ACCOUNTING INFORMATION AGGREGATION
AND MANAGERIAL COOPERATION
By
Eric Jonathan Marinich

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Business Administration—Doctor of Philosophy
2013

ABSTRACT
ACCOUNTING INFORMATION AGGREGATION
AND MANAGERIAL COOPERATION
By
Eric Jonathan Marinich

When managers face incentives against cooperation, accounting information can increase
managerial cooperation when it allows managers to perfectly verify the cooperativeness of
others’ prior actions. The extent to which accounting information facilitates perfect verification
of others’ prior actions, however, can depend on its aggregation. This dissertation provides
theory-consistent experimental evidence of the effects of accounting information aggregation on
managerial cooperation when managers face incentives against cooperation. Based on the
psychology theory of non-consequential reasoning, I predict and find that managerial
cooperation is higher when accounting information is aggregated and individuals cannot
perfectly verify others’ prior actions than when accounting information is disaggregated and
individuals can perfectly verify others’ prior actions. The experimental results indicate that
individuals are more likely to use non-consequential reasoning when accounting information is
aggregated than when it is disaggregated. As a result, they are more likely to frame decisions as
group decisions and cooperate because it is the only action that leads to the best group outcome.

Copyright by
ERIC JONATHAN MARINICH
2013

To my grandmother, who was a strong woman

iv

ACKNOWLEDGMENTS

I gratefully acknowledge the guidance and support of my dissertation committee: Mike
Shields (Chair), Ranjani Krishnan, Joan Luft, and Joseph Cesario. I will be forever indebted to
them for their intellectual and personal mentorship. I thank Matt DeAngelis, Susanna Gallani,
Michelle Lau, Kathy Petroni, Joe Schroeder, Karen Sedatole, Amy Swaney, and Tyler Thomas
for their helpful comments on drafts of my dissertation. I thank the Department of Accounting
and Information Systems at Michigan State University for providing financial support for my
dissertation.
I also would not have been able to complete my dissertation without the love and support
of my family and friends. I thank my husband, Corey, for his never-ending kindness, constant
encouragement, and inexhaustible patience. Corey, you are my love, my happiness, and my best
friend every single day of my life. I thank my mother, Ruth Ann, for inspiring me to persist in
achieving my goals, for teaching me the value of hard work, and for always believing in me even
when I doubted myself. I thank my sister, Fran, for being a loyal sister and for being one of my
best friends; we shared some of the happiest moments in my life and I’ll cherish those memories
forever. I thank my aunt for her humor and friendship; the laughter we’ve shared has eased my
pain and lightened my life at the darkest times, just when I needed it most.
Finally, I thank my grandma, to whom my dissertation is dedicated. Grandma, not a day
passes that I don’t think about you. Thank you for loving me unconditionally, for always
believing in me, for always being proud of me, and for making me feel like I was the most
special boy on earth every single moment that I spent with you. Your love made me a better
person and you will be part of me forever.

v

TABLE OF CONTENTS

LIST OF TABLES.........................................................................................................................vii
LIST OF FIGURES…………………………………………………………………………......viii
INTRODUCTION………………………………………………………………………...............1
CHAPTER 1: LITERATURE REVIEW AND HYPOTHESIS DEVELOPMENT……………...8
1.1 Accounting and Managerial Cooperation……………………………………………..8
1.2 Accounting Information Aggregation………………………………………………..13
1.3 Consequential and Non-Consequential Reasoning…………………………………..16
1.4 Hypothesis Development………………………………………………………….....17
CHAPTER 2: EXPERIMENTAL DESIGN……….……………………………………………20
2.1 Participants……………………………………………………………………….…..20
2.2 Experimental Design and Treatment Conditions………………………………….....20
2.3 Procedure and Task…………………………………………………………………..21
2.3.1 Part 1: Instructions…………………………………………………………22
2.3.2 Part 2: Experimental Periods………………………………………………23
2.3.3 Part 3: Post-experiment…………………………………………………….25
CHAPTER 3: RESULTS…………….……………………………………………………….….26
3.1 Manipulation Checks………………………………………………………………...26
3.2 Descriptive Statistics and Preliminary Analysis…..…………………………….…...27
3.3 Hypothesis Test………………………………………………………………………29
3.4 Supplemental Analyses………………………………………………………………30
3.4.1 Expectations about the Likelihood of Cooperation……………………......30
3.4.2 Factors that Influence Participants’ Management Decisions………………31
CHAPTER 4: CONCLUSION……………………………………………………………….…34
4.1 Summary of Experiment and Results………………………………..………………34
4.2 Limitations and Suggestions for Future Research…………………………………...35
APPENDIX………………………………………………………………………………………37
BIBLIOGRAPHY…………………………………………………………………………….….51

vi

LIST OF TABLES

TABLE 1A: Accounting Information Aggregation (Aggregated Condition). At the beginning
of period 5, participants in the aggregated accounting information condition were provided with
aggregated accounting information about the organizational subunit with which there was profit
interdependence in that period…………………………………………………………………...43

TABLE 1B: Accounting Information Aggregation (Disaggregated-Cooperative Condition).
At the beginning of period 5, participants in the disaggregated-cooperative accounting
information condition were provided with disaggregated accounting information about the
organizational subunit with which there was profit interdependence in that period. The
organizational subunit with which participants in this condition were paired chose the
cooperative management action in period 4…………………………………………………….44

TABLE 1C: Accounting Information Aggregation (Disaggregated-Non-Cooperative
Condition). At the beginning of period 5, participants in the disaggregated-non-cooperative
accounting information condition were provided with disaggregated accounting information
about the organizational subunit with which there was profit interdependence in that period. The
organizational subunit with which participants in this condition were paired chose the noncooperative management action in period 4……………………………………………………..45

TABLE 2: Descriptive Statistics. Number (%) of participants who choose the cooperative
action in period 5.……………………………………………………………………………..…46
TABLE 3: Supplemental Analysis I. Participants’ expectations about the likelihood of
cooperative action in period 5……………………………………………………..…………….47

TABLE 4A: Supplemental Analysis II (Post-Experiment Questionnaire Item). In Part 3 of
the experiment, participants respond to a post-experiment questionnaire item about factors that
influenced their management decision in period 5. Participants who expected the other
participant to choose Action X (i.e., the cooperative action) in period 5 allocated 100 points to
seven factors…………………………………………………………………………………….48

TABLE 4B: Supplemental Analysis II (Decision Factors). Descriptive statistics about factors
that influenced participants’ decisions in period 5. This table reports the mean (standard
deviation) points allocated by condition for participants who expected the other participant to
choose the cooperative action…………………………………………………………………..49

vii

LIST OF FIGURES
Figure 1: Profit Interdependence as a Prisoner’s Dilemma Game. Normal-form
representation of division profit structure as a prisoner’s dilemma game..……………………...38

Figure 2A: Order of Events. Order of events for Part 1, Part 2, and Part 3 of the experiment...39
Figure 2B: Task Overview (Periods 1 – 4). Experimental task for periods 1 – 4……………..40
Figure 2C: Task Overview (Period 5). Experimental task for period 5……………………….41

Figure 3: Preliminary Analysis. Mean managerial cooperation by accounting aggregation
condition………………………………………………………………………………………...42

viii

INTRODUCTION
Managers are often individually responsible and rewarded for their organizational
subunits’ profits (Antle and Demski 1988; Briers and Chua 2001; Melumad, Mookherjee, and
Reichelstein 1992; Pick 1971), but their profits often depend on the actions of other
organizational subunits. Cooperation among managers is important in these situations because it
can lead to positive financial and psychological outcomes for organizations and their managers
such as increased innovation, information sharing, performance, and trust (Coletti, Sedatole, and
Towry 2005; Christ et al. 2008; Drake and Haka 2008; Drake, Haka, and Ravenscroft 1999;
Rowe 2004; Rowe, Shields, and Birnberg 2008; Tayler and Bloomfield 2011; Young, Fisher, and
1

Lindquist 1993). Managers, however, often face incentives against cooperation; for example,
managers face incentives against cooperation when they can maximize their individual profits by
choosing non-cooperative actions such as free-riding (Coletti et al. 2005; McCarter and
Northcraft 2007; Rowe 2004; Tayler and Bloomfield 2011; Schwartz and Young 2002;
Schwartz, Young, and Zvinakis 2000; Zeng and Chen 2003).
Accounting research, however, provides evidence that non-trivial levels of managerial
cooperation are common, even when managers face incentives against cooperation (Coletti et al.
2005; Drake et al. 1999; Drake and Haka 2008; Rowe 2004; Rowe et al. 2008; Schwartz and
Young 2002; Schwartz et al. 2000; Tayler and Bloomfield 2011). For example, accounting
information can increase managerial cooperation by changing managers’ implicit incentives to
1

Collusion is a form of cooperation that can lead to negative financial outcomes for
organizations (e.g., decreased effort and performance) (cf. Hannan, Towry, and Zhang 2013 and
Towry 2003). Collusion occurs when managers agree to make decisions that maximize their joint
profit but not organizational profit. In order to achieve collusion, however, managers must be
able to coordinate their decisions, agree to do so, and must be able to enforce their agreement
(e.g., through contracting). Collusion is not possible in my dissertation because managers
independently and simultaneously decide whether or not to cooperate and therefore face a
coordination problem.
1

cooperate when it allows them to infer with certainty whether or not others have previously
cooperated thereby allowing managers to build reputations for cooperation or non-cooperation
(Coletti et al. 2005; Schwartz and Young 2002; Schwartz et al. 2000). Accounting information,
however, is often a noisy signal of managers’ prior actions due to characteristics of the
information such as its aggregation. Little is known about the effects of accounting information
on managerial cooperation when its aggregation is such that managers cannot infer the
cooperativeness of other managers’ prior actions with certainty. This dissertation therefore
investigates the effects of accounting information aggregation on managerial cooperation when
managers face incentives against cooperation.
Managerial cooperation depends on managers’ explicit and implicit incentives for
cooperation and on their uncertainty about the likelihood that other managers will cooperate.
Accounting controls and information can influence managerial cooperation by increasing
managers’ incentives for cooperation and by influencing their beliefs about the likelihood that
others will cooperate. Accounting information, for example, can increase managers’ implicit
incentives for cooperation when it allows managers to build reputations for cooperation or noncooperation by allowing them to infer the cooperativeness of other managers’ prior actions with
certainty. In these situations, accounting information can be used as the basis for rewards
(punishments) for other managers’ prior cooperation (non-cooperation) (Dickhaut and McCabe
1997; Hales and Williamson 2010; King 1996; Mayhew 2001; Schwartz and Young 2002;
Schwartz et al. 2000). The ability to reward (punish) cooperation (non-cooperation) is

2

fundamental to motivating and sustaining cooperation because reward (punishment) for
cooperation (non-cooperation) increases the expected payoffs to cooperation (Axelrod 1984).

