Budgeting, Psychological Contracts, and Budgetary Slack Susanna Gallani Ranjani Krishnan

advertisement
Budgeting, Psychological Contracts,
and Budgetary Slack
Susanna Gallani
Eric J. Marinich
Ranjani Krishnan
Michael D. Shields
Working Paper 16-017
Budgeting, Psychological Contracts, and
Budgetary Slack
Susanna Gallani
Ranjani Krishnan
Harvard Business School
Michigan State University
Eric J. Marinich
Michael D. Shields
Miami University
Michigan State University
Working Paper 16-017
Copyright © 2015 by Susanna Gallani, Ranjani Krishnan, Eric J. Marinich and Michael D. Shields
Working papers are in draft form. This working paper is distributed for purposes of comment and discussion only. It may
not be reproduced without permission of the copyright holder. Copies of working papers are available from the author.
Budgeting, Psychological Contracts, and Budgetary Slack
SUSANNA GALLANI, Harvard Business School
RANJANI KRISHNAN, Michigan State University
ERIC J. MARINICH*, Miami University
MICHAEL D. SHIELDS, Michigan State University
July 2015
*Corresponding author: Eric J. Marinich
Address: 800 East High Street, 3094 FSB, Oxford, OH 45056
Phone: 1+(513) 529-8283, Fax: 1+(513) 529-6992
Email: marinie@miamioh.edu
An earlier version of this paper was presented at the 2012 Management Accounting Section
Research Conference and the 2012 American Accounting Association Annual Meeting. We
thank Jake Birnberg, Lynn Hannan, Joanna Ho, Steve Kachelmeier, Theresa Libby, Joan Luft,
Bill Tayler, Tom Vance, and workshop participants at Michigan State University, Miami
University, Monash University, and University of California-Irvine for providing helpful
comments on our paper. Funding for this research was provided by the Department of
Accounting and Information Systems at Michigan State University.
Budgeting, Psychological contracts,
and Budgetary Slack
Abstract: We study three types of budgeting (authoritative, consultative, and participative) that
differ in the influence that managers have on their approved budgets and the opportunities they
provide for managers to build budgetary slack into their budget requests. We provide evidence
that when organizations implicitly communicate that budgeting will be participative, it
establishes psychological contracts in managers. If managers subsequently experience budgeting
that is inconsistent with their psychological contracts, then their psychological contracts are
breached. This leads to feelings of violation and distrust, even when the terms of the managers’
economic contracts are fulfilled. We examine if managers whose psychological contracts are
breached seek redress by building additional budgetary slack into their budget requests.
Experimental results indicate that authoritative and consultative budgeting result in
psychological contract breach to a greater extent than participative budgeting. To seek redress,
managers build in more budgetary slack when budgeting is authoritative or consultative than
when it is participative. Furthermore, the effects of the extent of psychological contract breach
on built-in budgetary slack persist in the future when budgeting is participative.
Keywords: budgeting; budgetary slack; participative budgeting; psychological contracts.
1. Introduction
Organizations routinely use budgeting as an important basis for communicating,
planning, motivating, and evaluating performance. Three types of budgeting that are commonly
implemented in organizations (authoritative, consultative, and participative) vary in the level of
managers’ budgetary participation. Budgetary participation refers to a process in which managers
are involved with and have influence on their approved budgets (Shields and Shields 1998).
Research finds that budgetary participation can improve organizational outcomes (Baiman and
Evans 1983; Antle and Fellingham 1995; Covaleski et al. 2003). The benefits of budgetary
participation are especially salient in the presence of information asymmetry, i.e., when
managers have better information than their superiors about costs, revenues, or cash flows.
Considerable research in accounting investigates economic and psychological mechanisms by
which effects of budgetary participation occur (Shields and Young 1993; Shields and Shields
1998; Libby 1999; Covaleski et al. 2003).
While budgetary participation can benefit organizations, it can also be the basis of
psychological contracts between managers and their organizations. Psychological contracts arise
from managers’ beliefs about the obligations and responsibilities of parties in their exchange
relationships (Rousseau 1995) and can exist either in the presence or absence of economic
contracts.1 We investigate whether and how the three commonly implemented types of budgeting
(authoritative, consultative, and participative) influence managers’ evaluations of their
psychological contracts of participation.
