (p = 0.053), indicating that the per-
centage of surplus retained by subordinates differs signi?-
cantly across conditions and periods.
Panel B of Table 1 presents the pairwise comparisons
between all ranking conditions.
18
These comparisons pro-
vide support for H1 and H2. Panel B of Table 1 shows that
subordinates who were ranked on their own compensation
retained signi?cantly more surplus than subordinates who
were ranked randomly (p = 0.024). Panel B of Table 1 also
shows that subordinates who were ranked based on ?rm
pro?t retained signi?cantly less surplus than subordinates
who were ranked randomly (p = 0.051), providing support
for H2.
19
Collectively, our results indicate that, even though the
rankings did not affect remuneration in any way, subordi-
nates’ budget requests are signi?cantly in?uenced by the
ranking metrics. We ?nd that the percentage of surplus
claimed by subordinates almost doubles when rankings
are based on own compensation (84%) compared to the
condition where rankings are based on ?rm pro?t (45%).
Our ?ndings suggest that ranking metrics are powerful
motivators that ?rms should carefully consider when
designing their control systems. In particular, ?rms may
bene?t from ranking on ?rm-based metrics (e.g., pro?tabil-
ity-based metrics), but they may exacerbate control prob-
lems when ranking on individual-compensation metrics.
Research question – multiple rankings
Panels A and B of Fig. 1 show that the percentage of sur-
plus retained by subordinates receiving both rankings
(55.6%) is closer to the percentage of surplus retained by
subordinates receiving the ?rm pro?t ranking (45%) than
Table 1
ANOVA on percentage of surplus retained by subordinates.
Source df M.S. F-statistic p-Value
Panel A: Overall ANOVA
Between subjects
Condition 3 5.32 5.03 0.003
Participant 86 1.06
Within subjects
Period 9 0.17 1.87 0.053
Period  Condition 27 0.09 1.02 0.433
Residual 774 0.09
Total 899 0.20
Ranking conditions F-statistic p-Value
Panel B: Follow-up ANOVA’s – pairwise comparisons
Own compensation: Random 4.20 0.024
*
Firm pro?t: Random 2.81 0.051
*
Firm pro?t: Own compensation 8.82 0.003
*
Own compensation: Both rankings 14.85 0.001
Firm pro?t: Both rankings 0.97 0.330
Random: Both rankings 2.47 0.122
Panel A presents the repeated-measures ANOVA for the percentage of surplus retained by subordinates. In Panel B, we present the pairwise comparisons
between ranking conditions – for these tests, we replicate the repeated-measures ANOVA using the two conditions of interest.
*
One-tailed p-value.
17
An ANOVA revealed no signi?cant order effects in the both rankings
condition. That is, the percentage of surplus retained by subordinates did
not differ depending on whether subordinates received the own compen-
sation rank or the ?rm pro?t rank ?rst or second. As such, we do not
distinguish between these two treatment conditions.
18
In Panel B, we replicate the repeated-measures ANOVA in Panel A using
the two conditions of interest. None of the period-by-condition pairwise
interaction effects are signi?cant (smallest p > 0.19).
19
Panel B of Table 1 also reports that the surplus retained by subordinates
who were ranked on ?rm pro?t is less than (p = 0.003) the surplus retained
by subordinates who were ranked based on their own compensation.
J.L. Brown et al. / Accounting, Organizations and Society 39 (2014) 237–246 243
to the percentage of surplus retained by subordinates
receiving the own compensation ranking (84%). Pairwise
comparisons bear this observation out. As documented in
Panel B of Table 1, subordinates in the both rankings con-
dition submitted signi?cantly lower budget requests than
subordinates in the own compensation ranking condition
(p = 0.001). The budget requests of subordinates in the
both rankings condition, however, did not differ signi?-
cantly from the budget requests of subordinates in the ?rm
pro?t ranking condition (p = 0.330). This ?nding is some-
what unexpected as one might posit that having both rank-
ings would lead subordinates to anchor on the own
compensation ranking and internally ‘‘justify’’ their report-
ing decisions via this ranking metric.