2

Accounting information, however, can be a noisy signal of managers’ prior actions and,
as such, may not allow managers to perfectly verify the cooperativeness of others’ prior actions.
Accounting information, for example, can be aggregated in its natural form (e.g., annual lease
expense). Alternatively, accounting systems may aggregate information that is disaggregated in
its natural form in order to facilitate managerial decision-making. While aggregation can reduce
the cognitive demands that large amounts of disaggregated information can place on users with
limited cognitive resources, it can also introduce noise and decrease the informativeness of
accounting information (Datar and Gupta 1994; Ijiri 1967; Lev 1969; Sorter 1969). If managerial
cooperation depends on managers’ ability to use accounting information as the basis for
rewarding (punishing) other managers for their prior cooperation (non-cooperation), then
aggregated accounting information that obfuscates managers’ prior actions will reduce the
implicit incentive effect of accounting information on managerial cooperation. The effects of
accounting information aggregation on managerial cooperation, however, are largely unexplored
in the accounting literature.

3

Psychology research on non-consequential reasoning provides a theoretical basis for
predictions about the effects of aggregated accounting information on individuals’ decisions to

2

I use reputation in a manner that is consistent with prior accounting research on managerial
cooperation (Hales and Williamson 2010; Mayhew 2001; Schwartz and Young 2002; Schwartz
et al. 2000). Reputation-building is possible in situations in which managers can form
expectations about other managers’ actions based on their prior actions.
3

One exception is Nikias, Schwartz, Spires, Wollscheid, and Young (2010), who use an
experiment to investigate the effects of cost information aggregation and budget request timing
on the level of slack contained in subordinates’ budget requests.
3

cooperate when they face incentives against cooperation (Croson 1999; Shafir and Tversky 1992;
Tversky and Shafir 1992). Expected utility theory assumes that individuals use consequential
reasoning to make decisions (i.e., they evaluate all decision alternatives with respect to their
consequences and expected likelihoods) (Hammond 1988; Levi 1991; Savage 1954). Individuals
who make decisions under uncertainty, however, sometimes use non-consequential reasoning to
make their decisions and fail to appropriately consider all relevant decision outcomes. If
aggregated accounting information creates uncertainty about others’ prior actions, then managers
who rely on aggregated accounting information when they make cooperation decisions may be
more likely to use non-consequential reasoning and make cooperation decisions that are
inconsistent with the assumptions of expected utility theory.
Shafir and Tversky (1992) propose that non-consequential reasoning explains non-trivial
cooperation in prisoner’s dilemma games in which individuals face incentives against
cooperation. They predict and find that a player who is uncertain about the other player’s action
choice when she chooses her own action is more likely to choose cooperation than a player who
knows the other player’s action choice before she chooses her own action, regardless of whether
the other player chose cooperation or non-cooperation. When there is uncertainty about the other
player’s action choice, the game’s outcome depends on both players’ decisions. As a result,
players frame their decision as a group decision and focus their attention on the action that can
lead to the highest joint payoff (i.e., cooperation). In contrast, when there is certainty about the
other player’s action choice, the game’s outcome depends only on the decision of an individual
player. As a result, players frame their decision as an individual decision and focus their
attention on the action that leads to the highest individual payoff (i.e., non-cooperation).

4

I use the psychology theory of non-consequential reasoning to predict that aggregated
accounting information will increase managerial cooperation, even though it does not allow
managers to infer the cooperativeness of others’ prior actions with certainty. Based on the theory
of non-consequential reasoning, I expect that individuals who are provided with aggregated
(disaggregated) accounting information will use non-consequential (consequential) reasoning to
make their decisions because they cannot (can) use accounting information to infer the
cooperativeness of others’ prior actions with certainty. As a result, I predict that individuals who
are provided with aggregated accounting information will frame the decision to cooperate as a
group decision and prefer cooperation because it is the only action that leads to the highest joint
profit. In contrast, I predict that individuals who are provided with disaggregated accounting
information will frame the decision to cooperate as an individual decision and prefer noncooperation because it is the only action that leads to the highest individual profit. Thus, I
hypothesize that managerial cooperation will be higher when accounting information is
aggregated than when it is disaggregated.
To test my hypothesis, I use an experiment in which accounting information aggregation
is manipulated at two levels: aggregated and disaggregated profit information. Participants play a
repeated one-shot prisoner’s dilemma game in which they assume the role of division managers
who choose either the cooperative or non-cooperative management action for their division in
each period. In the final period, participants are provided with accounting information about the
other participant’s prior-period profit when the other participant was matched with a different
participant. In the disaggregated (aggregated) condition, participants can (cannot) use accounting
information to infer the cooperativeness of the other participant’s prior-period action with

5

certainty. The dependent variable, managerial cooperation, is the proportion of participants who
choose the cooperative management action in a treatment condition in the final period.
The experimental results support my hypothesis. Managerial cooperation is higher when
accounting information is aggregated than when it is disaggregated. The results of two sets of
supplementary analyses provide evidence that the pattern of cooperation that I observe in my
experiment is consistent with the theory of non-consequential reasoning. Specifically,
participants are more likely to use non-consequential reasoning when accounting information is
aggregated than when it is disaggregated. Participants who are provided with aggregated
accounting information frame the cooperation decision as a group decision and therefore prefer
to maximize joint profit, while participants who are provided with disaggregated accounting
information frame the cooperation decision as an individual decision and therefore prefer to
maximize individual profit.
My dissertation contributes to the accounting literature by addressing an important
limitation of prior research, which investigates the effects of accounting information on
managerial cooperation only when it allows individuals to perfectly verify the cooperativeness of
others’ prior actions (Coletti et al. 2005; Schwartz et al. 2000; Schwartz and Young 2002). This
is a limitation of prior research because accounting information that is provided to managers in
the natural environment is often a noisy signal of other managers’ prior actions and, as such, may
not eliminate managers’ uncertainty about the cooperativeness of others’ prior actions. I address
this limitation by providing evidence of the effects of accounting information on managerial
cooperation when it’s aggregation is such that it does not allow individuals to infer the
cooperativeness of others’ prior actions with certainty (i.e., when accounting information is a
noisy signal of other managers’ prior actions). Rather than aggregation decreasing managerial

6

cooperation by hindering reputation-building and preventing individuals from inferring others’
prior actions with certainty, my dissertation provides evidence that aggregated accounting
information can increase managerial cooperation by influencing how individuals frame their
cooperation decisions.
The remainder of my dissertation is organized as follows. Chapter 1 reviews relevant
literature and develops the hypotheses. Chapter 2 describes the experimental design. Chapter 3
reports the results and Chapter 4 concludes.

7

CHAPTER 1: LITERATURE REVIEW AND HYPOTHESIS DEVELOPMENT
This chapter first reviews the accounting literature on the effects of accounting
information on managerial cooperation. Next, it reviews the accounting literature on the effects
of aggregation on managerial decision-making and the psychology literature on nonconsequential reasoning. Finally, it develops a hypothesis about the effects of accounting
information aggregation on managerial cooperation.
1.1 Accounting and Managerial Cooperation
Interdependence among the profits of organizational subunits is ubiquitous and has
important implications for the design of accounting and for managerial decision-making
(Bouwens and Abernethy 2000; Caglio and Ditillo 2008; Dekker 2004; Galbraith 1973;
Thompson 1967). Managerial cooperation is important when there is profit interdependence
because it can lead to positive financial and psychological outcomes such as increased
organizational performance and interpersonal trust (Coletti et al., 2005; Schwartz and Young,
2002; Schwartz et al. 2000). Managerial cooperation occurs when managers make decisions that
lead to the highest joint profit for the group of managers among whom there is profit
interdependence. Accounting research investigates managerial cooperation in a variety of intraand inter-firm settings such as cross-functional teams, joint investments, joint production, and
strategic alliances (Caglio and Ditillo 2008; Coletti et al. 2005; Dekker 2004; Drake and Haka
1999; Drake et al. 1998; Rowe 2004; Rowe et al. 2008; Schwartz et al. 2000; Schwartz and
Young 2002; Tomkins 2001).
Managerial cooperation depends on at least two factors: managers’ incentives for
cooperation and their uncertainty about whether or not other managers will cooperate. Managers,
however, often face incentives against cooperation such that the actions that lead to the highest

8

joint profit are not the same actions that lead to the highest individual profits for the managers.
As a result, managerial cooperation can be difficult to achieve and sustain (Das and Teng 1996,
1999, 2000; Devetag and Ortmann 2007). Cooperation problems occur when at least one
manager does not cooperate. Cooperation problems are common and can lead to negative
outcomes such as decreased effort, performance, and trust (Christ et al. 2008; Coletti et al. 2005;
Drake et al. 1999; Kachelmeier and Shehata 1997; Drake and Haka 2008; Rowe 2004; Rowe et
al. 2008; Tayler and Bloomfield 2011; Young et al. 1993).
Cooperation decisions are often represented in the social science literatures as a
prisoner’s dilemma game in which individuals face incentives against cooperation and make
decisions that can maximize either individual or joint profit (Axelrod 1984; Kollock 1993;
4

McCarter and Northcraft 2007; Taylor 1976, 1987; Zeng and Chen 2003). In prisoner’s
dilemma, non-cooperation is each player’s dominant Nash equilibrium strategy, but it is not the
Pareto optimal strategy. If both players chose cooperation, then each player could earn higher
profits than they would earn at the Nash equilibrium. Representing managers’ incentives for
cooperation as a prisoner’s dilemma captures the fundamental management control challenge
that arises when managerial cooperation is desirable, but managers face financial incentives to
choose non-cooperative actions (e.g., free-riding, misrepresentation of private information).
Accounting information can increase managerial cooperation even when managers face
incentives against cooperation by allowing managers to build reputations for cooperation or noncooperation, thereby increasing managers’ implicit incentives for cooperation (Coletti et al.