1
Throughout this study, we refer to contracts that are designed using an agency theory perspective as economic
contracts. Economic contracts are also referred to as employment contracts (Baiman 1990) and compensation
contracts (Lambert 2001, 2007). Economic contracts have three distinguishing features (Baiman and Rajan 2002).
First, all information is jointly observable by contracting parties. Second, such information is verifiable by a court.
Third, they are costless to write and enforce.
1 The aforementioned three types of budgeting vary in two important ways. First, they
vary with respect to managers’ participation because they provide managers’ with different
levels of influence on their approved budgets. Second, they vary with respect to managers’
economic incentives for building budgetary slack. Under participative budgeting, managers have
the highest influence on their approved budgets and therefore the strongest economic incentives
to build budgetary slack into their budget requests. Under authoritative budgeting, managers
have the lowest influence on their approved budgets and therefore the weakest economic
incentives to build budgetary slack. Under consultative budgeting, managers have some
influence on their approved budgets and therefore moderate economic incentives to build
budgetary slack.
Regardless of the type of budgeting that is implemented, organizations often
communicate to managers that they will participate in budgeting. When organizations
communicate that managers will participate in budgeting, we argue that managers will expect
that their involvement in budgeting will influence their approved budgets. We expect that this
will establish psychological contracts of participation in the managers.
When managers have less influence on their approved budgets than they expect,
budgeting is pseudo-participative (Argyris 1952, 1953). If managers’ experience budgeting that
does not meet their expectations of influence, then we expect that they will conclude that their
psychological contracts of participation are breached. Therefore, we first predict that when
organizations communicate that managers will participate in budgeting and then implement
authoritative or consultative budgeting, it will cause managers to believe that budgeting is
pseudo-participative. This will lead managers to conclude that their psychological contracts of
participation are breached.
2 Next, we investigate the effects of the extent of psychological contract breach on the
budgetary slack that managers build into their budget requests. Unlike economic contracts,
psychological contracts do not specify the consequences of breach of contract. Managers whose
psychological contracts are breached are likely to recalibrate their beliefs about their exchange
relationships with their superiors and conclude that these relationships are transactional and
governed by self-interest by both parties. As a result, we expect that managers under
authoritative and consultative budgeting will build more budgetary slack into their budget
requests than will managers under participative budgeting. We also expect that managers’
feelings of violation that occur as a result of psychological contract breach under authoritative
and consultative budgeting will persist in the future even when budgeting is participative.
We use a laboratory experiment in which budgeting is operationalized at three levels that
vary in the influence that managers have on their approved budgets. As a consequence of varying
degrees of managerial influence, each type of budgeting also varies in the managers’ economic
incentives to build in budgetary slack.2 Consistent with our predictions, the experimental results
indicate that participants under authoritative and consultative budgeting believe that their
psychological contracts are breached to a greater extent compared to participants under
participative budgeting. As a result, participants under authoritative and consultative build more
budgetary slack into their budget requests and believe that their superiors are less trustworthy.
Consistent with the theory of psychological contracts, we also find that participants under
authoritative and consultative budgeting experience feelings of violation and distrust that are
2
We manipulate whether or not subordinates’ psychological contracts are breached rather than the presence or
absence of psychological contracts because it is rare to find an organization in which employees do not have
psychological contracts (Rousseau 1995; Morrison and Robinson 1997). Psychological contracts are pervasive in
organizations because employees work interdependently over time within a social context and, as a consequence,
form beliefs about their social exchanges which become the basis of psychological contracts.
3 predicted to occur after psychological contract breach. These results support our prediction that
different types of budgeting influence whether or not subordinates’ psychological contracts of
participation are breached and consequently the budgetary slack that they build into their budget
requests.
Results also indicate that the budgetary slack that participants build during budgeting is
in the opposite direction of their economic incentives to build in budgetary slack under each type
of budgeting. Although participants have the strongest economic incentives to build budgetary
slack under participative budgeting, we find that built-in budgetary slack is significantly lower
under participative than authoritative or consultative budgeting under which participants have
weaker economic incentives to build in budgetary slack. Finally, we find that the effects of
psychological contract breach on built-in budgetary slack that occur under authoritative and
consultative budgeting persist in the future when budgeting is participative.