Subordinates’ budget requests throughout the experiment
In our tests of H1 and H2, we ?nd that the period effect is
signi?cant in the overall ANOVA. This is attributable primar-
ily to the period-by-period reporting variance resulting from
the actual cost differing across periods rather than to report-
ing trends over time. To more closely examine the period ef-
fect, we compared the reporting behavior of subordinates in
the ?rst half of the experiment to the reporting behavior of
subordinates in the second half of the experiment. For each
of the ranking conditions, the surplus retained by subordi-
nates in the ?rst half of the experiment did not differ signi?-
cantly from the surplus retained by subordinates in the
second half of the experiment (the largest difference is less
than 4%, and the smallest p > 0.15). Thus, the effects of the
rankings neither increased nor dissipated over time.
Additional analysis – subordinates’ strategies
Finally, we examine subordinates’ reporting strategies
across ranking conditions. We classify subordinates who
claim between: (1) 67% and 100% of the surplus as primar-
ily focused on maximizing their own compensation; (2)
33% and 67% of the surplus as primarily focused on a more
equal (equitable) split; and (3) 0% and 33% of the surplus as
primarily focused on ?rm pro?t. Using these three catego-
ries, Fig. 2 shows subordinates’ strategies in each of the
four ranking conditions.
Fig. 2 reveals that the strategies chosen by subordinates
across the four ranking conditions vary markedly. In the
random ranking condition, 56% of the subordinates are fo-
cused on their own compensation, 39% are focused on a
more equal split, and only 6% are focused on ?rm pro?t.
When rankings are based on own compensation, the per-
centage of subordinates in these categories is 83%, 17%,
and 0%, respectively. A striking result in the own compen-
sation ranking condition is that no subordinates are de-
?ned as ?rm-focused.
In the ?rm pro?t ranking condition, subordinates were
split evenly across the three categories. This condition also
had the highest amount of subordinates who were classi-
?ed as ?rm-focused. In the both rankings condition, 33%
of the subordinates are focused on their own compensa-
tion, 56% are focused on a more equal split of the surplus,
and 11% are focused on ?rm pro?t. Collectively, these re-
sults show how subordinates’ reporting decisions were af-
fected by our ranking conditions. Moreover, these results
reveal why we observe differences in the levels of honesty
among our ranking conditions.
Conclusion
In this study, we examine whether rankings affect the
honesty of subordinates’ budget requests. In our experi-
ment, subordinates are requested to truthfully report a
project’s actual cost. Subordinates, however, have ?nancial
We classify subordinates’ strategies using three categories based on the percentage of surplus subordinates retained.
Specifically, we classify subordinates who claim between: (1) 67% and 100% of the surplus as primarily focused on
maximizing their own compensation; (2) 33% and 67% of the surplus as primarily focused on a more equal
(equitable) split; and (3) 0% and 33% of the surplus as primarily focused on firm profit.
Fig. 2. Subordinates’ strategies by ranking condition.
244 J.L. Brown et al. / Accounting, Organizations and Society 39 (2014) 237–246
incentives to maximally misrepresent the project’s true
cost and, as such, claim 100% of the project’s surplus. We
rank subordinates in one of four ways: based on (1) own
compensation, (2) ?rm pro?t, (3) both own compensation
and ?rm pro?t, and (4) randomly. None of the rankings af-
fect subordinates’ remuneration.
We ?nd that subordinates’ budget requests (honesty)
are signi?cantly affected by the rankings. Subordinates
ranked on ?rm pro?t claim the least surplus (45%), fol-
lowed by subordinates ranked on both individual compen-
sation and ?rm pro?t (55.6%), subordinates ranked
randomly (68%) and, ?nally, subordinates ranked on own
compensation (84%). Moreover, the surplus claimed by
subordinates in the both rankings condition is signi?cantly
lower than the surplus claimed by subordinates in the own
compensation condition but not signi?cantly different
from the surplus claimed by subordinates in the ?rm pro?t
condition. We also ?nd that our results are stable across
periods, suggesting that the ranking effects we observe
neither increase nor dissipate over time.
Our results have several important implications. First,
complementing extant research that documents the effects
of rankings on effort and performance (e.g., Frederickson,
1992; Hales et al., 2012; Hannan et al., 2008, 2013; Tafkov,
2013), we show how rankings signi?cantly affect another
variable vital to organizational success, managerial hon-
esty. Collectively, our ?ndings suggest that organizations
can use ?rm-based rankings as a low-cost, informal control
to constrain opportunistic reporting. Our results also
suggest, however, that compensation-, or even perqui-
site-, based rankings could lead employees to engage in
dishonest reporting and, as a result, ?rms should consider
carefully how they rank their employees and also be aware
of how employees may use accounting information to rank
themselves.