4

The analysis of prisoner’s dilemma in this chapter assumes that the profits associated with the
cooperative and non-cooperative actions represent the actual utility payoffs to managers. That is,
this analysis assumes that managers have only monetary preferences. If managers have nonmonetary preferences, then the utility payoffs to cooperative actions can increase and
cooperation can be a Nash equilibrium strategy.
9

2005; Hales and Williamson 2010; Schwartz et al. 2000; Schwartz and Young 2002). For
example, when accounting information allows managers to perfectly verify whether or not other
managers have previously cooperated, it enables managers to condition their cooperation
decisions on other managers’ prior actions and to reward (punish) other managers for their prior
cooperation (non-cooperation). The ability to reward and punish individuals for the
cooperativeness of their prior actions is important for achieving and sustaining cooperation
(Axelrod 1984; Dal Bo 2005; Charness and Rabin 2002; Heide and Miner 1992; Skaperdas and
Syropoulos 1996). When managers anticipate a future in which perfect accounting information
will be available, they anticipate future financial rewards (punishments) for cooperation (noncooperation). As a result, their current incentives for cooperation increase in the amount of the
present value of the expected future payoffs to cooperation.
Schwartz et al. (2000) use a modified repeated prisoner’s dilemma game to provide
experimental evidence that accounting information can increase individuals’ implicit incentives
for cooperation by facilitating reputation-building, even when they cannot build reputations
through repeated interaction. They predict and find that individuals in a joint investment decision
are more likely to cooperate (i.e., invest resources in a joint investment rather than an individual
investment) when accounting information allows others to perfectly verify the cooperativeness of
their prior actions. Managerial cooperation is higher when accounting information is present than
when it is absent, regardless of the timeliness of the information, because the expected payoffs to
cooperation are higher when accounting information is present than when it is absent. When
perfect accounting information is present, individuals anticipate future financial rewards for
cooperation and future financial punishments for non-cooperation. When perfect accounting
information is absent, however, individuals do not anticipate future financial rewards or

10

punishments for the cooperativeness of their actions because other individuals cannot use
accounting information to infer their actions. As a result, the expected payoffs to cooperation are
higher (and managerial cooperation is higher) when accounting information is present than when
it is absent.
Schwartz and Young (2002) also use a modified repeated prisoner’s dilemma game to
corroborate and extend the findings of Schwartz et al. (2000). They provide experimental
evidence about the incremental effects of accounting information on managerial cooperation
when managers can also build reputations through repeated interaction with the same managers.
They investigate individuals’ ex ante disclosures of their private information in a setting in which
joint-profit maximization depends on truthful disclosures, but individuals face incentives to
misrepresent their private information. Consistent with Schwartz et al. (2000), they find that
when individuals do not interact repeatedly, cooperation is higher when perfect accounting
information is present than when it is absent because it facilitates reputation-building and
increases implicit incentives for cooperation. Schwartz and Young (2002), however, also find
that accounting information does not increase cooperation when individuals interact repeatedly.
Although cooperation is higher when individuals interact repeatedly than when they do not,
individuals who interact repeatedly are no more likely to cooperate when accounting information
is present than when it is absent because repeated interaction serves as a substitute for perfect
accounting information in reputation building.
Coletti et al. (2005) provide experimental evidence that accounting can also increase
managerial cooperation by increasing managers’ explicit incentives to cooperate and by building
trust among managers. Coletti et al. (2005) investigate the effects of accounting controls and
information on managerial cooperation over time in an existing relationship in which managers

11

interact repeatedly and face incentives against cooperation (i.e., in a repeated prisoner’s dilemma
game). Specifically, they investigate a setting in which audits allow organizations to perfectly
monitor the cooperativeness of managers’ prior actions and enforce cooperation by imposing
financial penalties for non-cooperation. Coletti et al. (2005) predict and find that individuals are
more likely to allocate high resources to a joint research and development project (i.e.,
cooperate) when perfect accounting information allows organizations to impose financial
penalties for non-cooperation that increase the payoffs to cooperative actions such that
cooperation maximizes not only joint profit, but also individual profits. Further, they predict and
find that although increased cooperation is attributable to financial penalties for non-cooperation
(a situational factor), individuals attribute increased cooperation to others’ trustworthiness (a
dispositional factor). As a result, individuals judge others to be more trustworthy when
accounting is present than when it is absent and cooperate with others more even after the
financial penalties for non-cooperation are removed.
While prior accounting research investigates the effects of accounting information on
managerial cooperation when it allows individuals to infer the cooperativeness of others’ prior
actions with certainty, experimental economics research provides evidence that cooperation
depends on the certainty with which individuals can infer the cooperativeness of others’ prior
actions (Bolton, Katok, and Ockenfels 2005; Nowak and Sigmund 2005). One factor that can
influence individuals’ certainty about others’ prior actions is the aggregation of information
about others’ prior actions. When individuals are provided with accounting information that is a
noisy signal of others’ prior actions due to characteristics of the information such as its
aggregation, I expect the effects of accounting information on managerial cooperation to differ

12

from those observed in prior accounting research on accounting information and managerial
cooperation.
1.2 Accounting Information Aggregation
The extent to which accounting information eliminates individuals’ uncertainty about the
cooperativeness of others’ prior actions depends on characteristics of the information, such as its
aggregation (Banker and Datar 1989; Holmstrom 1979). Aggregation can change the
informativeness (i.e., noise) of accounting information because it can change the mapping of
actions to their financial outcomes (e.g., profits). Specifically, aggregation can increase the
number of actions that are mapped to an outcome. Aggregation is a continuum with fully
aggregated accounting information and fully disaggregated accounting information as endpoints. When accounting information is fully aggregated, all actions are mapped to a single
financial outcome (e.g., joint profit). When accounting information is fully disaggregated,
individual actions are mapped to separate financial outcomes (e.g., division profits).

5

The aggregation of accounting information can vary due to measurement capabilities, the
costs of aggregation and disaggregation, and to suit the needs of managers with different
information demands and decision models. Aggregation is a design choice for organizations that
trade-off the costs and benefits of aggregation and disaggregation (Horngren et al. 2012; Lev

5

Note that division profit may not provide a one-to-one mapping of managers’ actions to their
financial outcomes. For example, division profit may not provide a one-to-one mapping of
managers’ actions to their financial outcomes when there is profit interdependence among
divisions. In my dissertation, I assume that the disaggregation process sufficiently filters out the
effects of profit interdependence such that managers can use disaggregated accounting
information to infer other managers’ prior actions with certainty.
13

6

1969; Sorter 1969). Aggregation can benefit organizations and their managers by summarizing
financial outcomes thereby reducing the cognitive demands that large amounts of detailed,
disaggregated accounting information can place on managers with limited cognitive resources.
Aggregated accounting information, for example, may better suit the needs of centralized
managers who require summarized information about organizational subunits’ costs and
revenues in order to prepare master budgets or evaluate the financial reporting segments’
performance.
Empirical accounting research, for example, provides evidence that aggregated
accounting information can increase managerial cooperation in cross-functional team settings,
even when managers face incentives against cooperation (Rowe 2004; Rowe et al. 2008). Rowe
(2004) investigates a public goods dilemma in which information sharing can lead to increased
team performance (i.e., information sharing is a cooperative strategy that can lead to Pareto
outcomes), but it is costly and individuals can be better off by free-riding (i.e., non-cooperation
is the only Nash equilibrium strategy). Using an experiment, he provides evidence that
accounting report structure (aggregated, disaggregated) and team structure (face-to-face,
dispersed) influence whether individuals frame their information-sharing decisions as either
group or individual decisions. Consistent with his predictions, he finds that managerial
cooperation is higher when aggregated accounting information and face-to-face team structure
align to provide group framing than when accounting information and/or team structure provide
individual framing.

6

The American Institute of Certified Public Accountants provides guidance for organizations on
the aggregation of accounting information that is appropriate for organizations to use in order for
the accounting information to be presented fairly. The Statement on Auditing Standards No. 5
states that in order for accounting information to be fairly presented, it should be “summarized in
a reasonable manner that is neither too detailed nor too condensed.” (Nikias et al. 2010, 68)
14

Experimental economics research provides evidence that organizations can provide
aggregated accounting information in order to establish and maintain confidentiality among
managers, which can increase managerial cooperation (John, Loewenstein, and Rick 2013). For
example, when managers’ performance evaluations and compensation depend on disaggregated
accounting measures such as division profit, organizations may choose to provide aggregated
accounting information such as financial reporting segment profits to managers in order to
prevent them from inferring other managers’ compensation. When they can (cannot) infer other
managers’ compensation and their own compensation is lower than the other managers’
compensation, managers are more (less) likely to engage in upward social comparison and
choose non-cooperative actions such as dishonest communication and free-riding.
Although aggregation can benefit organizations and their managers, it can also be costly
to organizations when it decreases the informativeness of accounting information by introducing
noise (Datar and Gupta 1994; Ijiri 1967; Lev 1969; Sorter 1969). When it introduces noise,
aggregation can obfuscate properties of profit distributions that may be useful inputs to
managers’ decision models. To the extent that the aggregation of accounting information
decreases its informativeness about prior outcomes, managers may not be able to use the
accounting information to infer with certainty the cooperativeness of other managers’ prior
actions.
The effect of accounting information on managerial cooperation when it creates
uncertainty for managers about the cooperativeness of others’ prior actions is an unanswered
question. Next, I review psychology literature on consequential and non-consequential reasoning
and consider the possible effects of uncertainty on managerial cooperation when managers face
incentives against cooperation.

15

1.3 Consequential and Non-Consequential Reasoning
Expected utility theory assumes that individuals use consequential reasoning when they
make decisions (i.e., they appropriately evaluate each possible decision outcome and its expected
likelihood) (Hammond 1988; Levi 1991). When they use consequential reasoning, their
decisions satisfy Savage’s sure-thing principle, which states that if individuals prefer x over y for
every possible decision outcome, then they should prefer x over y when the decision outcome is
uncertain (Savage 1954). To illustrate, consider an individual who decides whether or not to
cooperate in prisoner’s dilemma. If the individual prefers non-cooperation when she is certain
that the other player will cooperate and if she also prefers non-cooperation when she is certain
that the other player will not cooperate, then the sure-thing principle states that she should prefer
non-cooperation when she is uncertain about whether or not the other player will cooperate.
Psychology research on non-consequential reasoning, however, provides evidence that
individuals sometimes fail to use consequential reasoning when they make decisions under
uncertainty (Bastardi and Shafir 2000; Croson 1999; Shafir and Tversky 1992; Tversky and
Shafir 1992). When they use non-consequential reasoning to make decisions, individuals do not
appropriately consider all of the relevant decision outcomes and their expected likelihoods. As a
result, they weight decision outcomes differently and make different decisions than when they
use consequential reasoning. Non-consequential reasoning manifests in a pattern of preferences
that violates Savage’s sure-thing principle. Specifically, individuals who use non-consequential
reasoning reveal that they prefer x over y for each possible decision outcome when outcomes are
certain, but reveal that they do not prefer x over y when there is uncertainty over outcomes.
Shafir and Tversky (1992) propose that the effects of non-consequential reasoning on
decision-making are attributable to the effects of uncertainty on individuals’ self-assessments of

16

their preferences. Specifically, they argue that individuals assess their preferences differently
when they make decisions under certainty because they use non-consequential reasoning and do
not appropriately consider all relevant decision outcomes. Shafir and Tversky (1992) also
propose that non-consequential reasoning can explain non-trivial levels of cooperation in
prisoner’s dilemma. They predict and find that players in prisoner’s dilemma who do not know
the other player’s action choice when they choose their own action are more likely to cooperate
than players who know with certainty what the other player’s action choice was when they
choose their own action. When there is uncertainty about the other player’s action choice, the
game’s outcome depends on both players’ decisions and thus players frame the decision as a
group decision. As a result, they prefer cooperation because it leads to the best group outcome.
In contrast, when there is certainty about the other player’s action choice, the game’s outcome
depends only on the decision of an individual player and players thus frame the decision as an
individual decision. As a result, they prefer non-cooperation because it leads to the best
individual outcome.
1.4 Hypothesis Development
I use the psychology theory of non-consequential reasoning (Shafir and Tversky 1992;
Tversky and Shafir 1992) to develop a hypothesis about the effect of accounting information
aggregation on managerial cooperation when individuals face incentives against cooperation. I
consider a setting in which individuals play a repeated one-shot prisoner’s dilemma game and
payoffs (i.e., profits) are common knowledge. Individuals are randomly re-matched in each
period such that they are matched with the same individual only once and they do not know the
number of periods of play. In the final period, each individual is provided with accounting

17

information about the other’s prior-period profit when there was a cooperation problem (i.e.,
when one but not both individuals chose the cooperative action).