Our paper makes several contributions to accounting research and practice. First, we
provide evidence that budgeting can establish psychological contracts of participation in
managers. When managers’ experience budgeting that does not meet their expectations about
participation, their psychological contracts are breached and this has implications for their slackbuilding behavior. Although prior research provides evidence that budgetary slack is driven by
economic motives (Baiman and Lewis 1989) and psychological motives such as preferences for
reciprocity, honesty, and distributional equity (Evans et al. 2001; Kuang and Moser 2009), we
provide evidence that budgetary slack is also driven by the extent to which managers experience
psychological contract breach during budgeting. Psychological contracts are the result of an
underlying mechanism that is theoretically distinct from other psychological motives for building
budgetary slack (e.g., preferences for reciprocity, honesty, and distributional equity). We provide
4 evidence that psychological contracts have powerful effects on managerial and organizational
outcomes.
Second, we find that consultative budgeting is likely to lead to poor financial outcomes.
Although managers have some influence on their approved budgets under consultative
budgeting, we find that they do not believe that their influence meets their expectations. As a
result, they believe that their psychological contracts of participation are breached and they
respond by building budgetary slack, which leads to lower firm profit than under participative
budgeting. Interestingly, although our results indicate that authoritative and consultative
budgeting each lead managers to believe that their psychological contracts are breached to a
similar extent, consultative budgeting has lower firm profit.
Third, accounting research provides evidence that pseudo-participative budgeting is
associated with negative psychological outcomes such as decreased job satisfaction,
organizational commitment, motivation, and performance, as well as increased stress and
interpersonal conflict (Argyris 1952; Libby 1999; Shields and Shields 1998). We contribute to
this literature in two ways. First, although some accounting research considers authoritative and
consultative budgeting to be participative (Baiman and Evans 1983; Antle and Fellingham 1995),
we provide evidence that managers consider them to be pseudo-participative. Second, we
provide evidence of additional negative effects of pseudo-participative budgeting. Specifically,
we provide evidence that pseudo-participative budgeting is a breach of managers’ psychological
contracts of participation and, as a result, leads to higher built-in budgetary slack and lower firm
profit. Finally, we provide evidence that the negative economic and psychological and
consequences of psychological contract breach that occur under authoritative and consultative
budgeting persist in the future when budgeting is participative.
5 The theory and hypotheses are developed in the following section. The experiment is
described next, followed by the results and conclusions.
2. Theory and hypothesis development
Budgeting
Participation in budgeting is valuable because it allows superiors to elicit managers’
private information. At the same time, participative budgeting also provides managers with
economic incentives to build in budgetary slack during budgeting to increase their resource
allocations. As a result, organizations often implement budgeting that limits managers’ influence
on their approved budgets to constrain built-in budgetary slack.
We investigate three types of budgeting that lie on a continuum; they vary in managers’
influence on their approved budgets3 and in managers’ economic incentives to build budgetary
slack into their budget requests (Cotton et al. 1988; Locke and Schweiger 1979; Wagner 1994;
Wagner et al. 1997). These three types of budgeting also have psychological (cognitive and
motivational) effects on managers (Locke and Schweiger 1979; Wagner et al. 1997; Shields and
Shields 1998). At one end of the continuum is participative budgeting, which is implemented in
order to facilitate the bilateral flow of information. Under participative budgeting, managers
submit budget requests that are considered by their superiors when their approved budgets are
decided. If managers’ approved budgets are different from their budget requests, then superiors
provide explanations for these differences (Libby 1999). Managers’ budget requests under
participative budgeting have more influence on their approved budgets than under the other two
For example, a recent survey of information technology managers reported that “[j]ust 22 percent of respondents to
our survey say they have a major influence on the IT budget. Another 35 percent say they have some influence …
[b]ut 43 percent say they have little or no influence on budget allocations” (Feldman, 2012, 35).
3
6 types of budgeting that we investigate and, consequently, managers have the strongest economic
incentives to build in budgetary slack.
At the other end of the continuum is authoritative budgeting,4 which is implemented in
order to facilitate unilateral information flows from managers to their superiors (Baiman and
Evans 1983; Antle and Fellingham 1995). That is, the primary purpose of authoritative budgeting
is for superiors to elicit information from managers in the form of budget requests although the
managers’ approved budgets are often already decided or pre-determined by the superiors. Under
authoritative budgeting, approved budgets can be markedly different from budget requests
because a superiors’ goal is to minimize slack in budgets. Managers’ budget requests under
authoritative budgeting have the lowest influence on their approved budgets relative to the other
two types of budgeting that we investigate. As a result, managers have the weakest economic
incentives to build in budgetary slack under authoritative budgeting.