Second, we examine a setting in which subordinates re-
ceive two con?icting rankings. We ?nd that subordinates
in this condition respond more to the ?rm pro?t ranking
than to the own compensation ranking. This result is some-
what counter-intuitive, as one might reasonably posit that
having both rankings would lead subordinates to engage in
more self-regarding behavior (because subordinates could
internally justify their reporting decisions by focusing on
the own compensation ranking). This particular ?nding
suggests that, as long as ?rm-based rankings are provided,
the control loss associated with having access to a ranking
that encourages self-interested behavior may not be as
large as expected.
Certain limitations of our study provide opportunities
for further inquiry. For example, we chose to have a real
participant play the role of the superior because account-
ing research examining preferences for honesty has used
real participants for both subordinate and superior roles
and because this setting most closely mirrors the natural
environment where one’s actions invariably affect the wel-
fare of others. However, as noted by an anonymous Re-
viewer, this design choice may have prodded participants
receiving both the ?rm pro?t and own compensation rank-
ings to, for other-regarding reasons, focus more on the ?rm
pro?t ranking than on the own compensation ranking. Fu-
ture research could examine the robustness of our results
to this speci?c design choice, either by having a hypothet-
ical superior and/or by examining a setting where the ac-
tions of the subordinate have a more tenuous/indirect
relation to the superior’s remuneration.
As suggested by another anonymous Reviewer, future
research also could examine how cardinal ranking infor-
mation affects the veracity of employees’ reports. In our
experiment, subordinates were only provided with ordinal
ranks. We made this design choice so as not to provide
subordinates in our baseline (random rankings) setting
information that could be used to rank themselves. We
also made this design choice to examine the starkest form
of rankings. That said, employees frequently know the le-
vel of a performance variable (e.g., the amount of ?rm prof-
it) and, as such, could use this information to rank
themselves. Moreover, employees often are provided with
a formal rank coupled with the level of performance. Re-
search that extends our work to settings with different
types of ranking information and feedback could prove
quite fruitful.
References
Anderson, D., Crowell, C., Sponsel, S., Clarke, M., & Brence, J. (1982).
Behavior management in the public accommodations industry: A
three-project demonstration. Journal of Organizational Behavior
Management, 4, 33–66.
Arrow, K. J. (1974). Gifts and exchanges. Philosophy and Public Affairs I, 4,
343–362.
Azmat, G., & Iriberi, N. (2009). The provision of relative feedback
performance information: An experimental analysis of performance
and happiness. Working paper, Universitat Pompeu Fabra and
Barcelona GSE.
Baker, G. P., Jensen, M. C., & Murphy, K. J. (1988). Compensation and
incentives: Practice vs. theory. The Journal of Finance, 43, 593–616.
Beach, S., & Tesser, A. (1995). Self-esteem and the extended self-
evaluation maintenance model: The self in social context. In M.
Kernis (Ed.), Ef?cacy, agency, and self-esteem (pp. 145–170). New York:
NY Plenum Press.
Bebchuk, L. A., & Fried, J. M. (2003). Executive compensation as an agency
problem. Journal of Economic Perspectives, 17, 71–92.
Bick, I., Palmon, O., & Wald, J. (2006). CEO compensation, director
compensation, and ?rm performance: Evidence of cronyism? Journal
of Corporate Finance, 12, 403–423.
Blanchard, O. J., Lopez-de-Silanes, F., & Shleifer, A. (1994). What do ?rms
do with cash windfalls? Journal of Financial Economics, 36, 337–360.
Blanes i Vidal, J., & Nossol, M. (2011). Tournaments without prizes:
Evidence from personnel records. Management Science, 5, 1721–1736.
Buunk, A. P., & Gibbons, F. X. (2007). Social comparison: The end of a
theory and the emergence of a ?eld. Organizational Behavior and
Human Decision Processes, 102, 3–21.
Buunk, B. P., Kuyper, H., & van der Zee, Y. G. (2005). Affective response to
social comparison in the classroom. Basic and Applied Social
Psychology, 27, 229–237.