7

When accounting information is aggregated (disaggregated), each individual is provided
with the other individual’s prior-period joint (individual) profit. Aggregated accounting
information does not allow the individuals to infer the other individual’s prior-period action with
certainty because the aggregation of profits in a period in which there was a cooperation problem
obfuscates individual profits in the prior period. Disaggregated accounting information, however,
allows each individual to infer with certainty the other’s prior-period action because
disaggregation reveals prior-period individual profits thereby allowing each individual to
attribute prior-period actions to the individuals who chose them.
When disaggregated accounting information eliminates uncertainty, I argue that
individuals will use consequential reasoning to make their decisions. I expect that they will
frame the decision to cooperate as an individual decision and choose non-cooperation because it
is the only action that leads to the highest individual profit, regardless of whether or not they
expect the other individual to cooperate. In contrast, when aggregated accounting information
does not allow individuals to infer the cooperativeness of the other’s prior-period action with
certainty, I argue that they will use non-consequential reasoning to make their decisions.
7

I restrict my hypothesis development to situations in which there are prior-period cooperation
problems because cooperation problems present an important challenge for individuals who wish
to condition their cooperation decisions on the cooperativeness of other individuals’ prior
actions. In the event of mutual prior-period cooperation or non-cooperation, the role of
accounting information in reducing uncertainty about the cooperativeness prior actions is muted
if not irrelevant. This is because mutual cooperation (mutual non-cooperation) yields the highest
(lowest) joint profit that individuals can earn. In these situations, individuals do not need
disaggregated accounting information to be able to infer with certainty the cooperativeness of the
other individual’s prior-period actions. When there are prior-period cooperation problems,
however, individuals cannot infer from prior-period joint profit whether the other individual’s
prior-period action was cooperative or non-cooperative. As a result, the role of accounting is
likely to be especially important when there are prior-period cooperation problems.
18

Consistent with the psychology theory of non-consequential reasoning, I expect that they will
frame the decision to cooperate as a group decision and choose cooperation because it is the only
action that leads to the highest joint profit.
HYPOTHESIS: When individuals face incentives against cooperation,
cooperation will be higher when accounting information is aggregated than when
it is disaggregated.

19

CHAPTER 2: EXPERIMENTAL DESIGN
2.1 Participants
Forty one upper-level, undergraduate business students at a large public university
participate in the experiment, which is conducted in seven sessions on multiple days. A
comparison of the dependent variable among the seven sessions does not provide evidence of a
session-specific effect (p = .760, two-tailed). Participants in each session work independently in
a computer laboratory using personal computers and each session lasts approximately 40
minutes. Communication among participants during the sessions is prohibited.
2.2 Experimental Design and Treatment Conditions
The experiment has a 2 x 5 mixed-factorial design. Accounting information aggregation
is manipulated between-subjects at two levels (aggregated and disaggregated) and period is
manipulated within-subjects at five levels. The dependent variable, managerial cooperation, is
measured as the proportion of participants in a treatment condition who choose the cooperative
action in period 5, which is the only period in which participants are provided with accounting
information about the prior period.
Profit interdependence is operationalized using a repeated one-shot prisoner’s dilemma
game in which participants simultaneously and independently choose either the cooperative or
non-cooperative action (Figure 1). The profit structure is such that each participant is better off
choosing the non-cooperative action no matter what she thinks the other participant will choose.
If both participants choose the non-cooperative action, however, then they will each earn lower
profits ($2, $2) than they would have earned if each had chosen the cooperative action ($6, $6).
The crucial feature of the game is that each participant can earn the highest profit ($8) when she
chooses the non-cooperative action while the other participant chooses the cooperative action.

20

The action labels are purposely abstract because they can influence participants’ decisions to
cooperate in laboratory experiments (Haynes and Kachelmeier 1998; Liberman, Samuels, and
Ross 2004; Rankin, Van Huyck, and Battalio 2000). Abstract operationalization of the action
labels minimizes the potentially confounding influence of value-laden labels such as “cooperate”
on the dependent variable.
Accounting information aggregation is operationalized in period 5 by manipulating the
mapping of participants’ period 4 actions to their period 4 profits. When accounting information
is aggregated, the actions of both participants in period 4 are mapped to one profit amount (i.e.,
joint profit) (Table 1A). When accounting information is disaggregated, the actions of each
participant in period 4 are mapped to one profit amount (i.e., division profit). As a result, each
participant in period 5 can infer the cooperativeness of the other participant’s prior-period action
with certainty when accounting information is disaggregated but not when it is aggregated.
Because disaggregated accounting information can reveal either cooperative or non-cooperative
prior-period actions, participants in the disaggregated condition in period 5 are randomly
matched with participants whose period 4 action was either cooperative or non-cooperative
(Table 1B and 1C).
2.3 Procedure and Task
Figure 2 illustrates the order of events in the experiment and provides an overview of the
task by period. The experiment has three parts. In Part 1, participants read instructions that
explain their role in the experiment and their compensation. In Part 2, they make management
decisions in each of five experimental periods. In Part 3, they respond to post-experiment
questions related to demographic information, factors that influenced their management
decisions, and the manipulation of the independent variable (Figure 2A).

21

2.3.1 Part 1: Instructions
Participants assume the role of division managers who choose a management action for
their own division in each of five experimental periods in which there are incentives against
managerial cooperation. Their compensation is comprised of a salary of $10 and a bonus of 50%
of their division’s profit in two periods, which are pre-determined by the experimenter but not
identified to participants until the beginning of those periods (i.e., before participants choose an
action for their division in those periods).8 Before the experimental periods commence,
participants are informed that their division’s profit in each period depends not only on the
management action that they choose for their own division, but also on the management action
that another division manager chooses for his or her division. Participants are also informed that
that there will not be profit interdependence between the same two divisions in more than one
period (i.e., that they will not be matched with the same participant more than once) and they are
provided with an example of a profit table in order to illustrate profit interdependence.

9

Next, participants are informed that another participant in the experiment will assume the
role of the other division manager in each period except for one of the periods. In one of the
periods, participants are informed that the role of the other division manager will be assumed by
8

Participants’ compensation is a function of their action choices in periods 4 and 5 (i.e., the final
period and the period preceding the final period). Providing incentive compensation for their
decisions in these two periods is intended to motivate participants to allocate as much attention
and effort as possible to their decisions in these periods. Participants’ decisions in these periods
are important because they directly influence the dependent variable.
9

Participants are informed that the profit table example that they are provided illustrates the
format in which division profits will be presented in the five experimental periods in which there
is profit interdependence. The example is intended to familiarize participants with the normal
form representation of division profits. All participants are provided with the same profit table
example. In order to minimize carryover effects, the profit table example that is provided to
participants illustrates a division profit structure that is not a prisoner’s dilemma.
22

a computer that is programmed to choose the management action that other participants facing
the same conditions made in a prior administration of the experiment.

10

Participants are

informed that the period in which a computer will assume the role of the other manager will not
be identified. In order to control for end-game effects, participants are not informed that the
experiment has five periods or that accounting information about the other participant’s priorperiod profit will be available in any period.
After reading the instructions but before the experimental periods begin, participants
complete a pre-experiment quiz on which they are required to score 100% to ensure that they
understand the task, their compensation, and the form of profit interdependence (i.e., the
structure of division profits in each period). After successful completion of the quiz, the first of
five experimental periods begins.
2.3.2 Part 2: Experimental Periods
In periods 1 – 4, participants make management decisions for their divisions without
accounting information about the other participant’s prior-period profit. In each of these periods,
each participant’s profit depends on the decision of another participant who makes a
management decision for another division (Figure 2B). At the beginning of each period, the
division with which each participant’s profit is interdependent is identified by division name
10

In the period in which a computer assumes the role of the other division manager, the
computer’s decision rule is determined by data that is collected in a pilot study. The pilot study
has a 2 x 4 mixed-factorial design in which accounting information aggregation is manipulated
between-subjects at two levels (aggregated and disaggregated) and period is manipulated withinsubjects at four levels. The pilot study has two purposes. First, it allows me to observe the
frequency with which each of the four possible decision outcomes occurs in each period. Second,
it allows me to match participants in the disaggregated condition in period 5 of my dissertation
experiment with a mechanical division manager that is programmed to make the same decision
that participants who chose either the cooperative or non-cooperative action in period 4 of the
pilot study made when facing the same accounting information aggregation and profit
interdependence conditions.
23

only (e.g., Division B), which ensures that participants are matched anonymously in each period.
Each participant is then provided with a profit table that illustrates the division profit structure
for that period. Next, each participant chooses a management action (i.e., Action X or Y) for his
or her own division. After submitting their action choice, participants are directed to a
confirmation screen that displays their action choice and the division profit they would earn
given their action choice if the other manager chose Action X for the other division and if the
other manager chose Action Y for the other division. Next, participants either confirm or revise
or their action choice. If they choose to revise their choice, then they are redirected to the action
choice screen. If they confirm their choice, then they are directed to a waiting screen if the other
participant has not yet chosen and confirmed a management action. After each participant
confirms her management action, she is informed of her own profit as well as the other
participant’s profit for that period.
In period 5, participants continue to assume the role of a division manager, but a
computer assumes the role of the other division manager. The task in period 5 is the same as the
task in periods 1 – 4, with the following two exceptions (Figure 2C). First, participants are
informed at the beginning of period 5 that they will be provided with accounting information
about the prior-period profit of the other division (i.e., the division with which there is profit
interdependence in period 5). Participants are then provided with the other division’s priorperiod profit table and either aggregated or disaggregated accounting information about the other
division’s prior-period profit. Second, after participants make their period 5 management
decisions but before they learn the period 5 outcome, they respond to three questions. First, they
respond to a question about the informativeness of the accounting information that was provided
to them at the beginning of period 5. Second, they respond to a question that is intended to elicit

24

their beliefs about the likelihood that the other division manager would choose the cooperative
action in period 5. Third, they respond to a question that is intended to elicit their beliefs about
the other participant’s beliefs that they would choose the cooperative action in period 5.
2.3.3 Part 3: Post-experiment
After participants complete all five experimental periods, they respond to postexperiment questions related to demographic information, factors that influenced their
management decisions, and the manipulation of accounting information aggregation. Participants
are paid immediately following the experiment and their average earnings are $13.83 for
approximately 40 minutes of participation.