One example of authoritative budgeting is a hurdle contract (Antle and Eppen 1985).
Although a hurdle contract has elements of participative budgeting because it uses information
from managers to establish the hurdle (Baiman and Evans 1983; Antle and Fellingham 1995),
managers are likely to consider it pseudo-participative because budget requests, on average, have
little influence on approved budgets. The approved budget in a hurdle contract is always equal to
the economically optimal hurdle cost as long as the managers’ budget requests are less than the
hurdle cost and fall within an ex ante agreed upon distribution of costs.
4
Authoritative budgeting is prevalent in practice. A report by McClenahen (1995, 13), a principal consultant at
PwC, comments on the prevalence of bottom-up budget creation, i.e., “the anachronistic and expensive practice of
repeatedly passing numbers up through successive layers only to have them come to rest with some senior
executives who had pretty much pre-determined what figures would be acceptable.” Also see footnote 3 in which
almost half of the managers in a survey report that they have little or no influence on their approved budgets.
7 Between the ends of the continuum is consultative budgeting. The purpose of
consultative budgeting is similar to participative budgeting in that it is designed to facilitate the
bilateral flow of information. However, managers under consultative typically have less
influence on their approved budgets and do not always receive explanations from their superiors
for their approved budgets. Instead, information flows upwards from managers to their superiors,
who use that information along with other information (some of which may be unknown to the
managers, such as the resources available and alternate uses of the resources) to approve budget
requests, which are then communicated downward to managers. Approved budgets under
consultative budgeting are frequently lower than managers’ budget requests.5 As a result,
managers have moderate economic incentives to build in budgetary slack under consultative
budgeting because their budget requests have some influence on their approved budgets.
Psychological Contracts and Breach
Psychological contracts represent managers’ beliefs about their exchange relationships
with their organizations (Blau 1964; Rousseau 1995).6 Psychological contracts are established
when managers believe that mutual obligations exist between them and their organizations
(represented by their superiors) and that the other party has either explicitly or implicitly
promised to fulfill those mutual obligations (Rousseau 1995). Various factors such as written or
For example, a majority of corporate respondents in a recent survey of information technology managers reported
that budgeting was the same old exercise in containing costs and ended with approved budgets that were
substantially lower than the managers’ budget requests, without much explanation from their superiors (Feldman
2012). Similar instances of consultative budgeting are found in the public sector. For example, in 2011 (2012),
President Obama requested a Federal budget of $3.83 ($3.73) trillion in outlays and was approved a budget of $3.63
($3.53) trillion. For 2013, President Obama had a series of negotiations with senators to try to avoid a large decrease
to the approved budget relative to the budget request (New York Times, March 6, 2013).
6
In the psychological contract literature, the term “contract” is used as a construct to refer to an individual’s beliefs
about the relationship between the organization and the individual. Although these beliefs are not formally
documented in a legal contract, they are like a contract to the individual. The individual expects the psychological
contract to be enforced during the exchange process (Rousseau 1995).
5
8 oral communications, organizational policies or procedures, or interpersonal interactions can lead
managers to believe that they have psychological contracts. Psychological contracts share some
common features with gift-exchange or reciprocity-based contracts, but they are theoretically
different.7
Psychological contracts are breached when managers evaluate their psychological
contracts and conclude that the other party “has failed to meet one or more obligations within
one’s psychological contract in a manner commensurate with one’s contributions” (Morrison and
Robinson 1997, 230). Breach leads to feelings of violation, defined as “emotional distress and
feelings of anger and betrayal arising from the realization that one’s organization has not fulfilled
highly salient promises” (Dulac et al. 2008, 1080). Through these evaluations, managers form
beliefs about their exchange relationships, which influence their subsequent behavior (Rousseau
1995, 2001). When managers experience psychological contract breach, they re-evaluate their
exchange relationships and conclude that they are transactional (i.e., based entirely on economic
currency) rather than relational (i.e., based on the socioeconomic currency of trust and loyalty)
(Thompson and Bunderson 2003).