Cai, T., Jo, H., & Pan, C. (2011). Vice or virtue? The impact of corporate
social responsibility on executive compensation. Journal of Business
Ethics, 104, 159–173.
Church, B. K., Hannan, R. L., & Kuang, X. (2012). Shared interest and
honesty in budget reporting. Accounting, Organizations and Society, 37,
156–167.
Cialdini, R., Reno, R., & Kallgren, C. (1990). A focus theory of normative
conduct: Recycling the concept of norms to reduce littering in public
places. Journal of Personality and Social Psychology, 58, 1015–1026.
Cookson, R. (2000). Framing effects in public goods experiments.
Experimental Economics, 3, 55–79.
Core, J. R., & Larker, D. (1999). Corporate governance, chief executive
of?cer compensation and ?rm performance. Journal of Financial
Economics, 51, 371–406.
Dana, J., Weber, R. A., & Kuang, J. X. (2007). Exploiting moral wiggle room:
Experiments demonstrating an illusory preference for fairness.
Economic Theory, 33, 67–80.
Duffy, J., & Kornienko, T. (2010). Does competition affect giving? Journal of
Economic Behavior & Organization, 74, 82–103.
J.L. Brown et al. / Accounting, Organizations and Society 39 (2014) 237–246 245
Evans, J. H., III, Hannan, R. L., Krishnan, R., & Moser, D. V. (2001). Honesty
in managerial reporting. The Accounting Review, 76, 537–559.
Festinger, L. (1954). A theory of social comparison process. Human
Relations, 7, 117–140.
Fischbacher, U. (2007). Z-Tree: Zurich toolbox for ready-made economic
experiments. Experimental Economics, 10, 171–178.
Fisher, J. G., Maines, L. A., Peffer, S. A., & Sprinkle, G. B. (2002). Using
budgets for performance evaluation: Effects of resource allocation
and horizontal information asymmetry on budget proposals, budget
slack, and performance. The Accounting Review, 77, 847–865.
Frederickson, J. R. (1992). Relative performance information: The effects
of common uncertainty and contract type on subordinate effort. The
Accounting Review, 67, 647–669.
Goldstein, N. J., & Cialdini, R. B. (2007). The spyglass self: A model of
vicarious self-perception. Journal of Personality and Social Psychology,
92, 402–417.
Grant, A. (2013). Givers take all: The hidden dimension of corporate
culture. McKinsey Quarterly (April).
Gruder, C. L. (1971). Determinants of social comparison choices. Journal of
Experimental Social Psychology, 7, 473–489.
Haisley, E. C., & Weber, R. A. (2010). Self-serving interpretations of
ambiguity in other-regarding behavior. Games and Economic Behavior,
68, 614–625.
Hales, J., Hobson, J. H., & Resutek, R. J. (2012). The dark side of socially
mediated rewards: How narcissism and social status affect
managerial reporting. Working paper, University of Illinois at
Urbana-Champaign.
Hannan, L. R., Krishnan, R., & Newman, A. H. (2008). The effects of
disseminating relative performance feedback in tournament and
individual performance compensation plans. The Accounting Review,
83, 893–913.
Hannan, L. R., McPhee, G. P., Newman, A. H., & Tafkov, I. D. (2013). The
effect of relative performance information on effort allocation and
performance in a multi-task environment. The Accounting Review, 88,
553–575.
Hannan, L. R., Rankin, F. W., & Towry, K. L. (2006). The effect of
information systems on honesty in managerial reporting: A
behavioral perspective. Contemporary Accounting Research, 23,
885–918.
Huguet, P., Dumas, F., Monteil, J. M., & Genestoux, N. (2001). Social
comparison choices in the classroom: Further evidence for
students’ upward comparison tendency and its bene?cial impact
on performance. European Journal of Social Psychology, 31, 557–
578.
Hyman, H. (1942). The psychology of subjective status. Psychological
Bulletin, 39, 473–474.
Kramer, R. M., & Brewer, M. B. (1984). Effects of group identity on
resource use in a simulated commons dilemma. Journal of Personality
and Social Psychology, 46, 1044–1057.
Krupka, E., & Weber, R. A. (2009). The focusing and informational effects
of norms on pro-social behavior. Journal of Economic Psychology, 30,
307–320.