25

CHAPTER 3: RESULTS
This chapter first reports the results of manipulation checks and preliminary analysis.
Next, it reports descriptive statistics and the results of the hypothesis test. Finally, it reports the
results of two supplementary analyses that provide evidence about my proposed causal
explanation for the hypothesized relation between accounting information aggregation and
managerial cooperation.
3.1 Manipulation Checks
To test whether the accounting information aggregation manipulation is successful,
participants indicate the extent to which they agree or disagree with the following statement:
“From the information I was given about the other division manager, I knew exactly what his or
her division profit was in period 4.” This question is asked in period 5 after participants choose a
management action for that period, but before they learn the outcome of their decision. The
question has a response scale from 1 (completely disagree) to 7 (completely agree). The result of
a t-test indicates that the extent to which participants agree with the statement is significantly
higher when accounting information is disaggregated than when it is aggregated (means = 6.05,
2.60 respectively; t = 8.02, df = 39, p < 0.01, two-tailed, untabulated). This provides support that
the accounting information aggregation manipulation is successful.
Participants also respond to questions about their game theory experience and knowledge.
Fifteen participants (37%) previously studied game theory and eleven participants (27%)
previously participated in a research study in which they were a player in an economic game
such as prisoner’s dilemma. Participants rate their knowledge of game theory by responding to
the following question: “How do you rate your knowledge of game theory and/or experimental
economics?” The question has a response scale from 1 (no knowledge) to 7 (expert knowledge).

26

In preliminary analysis, I find no significant effects of game theory experience or knowledge on
the dependent variable among treatment conditions (all p-values > 0.22, untabulated). Therefore,
I do not include these variables in the hypothesis test.
3.2 Descriptive Statistics and Preliminary Analysis
Table 2 reports descriptive statistics. Forty-one participants are assigned to the two
treatment conditions. When accounting information is disaggregated and participants can infer
the cooperativeness of the other participant’s prior-period action with certainty, three (14%)
participants choose the cooperative action. When accounting information is aggregated and
participants cannot infer the cooperativeness of the other participant’s prior-period action with
certainty, eight (40%) of participants choose the cooperative action. This pattern of proportions
is directionally consistent with the predicted effects of accounting information aggregation on
managerial cooperation.
When accounting information is disaggregated, it can reveal either prior-period
cooperation or non-cooperation. I therefore randomly match participants in period 5 with a
participant whose prior-period action was either cooperative or non-cooperative. Figure 3
illustrates the mean proportions of managerial cooperation in period 5 when accounting
information is aggregated and when accounting information is disaggregated and participants are
matched with a participant whose period 4 action was either cooperative or non-cooperative.
When accounting information is disaggregated and reveals prior-period cooperation, eleven
participants (18%) choose the cooperative action. When accounting information is disaggregated
and reveals prior-period non-cooperation, ten participants (10%) choose the cooperative action.
When accounting information is aggregated, eight participants (40%) choose the cooperative
action.

27

The observed pattern of managerial cooperation reveals preferences that are consistent
with non-consequential reasoning. Notice that 10% percent of participants prefer to cooperate
when they know that the other participant did not previously cooperate and 18% of participants
prefer to cooperate when they know that the other participant previously cooperated. If
participants use consequential reasoning to determine their cooperation decisions as expected
utility theory assumes, then, by Savage’s sure-thing principle, between 10% and 18% of
participants should prefer to cooperate when they are uncertain about whether or not the other
player previously cooperated. When participants are uncertain about whether or not the other
player previously cooperated, however, greater than 18% of participants choose to cooperate.
In order to provide preliminary evidence about the effects of the cooperativeness of
participants’ prior-period actions on managerial cooperation when accounting information is
disaggregated, I use Fisher’s exact test of independence to test the significance of the difference
in the proportions of managerial cooperation in the disaggregated condition when participants are
matched with another participant whose prior action is cooperative or non-cooperative. When
disaggregated accounting information reveals prior-period cooperative (non-cooperative) action,
18% (10%) of participants choose the cooperative action. The result of the Fisher’s exact test
indicates that the difference in the proportions of managerial cooperation between prior-period
action conditions is not significant (p > 0.75, two-tailed). This provides some support for a single
disaggregated accounting condition that combines the cooperative and non-cooperative prior
action conditions, which allows me to increase the power of the main hypothesis test.
Based on the result of the Fisher’s exact test, I do not distinguish between disaggregated
accounting information that reveals prior-period cooperative and non-cooperative actions for the
remainder of the data analysis. Caution should be taken, however, in interpreting the result of the

28

test due to low power (60%). Although I use Fisher’s exact test to accommodate my small
sample sizes, additional data is required in order to increase the power of the test and strengthen
my inference about the possible effects of the cooperativeness of prior-period actions on
managerial cooperation when accounting information is disaggregated.
3.3 Hypothesis Test
The hypothesis predicts that managerial cooperation will be higher when accounting
information is aggregated than when it is disaggregated. To test the hypothesis, I use Pearson’s
Chi-square test for independence to test the significance of the difference in the proportions of
managerial cooperation between the aggregated and disaggregated accounting information
conditions. The result of the Chi-square test indicates that the difference in these proportions is
2

statistically significant (χ = 3.45, p = 0.032, df = 1, one-tailed); due to small sample size,
however, there is a risk that statistical significance is over-estimated. To address this risk, I
report the results of a Fisher’s exact test for independence and a Chi-square test with Yates’
continuity correction.11 The Fisher’s exact test is intended to accommodate small sample sizes
and Yates’ continuity correction is intended to prevent over-estimation of statistical significance
in small samples (Yates 1934). The result of the Fisher’s exact test provides marginal support for
the hypothesis that the proportions of managerial cooperation between the two aggregation
conditions are independent (p = 0.065, one-tailed). The result of the Chi-square test using Yates’
continuity correction provides marginal support for the hypothesis that managerial cooperation is
2

higher when accounting information is aggregated than disaggregated (χ = 2.27, p = 0.066, df =

11

Although Pearson’s Chi-square test for independence can over-estimate statistical
significance for small sample sizes (Greenwood and Nikulin 1996), Fisher’s exact test for
independence and Yates’ continuity correction can under-estimate statistical significance for
small sample sizes (Sokal and Rohlf 1981). I report the results of each test for completeness.
29

1, one-tailed). Taken together, these results provide some support for the hypothesis, although a
larger sample is desirable to increase the power of the hypothesis test.
3.4 Supplemental Analyses
Two supplemental analyses are intended to provide evidence that supports my proposed
causal explanation for the relation between accounting information aggregation and managerial
cooperation. First, I analyze the effects of accounting information aggregation on participants’
expectations about the likelihood that other participants will choose the cooperative action in
period 5. Second, I analyze factors that influence participants’ management decisions in period 5.
3.4.1 Expectations about the Likelihood of Cooperation
In order to provide evidence of participants’ expectations about the likelihood that the
other participant will choose the cooperative action in period 5, each participant responds to the
following question in period 5 after choosing a management action but before learning the
outcome of their action choice: “Which action did you expect the other division manager to
choose this period (i.e., in Period 5)?” Participants choose “Action X”, “Action Y”, or “I didn’t
have an expectation about which action the other manager would choose.” Nine (45%)
participants in the aggregated accounting condition and nine (43%) participants in the
disaggregated accounting condition expect the other participant to choose the cooperative
management action in period 5 (Table 3). The result of a Pearson’s Chi-square test provides
2

evidence that this difference is not significant (χ = 0.019, p = 0.89, df = 1, two-tailed).
Although their expectations about the likelihood of cooperation are similar when
accounting information is aggregated and when it is disaggregated, the proportion of participants
who respond with cooperative action is significantly higher when accounting information is
aggregated than disaggregated. When accounting information is aggregated, 45% of participants
30

expect the other participant to choose the cooperative action and 89% of these participants
respond by choosing the cooperative action. When accounting information is disaggregated,
however, 43% of participants expect the other participant to choose the cooperative action and
33% of these participants respond by choosing the cooperative action. Despite similar
expectations about the likelihood of cooperation in the aggregated and disaggregated conditions
(i.e., 45% and 43% respectively), participants are more likely to choose the action that can lead
to the highest joint (individual) profit when accounting information is aggregated
(disaggregated). This is consistent with my proposed causal explanation for the relation between
accounting information aggregation and managerial cooperation.
3.4.2 Factors that Influence Participants’ Management Decisions
Shafir and Tversky (1992) propose that the pattern of managerial cooperation that I
observe in my experiment is attributable to the effect of uncertainty on individuals’ selfassessments of their preferences. Specifically, they propose that a player will be more likely to
frame the decision to cooperate in a prisoner’s dilemma game as a group decision when there is
uncertainty about the other player’s action choice. As a result, players will self-assess stronger
preferences for cooperation when there is uncertainty because cooperation is the only strategy
that can lead to the best group outcome (i.e., highest joint profit). When there is certainty over
prior actions, however, Shafir and Tversky (1992) argue that individuals will be more likely to
frame the decision to cooperate in prisoner’s dilemma as an individual decision. As a result,
these individuals will self-assess stronger preferences for non-cooperation because it is the only
strategy that leads to the best individual outcome (i.e., highest individual profit).
In order to provide evidence about participants’ self-assessments of their preferences,
they respond to the following post-experiment question about factors that influenced their

31

management decisions in period 5: “When you considered whether to choose Action X or Action
Y in this period (i.e., period 5), how important were the following factors?” Participants
allocated 100 points to the factors that influenced their management decision; they allocated
more points to factors that had more influence on their decision and fewer points to factors that
had less influence on their decision (Table 4A).
When accounting information is aggregated, 89% of participants who expect cooperation
respond by choosing the cooperative action. Consistent with these participants having stronger
preferences for the best group outcome, they allocate, on average, 92 out of 100 points to the
following decision factor: “Action X was attractive to me because I wanted the highest total
profit for me and the other division manager (i.e., Action X could maximize the sum of our two
divisions’ profits)” (Table 4B). When accounting information is disaggregated, however, only
33% of participants who expect cooperation respond by choosing the cooperative action. These
participants allocate, on average, 78 out of 100 points to this decision factor. The result of a t-test
indicates that the difference in mean points allocated to this decision factor is significantly higher
when accounting information is aggregated than when it is disaggregated (t = 3.07, df = 8, p <
0.01, one-tailed). This result is consistent with participants in the disaggregated condition having
weaker preferences for the best group outcome than participants in the aggregated condition.
When accounting information is disaggregated, 66% of participants who expect
cooperation respond by choosing the non-cooperative action. Consistent with these participants
having stronger preferences for the best individual outcome, they allocate, on average, 87 out of
100 points to the following factor: “Action Y was attractive to me because I wanted my division
to earn the highest profit.” When accounting information is aggregated, however, only one
participant (11%) who expects cooperation responds by choosing the non-cooperative action.