Managers can respond to psychological contract breach in a number of ways, including
termination of their exchange relationships, reduced effort, sabotage, or dishonesty. Empirical
research provides evidence that psychological contract breach leads to decreased performance,
7
Gift-exchange contracts pay higher wages than those predicted by agency theory in order to motivate managers to
reciprocate the gift of a higher-than-market wage with effort that is higher than the minimum effort required (Fehr
and Gachter 2000; Hannan et al. 2002; Kuang and Moser 2009). Managers under gift-exchange contracts do not
believe that they are obligated to provide high effort, but rather provide it as a gift to their firms because they prefer
reciprocity. Reciprocity is theoretically different from psychological contracts because while managers may prefer
reciprocity, they do not necessarily believe that it has been promised to them. When managers have psychological
contracts, however, they unilaterally believe that an obligation of reciprocity exists and that reciprocity has been
promised to them. Further, psychological contracts are an individual-level construct, whereas gift-exchange and
reciprocity-based contracts are organizational-level constructs (Rousseau 1989).
9 lower organizational commitment, and increased turnover (Morrison and Robinson 1997; Dabos
and Rousseau 2004).
Budgeting and Psychological Contracts
Research in accounting provides evidence that participative budgeting has positive effects
on organizational economic outcomes (Baiman and Evans 1983; Penno 1984; Antle and Eppen
1985; Antle and Fellingham 1995; Shields and Young 1993; Baiman, Larcker, and Rajan 1995;
Baiman and Rajan 2002; Rajan and Reichelstein 2009). In contrast, organizational research
warns that budgeting can have negative psychological effects on employees depending on how it
is implemented (Argyris 1952, 1953). When organizations explicitly or implicitly communicate
that managers will participate in budgeting, they expect that their participation will influence
their approved budgets. The managers’ expectation establishes their psychological contracts of
participation. If they subsequently experience budgeting in which their influence is less than their
expectations of influence, then they believe that their psychological contracts are breached.
We provide evidence that different types of budgeting lead to differences in the extent to
which managers believe that their psychological contracts are breached and that this has
implications for built-in budgetary slack. Consider a situation in which managers accept an
economic contract offered by an organization that either explicitly or implicitly communicates
that the managers will participate in budgeting. Although the organization provides no detail
regarding the influence that the managers will have, we predict that the mere use of the word
“participation” will lead them to expect that their involvement in budgeting will influence their
approved budgets. We predict that managers will consider authoritative and consultative
budgeting to be pseudo-participative and, as a result, they will believe that their psychological
10 contracts are breached to a greater extent under authoritative and consultative than participative
budgeting.
HYPOTHESIS 1. Compared to participative budgeting, managers under authoritative or
consultative budgeting will believe that their psychological contracts of
participation are breached to a greater extent.
Psychological Contract Breach and Built-In Budgetary Slack
Psychological contract breach leads managers to engage in an active process of cognitive
sense-making, whereby they seek to understand why their psychological contracts were breached
(Rousseau 1995). If they attribute psychological contract breach to factors under their superiors’
control, then the negative feelings associated with this breach are heightened. Specifically,
managers who experience psychological contract breach believe that their exchange relationships
are fundamentally undermined because their superiors have violated their trust (Rupp and
Cropanzano 2002).
We investigate whether and to what extent breach of managers’ psychological contracts
of participation affects the slack that they build in to their budget requests when the terms of
their economic contracts are fulfilled. Unlike breach of an economic contract, managers facing
breach of a psychological contract cannot seek redress from the legal system. Managers who
experience psychological contract breach, however, still require some type of redress to bridge
the discrepancy between the outcomes they believe they were promised and the outcomes they
experience. We argue that budgetary slack is one redress available to managers. By building
slack into their budget requests, managers can (depending on the influence that their budget
requests have on their approved budgets) obtain larger approved budgets and resource
allocations. These larger budgets and resource allocations can cause superiors to experience
11 economic losses: for example, if superiors approve budgets that include slack, then they have
fewer resources available for other budgets; furthermore organizational profits are likely to be
lower if approved budgets contain slack.