Lazarus, R. (1991). Emotion and adaption. NewYork, NY: Oxford University
Press.
Metee, D. R., & Smith, G. (1977). Social comparison and interpersonal
attraction: The case for dissimilarity. In J. M. Sulls & R. L. Miller (Eds.),
Social comparison processes: Theoretical and empirical perspectives
(pp. 69–101). Washington: Hemisphere.
Murphy, K. R. (1993). Honesty in the workplace. Pac?c Grove, CA: Brooks/
Cole.
Nordstrom, R., Lorenzi, P., & Hall, R. (1990). A review of public posting
of performance feedback. Journal of Organizational Behavior
Management, 11, 101–123.
Noreen, E. (1988). The economics of ethics: A new perspective on agency
theory. Accounting, Organizations and Society, 13, 359–369.
Nosanchuk, T. A., & Erickson, B. H. (1985). How high is up? Calibrating
social comparison in the real world. Journal of Personality and Social
Psychology, 48, 624–634.
Rankin, F. W., Schwartz, S. T., & Young, R. A. (2003). Management control
using nonbinding budgetary announcements. Journal of Management
Accounting Research, 15, 75–93.
Rankin, F. W., Schwartz, S. T., & Young, R. A. (2008). The effect of honesty
and superior authority on budget proposals. The Accounting Review,
83, 1083–1099.
Rowe, C. (2004). The effect of accounting report structure and team
structure on performance in cross-functional teams. The Accounting
Review, 79, 1153–1180.
Schatzberg, J., & Stevens, D. E. (2008). Public and private forms of
opportunism with the organization: A joint examination of budget
and effort behavior. Journal of Management Accounting Research, 20,
59–81.
Smith, R. H. (2000). Assimilative and contrastive emotional reactions to
upward and downward social comparisons. In J. Suls & L. Wheelers
(Eds.), Handbook of social comparison: Theory and research
(pp. 173–200). New York: Plenum.
Suls, J. M., & Miller, R. L. (1977). Social comparison processes: Theoretical
and empirical perspectives. Washington, DC: Hemisphere.
Tafkov, I. D. (2013). Private and public relative performance information
under different compensation contracts. The Accounting Review, 88,
327–350.
Tesser, A. (1988). Toward a self-evaluation maintenance model of social
behavior. Advances in Experimental Social Psychology, 21, 181–228.
Tesser, A., & Campbell, J. (1980). Self-de?nition: The impact of the relative
performance and similarity of others. Social Psychology Quarterly, 43,
341–347.
Tosi, H. L., Wermer, S., Katz, J. P., & Gomez-Mejia, L. R. (2000). How much
does performance matter? A meta-analysis of CEO pay studies.
Journal of Management, 26, 301–339.
Towry, K. L. (2003). Control in a teamwork environment – The impact of
social ties on the effectiveness of mutual monitoring contracts. The
Accounting Review, 78, 1069–1095.
West, K. L., & Mykerezi, E. (2011). Teachers’ unions and compensation:
The impact of collective bargaining on salary schedules and
performance pay schemes. Economics of Education Review, 30, 99–108.
Wheeler, L. (1966). Motivation as a determinant of upward comparison.
Journal of Experimental Social Psychology, 1, 27–31.
Wheeler, L., Koestner, R., & Driver, R. E. (1982). Related attributes in the
choice of comparison others: It’s there, but it isn’t all there is. Journal
of Experimental Social Psychology, 18, 489–500.
Wheeler, L., Shaver, K. G., Jones, R. A., Goethals, G. R., Cooper, J., Robinson,
J. E., et al. (1969). Factors determining choice of a comparison other.
Journal of Experimental Social Psychology, 5, 219–232.
Wikoff, M., Anderson, D., & Crowell, C. (1982). Behavior management in a
factory setting: Increasing work ef?ciency. Journal of Organizational
Behavior Management, 4, 97–127.
Wood, J. V. (1989). Theory and research concerning social comparisons of
personal attributes. Psychological Bulletin, 106, 231–248.
Zhang, M. (2008). The effects of perceived fairness and communication on
honesty and collusion in a multi-agent setting. The Accounting Review,
83, 1125–1146.
246 J.L. Brown et al. / Accounting, Organizations and Society 39 (2014) 237–246
doc_922341880.pdf