32

This participant allocates only 60 out of 100 points to the same decision factor. The result of a
one-sample t-test provides evidence that the mean points allocated to this decision factor is
significantly higher than when accounting information is disaggregated than aggregated (t =
4.33, df = 5, p < 0.004, one–tailed). Specifically, the result of the one-sample t-test provides
supports for rejecting the null hypothesis that there is no difference between the mean points
allocated to this decision factor when accounting is aggregated (i.e., 60 points) and when it is
disaggregated (i.e., 87 points). This result is consistent with participants in the disaggregated
condition having stronger preferences for the best individual outcome than in the aggregated
condition.
Taken together, these results provide support for my proposed causal explanation.
Although participants in the aggregated and disaggregated accounting conditions face the same
financial incentives and have the same expectations about the likelihood that the other participant
will cooperate, managerial cooperation is significantly higher when accounting is aggregated
than when it is disaggregated. Consistent with the psychology theory of non-consequential
reasoning, participants are more likely to prefer cooperation when accounting information is
aggregated and there is uncertainty about other participants’ prior-period actions because
participants frame their decisions to cooperate as group decisions and self-assess strong
preferences for actions that lead to the best group outcome.

33

CHAPTER 4: CONCLUSION
4.1 Summary of Experiment and Results
Accounting research provides evidence that accounting information can increase
managerial cooperation when it allows managers build reputations for cooperation or noncooperation by allowing them to infer the cooperativeness of other managers’ prior actions with
certainty (Coletti et al. 2005; Hales and Williamson 2010; Schwartz et al. 2000; Schwartz and
Young 2002). Reputation-building, however, may be hindered and managers may be uncertain
about whether or not other managers have previously cooperated when accounting information is
a noisy signal of prior actions due to its aggregation (Datar and Gupta 1994; Ijiri 1967; Lev
1969; Sorter 1969). I investigate the effects of accounting information on managerial cooperation
when its aggregation does not eliminate managers’ uncertainty about the cooperativeness of
other managers’ prior-period actions. To the extent that increased managerial cooperation
depends on managers’ ability to build reputations for cooperation or non-cooperation, prior
research has conjectured that noisy accounting information may lead to decreased managerial
cooperation by hindering reputation-building (Schwartz and Young 2002). In my dissertation,
however, I provide theory-consistent experimental evidence that noisy accounting information
can increase rather than decrease managerial cooperation by influencing whether individuals
frame their cooperation decisions as either individual or group decisions.
The pattern of managerial cooperation that I observe in my experiment is consistent with
psychology research on non-consequential reasoning (Croson 1999; Shafir and Tversky 1992;
Tversky and Shafir 1992). When individuals are provided with aggregated accounting
information that does not allow them to infer others’ prior-period actions with certainty, they
make their cooperation decisions using non-consequential reasoning and do not appropriately

34

consider all relevant decision outcomes when they make their decisions. In these situations,
individuals frame their decisions as group decisions and self-assess strong preferences for
cooperation because it is the only action that can lead to the best group outcome. In contrast,
when individuals are provided with disaggregated accounting information that allows them to
infer others’ prior-period actions with certainty, they make their cooperation decisions using
consequential reasoning and appropriately consider all relevant decision outcomes when they
make their decisions. In these situations, individuals frame their decisions as individual decisions
and self-assess strong preferences for non-cooperation because it is the only action that can lead
to the best individual outcome.
4.2 Limitations and Suggestions for Future Research
There are at least three limitations of my dissertation. First, I make the simplifying
assumption that managers know not only how their decisions will influence their own profits, but
also how their decisions will influence the profits of the other organizational subunit with which
their profits are interdependent (i.e., I assume that payoffs are common knowledge). Although
this assumption is conventional and consistent with prior accounting and experimental
economics research on managerial cooperation (Camerer 2003), it may limit the generalizability
of my results. In natural environments, managers may know the effects of their decisions on their
own organizational subunits’ profits, but it is unlikely that they will know exactly how their
decisions will influence the profits of other organizational subunits or how other organizational
subunit managers’ decisions will influence their own profits.
Second, I manipulate accounting information aggregation at the end-points of the
aggregation continuum. This is a strong manipulation and is therefore useful in a laboratory
environment because it provides a strong test of my hypothesis. Accounting information that is

35

provided to managers in natural environments, however, is unlikely to be either fully aggregated
or fully disaggregated. As a result, my aggregation manipulation may limit the generalizability of
my results. Further, it is possible that disaggregation may not increase the informativeness of
accounting information when the underlying information is aggregated in its natural form (e.g.,
when aggregated cost information such as annual lease expense is disaggregated by allocating
the aggregated cost to different cost objects).
Third, I hold profit interdependence constant at a level that assumes managers face
dominant financial incentives against cooperation. Managers, however, may not always face
dominant incentives against cooperation. For example, when managers can build reputations and
reward (punish) other managers for their prior cooperation (non-cooperation), incentives for noncooperation can decrease such that cooperation is no longer a strictly dominated strategy. In
these situations, a manager’s best response can include cooperation (Camerer 2003, 2006;
Rankin et al. 2000; Van Huyck, Battalio, and Bell 1990). Experimental economics research
suggests that situations in which cooperation can be a manager’s best response are common and
may be richer decision contexts that have greater generalizability than situations in which
individuals’ best response can never include cooperation (Harsanyi and Selten 1988). Future
accounting research could explore the effects of accounting information aggregation on
managerial cooperation when the form of profit interdependence is such that a manager’s best
response can include cooperation.

36

APPENDIX

37

FIGURE 1: Profit Interdependence as a Prisoner’s Dilemma Game
Normal-form representation of division profit structure as a prisoner’s dilemma game
Manager B
Action X

Action Y

Action X

$6, $6

$0, $8

Action Y

$8, $0

$2, $2

Manager A

Supplemental information: The first number in each cell represents Manager A’s profit and the
second number in each cell represents Manager B’s profit. The sum of the two numbers in each
cell is the joint profit that is associated with a decision outcome. This representation of division
profits assumes that the division profits that are associated with the cooperative and noncooperative actions (Actions X and Y, respectively) represent the actual utility payoffs to
managers who choose these actions. That is, this payoff structure assumes that managers do not
have non-monetary preferences. If managers have non-monetary preferences (e.g., for
cooperative outcomes), then these preferences may change the utility payoffs to the action
choices.

38

FIGURE 2A: Order of Events
Order of events for Part 1, Part 2, and Part 3 of the experiment

PART 1

PART 2
Periods 1 – 4

Instructions and Quiz

PART 3
Period 5

Aggregated Accounting
Information

No Accounting
Information

Disaggregated Accounting
Information

Post-Experiment

Supplemental information: After reading the instructions and completing a quiz about the instructions (PART 1), participants begin
the five experimental periods (PART 2). In each period, they make a management decision that influences their own division’s profit
as well as the profit of another division. In periods 1 – 4, participants make their management decision without accounting information
about the other division’s prior-period profit. In period 5, participants are provided with either aggregated or disaggregated accounting
information about the other division’s prior-period profit before they make their management decision. After completing the five
experimental periods, participants complete post-experiment questions (PART 3).

39

FIGURE 2B: Task Overview (Periods 1 – 4)
Experimental task for periods 1 – 4
SCREEN 1
Profit Interdependence
Participants are informed that their
division’s profit will depend on the
profit of another division, whose
manager will be another participant
who will choose a management action
for that division
Profit Table
Each participant is provided with a
profit table that represents the profit
structure created by profit
interdependence

SCREEN 2

SCREEN 3

SCREEN 4

Confirmation
Participants either
confirm or revise their
action choice

Waiting
Participants wait for
the other participant
to choose an action

Outcome
Participants learn their
own profit and the
other participant’s
profit for the period

If they confirm their
choice, then they are
directed to SCREEN 3
If they choose to revise
their action choice, then
they are redirected to
SCREEN 1

Action Choice
Participants independently choose
either Action X or Y for their division

Supplemental information: At the beginning of periods 1 – 4, participants are provided with the profit table and the name of the
division (e.g., Division B) with which profits are interdependent for that period. They are informed that the same profit table applies to
their division as well as the division with which there is profit interdependence for that period. Next, they choose either Action X
(cooperative action) or Action Y (non-cooperative action) for their own division. After they choose an action, they either confirm or
revise their action choice. If they choose to revise their action choice, then they are redirected to Screen 1. If they confirm their action
choice, then they are directed either to the waiting screen (Screen 3) or to the outcome screen (Screen 4). After both participants
choose and confirm an action, they are directed to the outcome screen, which provides information about each participant’s profit for
that period.
40

FIGURE 2C: Task Overview (Period 5)
Experimental task for period 5
SCREEN 1
Profit Interdependence
Participants are informed
that their division’s profit in
Period 5 will still depend on
the profit of another
division, but that accounting
information about the other
division’s prior-period
profit will be available to
them before they choose a
management action for
period 5.
Accounting information
Participants are provided
with the other division’s
period 4 profit table and
accounting information
about the other division’s
period 4 profit

SCREEN 2

SCREEN 3

Profit Table
Participants are
provided with
the period 5
profit table

Confirmation
Participants
either confirm or
revise their
action choice

Action choice
Participants
independently
choose either
Action X or Y
for their division
for period 5

If they confirm
their choice, then
they proceed to
SCREEN 4

SCREEN 4
Expectations
Participants answer
three questions about
their expectations
about the other
manager’s action
choice given the
accounting
information provided
on SCREEN 1

SCREEN 5
Outcome
Participants learn
their own profit
and the other
division’s profit
for the period

If they choose to
revise their
action choice,
then they are
redirected to
SCREEN 2

Supplemental information: The task in period 5 is the same as in periods 1 – 4, with the following two exceptions. First, each
participant is provided with the other division’s period 4 profit table and accounting information about the other division’s period 4
profit. Second, each participant answers three questions about his or her beliefs about 1) the likelihood that the other participant would
choose the cooperative action in period 5, 2) the likelihood that the other participant would expect him or her to choose the
cooperative action in period 5, and 3) the certainty with which he or she could infer the other participant’s period 4 action from the
accounting information that was provided.