To the extent that higher budgetary slack leads to higher approved budgets, managers can
use the slack as a way to seek redress for their superiors’ failure to fulfill their mutual obligations
and promises of participation to the managers. Therefore, based on the theory of psychological
contracts, we predict that built-in budgetary slack will be higher under authoritative and
consultative compared to participative budgeting, despite stronger economic incentives to build
in budgetary slack under participative budgeting.
HYPOTHESIS 2. Built-in budgetary slack will be higher under authoritative and
consultative than participative budgeting.
Psychological Contracts and Future Budgetary Slack
Finally, consistent with the theory of psychological contracts, we argue that the effects of
psychological contract breach that occur when budgeting is authoritative or consultative will
persist in the future when budgeting is participative. Managers who experience psychological
contract breach are likely to believe that their superiors have abused their trust, which influences
their current behavior. In addition, psychological contract breach can also influence managers’
future behavior in their exchange relationships because their current evaluations of their
psychological contracts allow them to form expectations about their future interactions with their
superiors (Rousseau 1995; Morrison and Robinson 1997). If managers believe that their
superiors have violated their trust, then we argue that they will be more likely to believe that
their superiors will violate their trust in the future. If managers currently under authoritative or
consultative budgeting believe that their psychological contracts are breached to a greater extent
12 than managers under participative budgeting, then we expect that their responses to
psychological contract breach (i.e., higher built-in budgetary slack) will persist in the future
when budgeting is participative.
HYPOTHESIS 3. Built-in budgetary slack under participative budgeting will be higher
when managers have previously experienced authoritative or consultative
budgeting.
3. Experimental method
Participants
Eighty-two full-time MBA students enrolled in an accounting course at a large public
university participated in the experiment. The experiment was conducted in two sessions on the
same day. Participants in each session worked independently using their own personal
computers. Statistical analysis of built-in slack does not indicate that there is a session-specific
effect (p = .82, two-tailed), therefore, we do not include session as a variable in our statistical
analyses.
Task Description and Procedure
The experimental task was adapted from Evans et al. (2001). We used a budgeting setting
in which participants assumed the role of managers with private production cost information that
they used to submit budget requests each period to their hypothetical (i.e., mechanical) superiors,
who decided the participants’ approved budgets.8 The use of hypothetical superiors allowed us
to determine in advance the superiors’ response to participants’ budget requests. This eliminated
variance in the operationalization of the three budgeting conditions that would have occurred if
participants had assumed the role of the superiors.
8
Participants were not aware that their superiors were hypothetical. 13 Participants first read instructions that explained their role in the experiment, the
experimental setting, and their compensation plan. Participants were informed that they were
managers of a production department in a division of a firm. Managers at their firm participated
in budgeting by submitting budget requests to their superiors who participants were told would
consider the requests when the superiors decided the approved budgets. Participants were also
told that their approved budgets could be greater than, less than, or equal to their budget requests.
Participants were also informed that, based on their firms’ forecasts, their actual production cost
would range from $2.00 and $8.00 in each period and that any cost within this range was equally
likely to occur. Participants were told that actual costs would never be below $2 or above $8 and
therefore if their budget requests were outside the range of $2 to $8, then they would get an error
message and receive a budget allocation of $0.
Participants were then informed that subsequent to accepting their economic contracts,
they had access to a private production-cost forecasting system, which provided information
about their department’s actual production costs with 100% accuracy. Participants were
informed that no one in their organization could ever obtain access to their actual cost
information or learn whether or not their budget requests equaled their actual costs in any period.
After participants observed their actual production cost in each period, they submitted a budget
request for that period to their superiors that would indicate the amount of money that they
wanted to receive in order to fund their production for the next period.
After reading the instructions, participants completed a pre-experiment quiz on which
they were required to score 100% to ensure that they understood their role, task, compensation
plan, and budgeting in their firm. The pre-experiment quiz also confirmed that participants
understood that the production cost information generated by their private forecasting system
14 could never be accessed by anyone in their firm besides themselves. After they successfully
completed the quiz, the first five experimental periods commenced. Participants were randomly
assigned to one of the three budgeting conditions in the first five periods. After the first five
periods, participants were informed that they had a different superior due to a personnel change
in their organization. Participants were then randomly assigned to one of the other two budgeting
conditions. In the last five periods, participants budgeted with their new superiors. In order to
control for end-game effects, participants were not aware that the experiment had ten periods or
that the type of budgeting or their superiors would change during the experiment.