41

FIGURE 3: Preliminary Analysis
Mean managerial cooperation by accounting aggregation condition
Managerial Cooperation
.90
.80
.70
.60
.50
.40
.30
.20
.10

.40

.18
.10

Disaggregated
Disaggregated
(Cooperative) (Non-Cooperative)
(n = 11)
(n = 10)

Aggregated
(n = 20)

Accounting Information Aggregation

42

TABLE 1A: Accounting Information Aggregation (Aggregated Condition)
At the beginning of period 5, participants in the aggregated accounting information condition
were provided with aggregated accounting information about the organizational subunit with
which there was profit interdependence in that period.
Division T
Action X

Action Y

Action X

$6, $6

$0, $8

Action Y

$8, $0

$2, $2

Division F

Participants were provided with the following supplemental information: In this period, you will
be paired with Division Manager F. You have not been paired with Division Manager F before,
but you will be provided with some information about this manager before you make your
management decision in this period. Specifically, you will be provided with information about
Division Manager F’s profit in period 4 when he or she was paired with someone other than you.
This information may or may not help you predict what Division Manager F’s management
decision will be in this period when he or she is paired with you. Division Manager F will also be
provided with information about your profit in period 4.
Division F was paired with Division T in period 4 and the following profit information is
available about period 4: the sum of the profits of Division F and Division T was $8.
In period 4, Divisions F and T were both provided with the division profit table below. This
means that Division Manager F could have chosen either Action X or Action Y in period 4.

43

TABLE 1B: Accounting Information Aggregation (Disaggregated-Cooperative Condition)
At the beginning of period 5, participants in the disaggregated-cooperative accounting
information condition were provided with disaggregated accounting information about the
organizational subunit with which there was profit interdependence in that period. The
organizational subunit with which participants in this condition were paired chose the
cooperative management action in period 4.
Division T
Action X

Action Y

Action X

$6, $6

$0, $8

Action Y

$8, $0

$2, $2

Division F

Participants were provided with the following supplemental information: In this period, you will
be paired with Division Manager F. You have not been paired with Division Manager F before,
but you will be provided with some information about this manager before you make your
management decision in this period. Specifically, you will be provided with information about
Division Manager F’s profit in period 4 when he or she was paired with someone other than you.
This information may or may not help you predict what Division Manager F’s management
decision will be in this period when he or she is paired with you. Division Manager F will also be
provided with information about your profit in period 4.
Division F was paired with Division T in period 4 and the following profit information is
available about period 4: Division F’s profit was $0 and Division T’s profit was $8.
In period 4, Divisions F and T were both provided with the division profit table below. This
means that Division Manager F chose Action X in period 4.

44

TABLE 1C: Accounting Information Aggregation
(Disaggregated-Non-Cooperative Condition)
At the beginning of period 5, participants in the disaggregated-non-cooperative accounting
information condition were provided with disaggregated accounting information about the
organizational subunit with which there was profit interdependence in that period. The
organizational subunit with which participants in this condition were paired chose the noncooperative management action in period 4.

Division T
Action X

Action Y

Action X

$6, $6

$0, $8

Action Y

$8, $0

$2, $2

Division F

Participants were provided with the following supplemental information: In this period, you will
be paired with Division Manager F. You have not been paired with Division Manager F before,
but you will be provided with some information about this manager before you make your
management decision in this period. Specifically, you will be provided with information about
Division Manager F’s profit in period 4 when he or she was paired with someone other than you.
This information may or may not help you predict what Division Manager F’s management
decision will be in this period when he or she is paired with you. Division Manager F will also be
provided with information about your profit in period 4.
Division F was paired with Division T in period 4 and the following profit information is
available about period 4: Division F’s profit was $8 and Division T’s profit was $0.
In period 4, Divisions F and T were both provided with the division profit table below. This
means that Division Manager F chose Action Y in period 4.

45

TABLE 2: Descriptive Statistics
Number (%) of participants who choose the cooperative action in period 5

Accounting Information Aggregation
Disaggregated

Aggregated

3
(14%)
(n = 21)

8
(40%)
(n = 20)

46

TABLE 3: Supplemental Analysis I
Participants’ expectations about the likelihood of cooperative action in period 5

Expect
Cooperative Action

Choose
Cooperative Action

Aggregated
Accounting
Information

9
(45%)
(n = 20)

8
(40%)
(n = 20)

Disaggregated
Accounting
Information

9
(43%)
(n = 21)

3
(14%)
(n = 21)

47

TABLE 4A: Supplemental Analysis II (Post-Experiment Questionnaire Item)
In Part 3 of the experiment, participants respond to a post-experiment questionnaire item about
factors that influenced their management decision in period 5. Participants who expected the
other participant to choose Action X (i.e., the cooperative action) in period 5 allocated 100 points
to seven factors.
Division F
Action X

Action Y

Action X

$6, $6

$0, $8

Action Y

$8, $0

$2, $2

Your Division

Participants were provided with the following supplemental information: In period 5 (i.e., the
final period), you were paired with Division Manager F. In period 5, you and Division Manager
F were provided with the division profit table at the bottom of this screen. When you considered
whether to choose Action X or Action Y in for your division period 5, how important were the
following factors? (Please allocate 100 points to the factors below. Allocate more points to
factors that had more influence on your decision and allocate fewer points to factors that had less
influence on your decision.)
1. Action X was attractive to me because I wanted the highest total profit for me and the other
division manager (i.e., Action X could maximize the sum of our divisions’ profits).
2. Action X was attractive to me because I wanted to reward the other division manager for his
or her prior period management decision.
3. Action X was attractive to me because I wanted both divisions to earn the same profit.
4. Action Y was attractive to me because I wanted my division to earn the highest profit.
5. Action Y was attractive to me because I didn’t want to risk earning zero profit.
6. Action Y was attractive to me because I wanted to punish the other division manager for his
or her prior period management decision.
7. Other (please type the other factor that influenced your decision): _____________________

48

TABLE 4B: Supplemental Analysis II (Decision Factors)
Descriptive statistics about factors that influenced participants’ decisions in period 5. This table
reports the mean (standard deviation) points allocated by condition for participants who expected
the other participant to choose the cooperative action.

Factor
COOP-JOINT
COOP-REWARD

Aggregated Accounting Information
Choose Cooperation Choose Non-Cooperation
(n = 8)
(n = 1)
91.87
(8.45)
5.63
(6.23)

NON-INDIVIDUAL

-

NON-RISK

-

NON-PUNISH

-

EQUITY

1.25
(3.53)
1.25
(3.53)

OTHER
Total

60.00
aaaaaaaaa ( - )
20.00
aaaaaaaaa ( - )
20.00
aaaaaaaaa ( - )
-

100.00

100.00

49

Marginal
(n = 9)
81.67
(31.62)
5.00
(6.12)
6.67
(20.00)
2.22
(6.67)
2.22
(6.67)
1.11
(3.33)
1.11
(3.33)
100.00

TABLE 4B (cont’d)

Factor
COOP-JOINT
COOP-REWARD

Disaggregated Accounting Information
Choose Cooperation Choose Non-Cooperation
Marginal
(n = 3)
(n = 6)
(n = 9)
78.33
26.11
(10.41)
(39.51)
3.33
1.11
(5.77)
(3.33)

NON-INDIVIDUAL

-

NON-RISK

-

NON-PUNISH

86.67
aaaaaaa (15.05 )
5.00
aaaaaaaa(8.36)
6.67
aaaaa(8.16)
1.67
(4.08)
100.00

6.67
(11.54)
8.33
(2.88)
3.33
(5.77)

EQUITY
OTHER
Total

100.00

57.77
(44.93)
3.33
(7.07)
6.67
(8.66)
2.77
(4.41)
2.22
(4.41)
100.00

Participants were provided with the following supplemental information: COOP-JOINT =
Action X was attractive to me because I wanted the highest total profit for me and the other
division manager (i.e., Action X could maximize the sum of our two divisions’ profits). COOPREWARD = Action X was attractive to me because I wanted to reward the other division
manager for his or her prior period management decision. NON-INDIVIDUAL= Action Y was
attractive to me because I wanted my division to earn the highest profit. NON-RISK = Action Y
was attractive to me because I didn’t want to risk earning zero profit for my division. NONPUNISH = Action Y was attractive to me because I wanted to punish the other division manager
for his or her prior period management decision. EQUITY = Action X was attractive to me
because I wanted both divisions to earn the same profit. OTHER = Please type the other factor
that influenced your decision in the space provided.

50

BIBLIOGRAPHY

51

BIBLIOGRAPHY

Antle, R., and J. Demski. 1988. The controllability principle in responsibility accounting. The
Accounting Review 63: 700-718.
American Institute of Certified Public Accountants (AICPA). 1975. The meaning of “present
fairly in conformity with Generally Accepted Accounting Principles” in the independent
auditor’s report. Statement on Auditing Standards No. 5. New York: AICPA
Axelrod, R. 1984. The Evolution of Cooperation. New York: Basic Books.
Banker, R. D., and S. M. Datar. 1989. Sensitivity, precision, and linear aggregation of signals for
performance evaluation. Journal of Accounting Research 27 (1): 21-39.
Bastardi, A., and E. Shafir. 2000. Nonconsequential reasoning and its consequences. Current
Directions in Psychological Science 9 (6): 216-219.
Bolton, G. E., E. Katok, and A. Ockenfels. 2005.Cooperation among strangers with limited
information about reputation. Journal of Public Economics 89: 1457-1468.
Bouwens, J., and M. A. Abernethy. 2000. The consequences of customization on management
accounting system design. Accounting, Organizations, and Society 25 (3): 221-241.
Briers, M., and W. F. Chua. 2001. The role of actor-networks and boundary objects in
management accounting change: A field study of an implementation of activity-based
costing. Accounting, Organizations and Society 26: 237-269.
Bushman, R. M., Indjejikian, R. J., and A. Smith. 1995. Aggregate performance measures in
business unit manager compensation: the role of intrafirm interdependencies. Journal of
Accounting Research Supplement 33: 101–134.
Caglio, A., and A. Ditillo. 2008. A review and discussion of management control in inter-firm
relationships: Achievements and future directions. Accounting, Organizations and
Society 33: 865-898.
Camerer, C. F. 2003. Behavioral Game Theory: Experiments in Strategic Interaction. New
Jersey: Princeton University Press.
Camerer, C. F. 2006. Behavioral Economics. In: Advances in Economics and Econometrics:
Theory and Applications, Ninth World Congress, Vol. II, edited by R. Blundell, W. K.
Newey, and T. Persson. Cambridge: Cambridge University Press.
Charness, G., and M. Rabin. 2002. Understanding social preferences with simple tests. The
Quarterly Journal of Economics 117: 817-869.