At the beginning of each period, participants privately learned their department’s actual
production cost for that period with certainty. Next, they sent their budget request for that period
to their superiors. They could choose budget requests that were equal or unequal to their actual
production costs. After they sent their budget requests to their superiors, their superiors informed
them of their approved budgets for that period.
After participants completed the ten experimental periods, they responded to postexperiment questions related to demographic information, factors that influenced their budget
requests, and the manipulation of the budgeting variable.
Design and Treatment Conditions
The experiment had three independent variables: budgeting, round, and period.
Budgeting was operationalized at three levels9, as described below.
Participative Budgeting: Under participative budgeting (hereafter, P), participants’
approved budgets equaled their budget requests each period when their budget requests were
9
Involvement in budgeting was held constant across the three budgeting conditions.
15 within the range of the ex ante determined cost distribution of $2-$8. Our operationalization of P
is consistent with the trust contract in Evans et al. (2001).
Authoritative budgeting: Under authoritative budgeting (hereafter, A), participants’
approved budgets always equaled the hurdle cost ($5) when their budget request were less than
or equal to the hurdle cost.10 If participants’ budget requests were greater than the hurdle cost,
then their approved budget was for $0. In this situation, participants received budgetary slack,
which equaled the difference between the hurdle cost and the actual cost, regardless of the
participants’ budget request and whether or not they were interested in obtaining slack. This is
consistent with the hurdle contract in Antle and Eppen (1985) and Evans et al. (2001), with one
important exception. Although the hurdle cost used to determine participants’ approved budgets
under A was economically optimal, participants were not aware of the hurdle cost ex ante.
Instead, they could learn about it as they gained experience with their superiors during
budgeting.
Consultative budgeting: Under consultative budgeting (hereafter, C), participants’
approved budgets equaled their budget requests minus a randomly determined number that was
uniformly distributed between $0.80 and $1.20 as long as their budget requests were within the
range of the ex ante determined cost distribution of $2-$8.11
10
If actual production cost is uniformly distributed between $2 and $8, then the optimal hurdle cost (H) in the A
condition is $5. This is because the firm chooses the hurdle cost per unit that maximizes its objective function:
$
$
$8
0,
$
$
subject to the constraint that $2.00  H  $8.00. Q is the quantity of production. The probability that production cost
$
is  H is
, and the corresponding firm profit per unit is ($8 - H). The probability that production cost is > H
$
$
, and the corresponding firm profit per unit is $0. The derivative of with respect to H is ($10 - $2H).
is
$
Setting this to zero for the first-order condition for maximization provides an optimal hurdle cost per unit (H) of $5.
11
We do not include a condition in which subordinates’ approved budgets are higher than their budget requests
because we are not aware of real-world settings in which superiors approve budgets that are higher than
subordinates’ budget requests.
16 Round was manipulated at two levels and there were five experimental periods in each
round. After completing the first round in one randomly assigned budgeting condition,
participants were randomly assigned to a different budgeting condition for the second round. For
example, participants who were randomly assigned to P in the first five periods were randomly
assigned to either A or C in the second five periods, which enabled testing H3.
Compensation
Participants were compensated using a budget-based compensation plan, which provided
them with economic incentives to overstate their actual costs of production (Evans et al. 2001).
Participants’ compensation was a function of their salary, actual production costs, approved
budgets, and donations to charity.12 Participants’ financial compensation for each period was
calculated as:
C=S-D
if A ≥ B
C = S + (B – A) – D
if A < B
(1)
where:
C = compensation;
S = salary;
D = donation;
A = actual cost; and
B = approved budget.
For participants who did not built-in budgetary slack, this compensation plan resulted in
expected compensation of $1.00 per period in each budgeting condition. Participants’ total
compensation for the experiment was the sum of their compensation in each of the ten
12
Participants had the option to anonymously donate to charity some or all of their compensation (e.g., budgetary
slack). Donations to charity are explained in detail in the next section. 17 experimental periods. Participants were paid approximately one week following the experiment
and their average earnings were $17.09 for approximately 50 minutes of participation.