52

Christ, M. H., K. L. Sedatole, K. L. Towry, and M. A. Thomas. 2008. When formal controls
undermine trust and cooperation. Strategic Finance January: 38-44.
Coletti, A. L., K. L. Sedatole, and K. L. Towry. 2005. The effect of control systems on trust and
cooperation in collaborative environments. The Accounting Review 80: 477-500.
Cooper, R., D. V. DeJong, R. Forsythe, and T. W. Ross. 1992. Communication in coordination
games. Quarterly Journal of Economics 107: 739-773.
Croson, R. T. A. 1999. The disjunction effect and reason-based choice in games. Organizational
Behavior and Human Decision Processes 80: 118-133.
Datar, S., and M. Gupta. 1994. Aggregation, specification, and measurement errors in product
costing. The Accounting Review 69: 567-591.
Dal Bo, P. 2005. Cooperation under the Shadow of the Future: Experimental evidence from
infinitely repeated games. The American Economic Review 95 (5): 1591-1604.
Das, T. K., and B. Teng. 1996. Risk types and inter-firm alliance structures. Journal of
Management Studies 33: 827-843.
Das, T. K., and B. Teng. 1999. Managing risks in strategic alliances. Academy of Management
Executive 13: 50-62.
Das, T. K., and B. Teng. 2000. Instabilities of strategic alliances: An internal tensions
perspective. Organization Science: A Journal of the Institute of Management Sciences 11:
77-101.
Dawes, R. 1980. Social dilemmas. Annual Review of Psychology 31: 169-193.
Dekker, H. C. 2004. Control of inter-organizational relationships: Evidence on appropriation
concerns and coordination requirements. Accounting, Organizations, and Society 29 (1):
27-49.
Devetag G., and A. Ortmann. 2007. When and why? A critical survey on coordination failures in
the laboratory. Experimental Economics 10: 331-344.
Dickhaut, J., and K. McCabe. 1997. The behavioral foundations of stewardship accounting and a
proposed program of research: What is accountability? Behavioral Research in
Accounting 9: 60-87.
Drake, A. R., and S. F. Haka. 2008. Does ABC information exacerbate hold-up problems in
buyer-supplier negotiations. The Accounting Review 83 (1): 29-60.

53

Drake, A. R., S. F., Haka, and S. P. Ravenscroft. 1999. Cost system and incentive structure
effects on innovation, efficiency, and profitability in teams. The Accounting Review 74:
323-346.
Galbraith, J. 1973. Designing Complex Organizations. Reading, MA: Addison-Wesley
Publishing
Greenwood, P. E., and M. S. Nikulin. 1996. A guide to chi-squared testing. New York: Wiley
Hales, J., and M. G. Williamson. 2010. Implicit employment contracts: The limits of
management reputation for promoting firm productivity. Journal of Accounting Research
48 (1): 51-80.
Hammond, P. 1988. Consequentialist foundations for expected utility. Theory and Decision 25:
25-78.
Hannan, R. K., K. L. Towry, and Y. Zhang. The effect of mutual monitoring on effort in
tournament contracts. Forthcoming in Contemporary Accounting Research.
Harsanyi, J. C., and R. Selten. 1988. A General Theory of Equilibrium Selection in Games.
Cambridge, MA: MIT Press.
Haynes, C. M., and S. J. Kachelmeier. 1998. The effects of accounting contexts on accounting
decisions: A synthesis of cognitive and economic perspectives in accounting
experimentation. Journal of Accounting Literature 17: 97-136.
Heide, J. B., and A. S. Miner. 1992. The shadow of the future: Effects of anticipated interaction
and frequency of contact on buyer-seller cooperation. The Academy of Management
Journal 35 (2): 265-291.
Holmstrom, B. 1979. Moral hazard and observability. The Bell Journal of Economics 10: 74-91.
Horngren, C. T., S. M. Datar, and M. Rajan. 2012. Cost Accounting: A Managerial Emphasis.
14th ed. Boston: Prentice Hall.
Ijiri, Y. 1967. The Foundations of Accounting Measurement: A Mathematical, Economic, and
Behavioral Inquiry. Boston: Prentice Hall.
John, L. K., G. Loewenstein, and S. I. Rick. 2013 Working Paper. Cheating more for less:
Upward social comparisons motivate the poorly compensated to cheat. Available at
SSRN: http://ssrn.com/abstract=2208139 or http://dx.doi.org/10.2139/ssrn.2208139.
Kachelmeier, S. J., and M. Shehata. 1997. Internal auditing and voluntary cooperation in firms:
A cross-cultural experiment. The Accounting Review 72: 407-432.

54

Keating, A. S. 1997. Determinants of divisional performance evaluation practices. Journal of
Accounting and Economics 24: 243–274.
King, R. 1996. Reputation formation for reliable reporting: An experimental investigation. The
Accounting Review 71 (3): 375-396.
Kollock, P. 1993. An eye for an eye leaves everyone blind: Cooperation and accounting systems.
American Sociological Review 58: 768-786.
Lev, B. 1969. Accounting and Information Theory. Studies in Accounting Research, No. 2.
American Accounting Association.
Levi, I. 1991. Consequentialism and sequential choice. In M. Bacharach and S. Hurley (Eds.),
Foundations of decision theory: Issues and advances (pg. 92-122). Oxford: Basil
Blackwell
Liberman, V., S. M. Samuels, and L. Ross. 2004. The name of the game: Predictive power of
reputations versus situational labels in determining prisoner’s dilemma game moves.
Personality and Social Psychology Bulletin 30: 1175-1185.
Mayhew, B. 2001. Reputation building and the demand for auditing. Journal of Accounting
Research 39 (3): 599-617.
McCarter, M. W., and G. B. Northcraft. 2007. Happy together? Insights and implications of
viewing managed supply chains as a social dilemma. Journal of Operations Management
25: 498-511.
Melumad, N., D. Mookherjee, and S. Reichelstein. 1992. A theory of responsibility centers.
Journal of Accounting and Economics 15: 445-484.
Nikias, A. D., S. T. Schwartz, E. E. Spires, J. R. Wollscheid, and R. A. Young. 2010. The effects
of aggregation and timing on budgeting: An experiment. Behavioral Research in
Accounting 22 (1): 67-83.
Nowak, M. A., and K. Sigmund. 2005. Evolution of indirect reciprocity. Nature 437: 1291-1298.
Pick, J. 1971. Is responsibility accounting irresponsible? The New York Certified Public
Accountant (July): 487-494.
Rankin, F. W., J. B. Van Huyck, and R. C. Battalio. 2000. Strategic similarity and emergent
conventions: Evidence from similar stag hunt games. Games and Economic Behavior 32:
315-337.
Rapoport, A. 1988. Experiments with n-person social traps I: Prisoner’s dilemma, weak
prisoner’s dilemma, volunteer’s dilemma, and largest number. Journal of Conflict
Resolution 32 (3): 457-472.
55

Rapoport, A. 1989. Prisoner’s Dilemma. In J. Eatwell, M. Milgate, and P. Newman (Eds.). The
New Palgrave Game Theory (pg. 199 – 204). New York: W.W. Norton.
Rapoport, A., and A. Chammah. 1965. Prisoner’s Dilemma. Ann Arbor: University of Michigan
Press.
Rowe, C. 2004. The effect of accounting report structure and team structure on performance in
cross-functional teams. The Accounting Review 79: 1153-1180.
Rowe, C., M. D. Shields, and J. G. Birnberg. 2008. Effects of organizational process change on
responsibility accounting and managers’ revelations of private knowledge. Accounting,
Organizations and Society 33: 164-198.
Savage, L. J. 1954. The Foundations of Statistics. New York: Wiley and Sons.
Schwartz, S. T., and R. A. Young. 2002. A laboratory investigation of verification and reputation
formation in a repeated joint investment setting. Contemporary Accounting Research 19:
311-342.
Schwartz, S. T., R. A. Young, and K. Zvinakis. 2000. Reputation without repeated interaction: A
role for public disclosures. Review of Accounting Studies 5: 351-375.
Sorter, G. H. 1969. An “events” approach to basic accounting theory. The Accounting Review 44:
12-19.
Shafir, E., and A. Tversky. 1992. Thinking through uncertainty: Non-consequential reasoning
and choice. Cognitive Psychology 24: 449-474.
Skaperdas, S., and C. Syropoulos. 1996. Can the shadow of the future harm cooperation? Journal
of Economic Behavior and Organization 29: 355-372.
Sokal, R. R., and F. J. Rohlf. 1981. Biometry: The Principles and Practice of Statistics in
Biological Research. Oxford: W. H. Freeman.
Tayler, W. B., and R. J. Bloomfield. 2011. Norms, Conformity, and Controls. Journal of
Accounting Research 49: 753-790.
Taylor, M. 1976. Anarchy and Cooperation. New York: Wiley.
Taylor, M. 1987. The Possibility of Cooperation. Cambridge: Cambridge University Press.
Thompson, J. D. 1967. Organizations in Action. New York: McGraw-Hill.
Tomkins, C. 2001. Interdependencies, trust and information in relationships, alliances and
networks. Accounting, Organizations and Society 26: 161-191.
56

Towry, K. 2003. Control in a teamwork environment: The impact of social ties on the
effectiveness of mutual monitoring contracts. The Accounting Review 78 (4): 1069-1095.
Tversky, A., and E. Shafir. 1992. The disjunction effect in choice under uncertainty.
Psychological Science 3: 305-309.
Van Huyck, J. B., R. C. Battalio, and R. O. Bell. 1990. Tacit coordination games, strategic
uncertainty, and coordination failure. American Economic Review 80: 234-248.
Yates, F. 1934. Contingency table involving small numbers and the χ2 test. Supplement to the
Journal of the Royal Statistical Society 1: 217-235.
Young, M., J. Fisher, and T. Lindquist. 1993. The effects of intergroup competition and
intragroup cooperation on slack and output in a manufacturing setting. The Accounting
Review 68: 466-481.
Zeng, M., and X. P. Chen. 2003. Achieving cooperation in multiparty alliances: A social
dilemma approach to partnership management. Academy of Management Review 28: 587605.

57