Donations to charity: We included an option to donate to charity for three reasons. First,
in under C, participants’ decisions to build in budgetary slack could be driven by a need for
redress for their superiors’ breach of their psychological contracts of participation. That is,
participants’ decisions to build in budgetary slack could be driven by psychological motives
rather than economic motives. Second, higher built-in slack under C could simply be driven by
the desire not to have production suspended, which in our setting occurs when approved budgets
are sufficient to fund production. Therefore, participants under C could have submitted budget
requests that were higher than their actual costs to ensure that production occurred even if their
approved budgets were less than requested. If participants under C decided to build in slack only
to fund production rather than to garner slack, then it would be possible that they would receive
slack that they did not wish to consume. In this situation, they were provided the option to donate
some or all of their slack to charity.
Third, under A, it was possible for slack to be imposed on participants even when they
did not want it (i.e., when they did not submit budget requests with budgetary slack). This could
occur when participants’ approved budgets (i.e., the hurdle cost) were higher than their actual
production costs. We therefore included an option to donate to charity in order to allow
participants under A to donate when slack was imposed on them if they did not want to have
slack. The option of donating to charity allows us to disentangle economic incentives for
budgetary slack from three alternate reasons for building slack, i.e., to psychologically redress
for their superiors’ breach of their psychological contracts without benefitting monetarily from
the redress, to fund production (under C), and to give away the imposed slack (under A).
18 Dependent Variables
There are two dependent variables: extent of breach of psychological contract and built-in
budgetary slack. The extent of breach of psychological contract was measured by participants’
answers to four questions that are designed consistent with the theory of psychological contracts
(Rousseau 1989) and prior empirical research in psychology on psychological contract breach
(Robinson and Rousseau 1994). These questions capture the extent to which participants believe
that their superiors: (a) fulfill their obligations, (b) keep their promises, (c) are honest, and (d) are
trustworthy. Fulfillment of obligations was measured on a 7-point scale: “To what extent did
your division manager fulfill his or her obligations to you?”, with scale anchors of 1=not at all
fulfilled and 7=completely fulfilled. Keep promises was measured on a 7-point scale: “To what
extent did your division manager keep his or her promises to you?”, with scale anchors of 1=not
at all and 7=completely. Superior’s honesty was measured on a 7-point scale: “To what extent do
you believe that your division manager was honest?”, with scale anchors of 1= completely
honest and 7=completely dishonest. Finally, superior’s trustworthiness was measured on a 7point scale: “To what extent did you trust your division manager?”, with scale anchors of 1=
completely trust and 7=completely distrust.
The dependent variable, built-in budgetary slack, is measured as the percentage of the
maximum amount of available budgetary slack that participants include in their budget requests.
Budgetary slack is a continuous variable that ranges from 0% to 100%. Specifically, it is
measured as %BUILT_IN_SLACK = built-in budgetary slack
maximum available budgetary
slack. For example, consider participants under P whose actual cost of production was $5.00.
Their budgetary slack would be maximized when their budget request equaled $8.00 (i.e., the
highest possible production cost) because they could keep the difference between the actual
19 production cost and their approved budget. Thus, the maximum amount of available budgetary
slack in this example is $3.00. If participants submit budget requests of $8.00, then the budgetary
slack they built in is $3.00 (i.e., $8.00 budget request – $5.00 actual cost = $3.00) and
%BUILT_IN_SLACK = $3.00
$3.00 = 100%. Alternatively, if participants submitted budget
requests of $5.00, then the budgetary slack they built in is $0 (i.e., $5.00 budget request – $5.00
actual cost = $0) and %BUILT_IN_SLACK = ($0 $5.00) = 0%.
4. Results
Participants mean work experience is 45 months. Thirty-nine (48 percent) participants
have work experience that includes budgeting as a subordinate manager and 21 (26 percent)
participants have work experience that includes budgeting as a superior manager. There are no
significant effects of work experience on the extent of breach of psychological contract or builtin budgetary slack (all p-values > .50, untabulated); therefore, we do not include these variables
in the hypotheses tests.13
Descriptive Statistics
Extent of psychological contract breach
For all four question in each round, the means are highest for P, lowest for A, and in
between P and A for C (Table 1, Panels A and B). Specifically, the extent to which superiors’
fulfill their obligations to the managers under P, A, and C are 6.04, 3.00, and 4.00, respectively,
in round 1; and 5.89, 2.93, and 3.33, respectively, in round 2. The mean of the extent to which
superiors’ keep their promises under P, A, and C are 6.11, 3.36, and 3.59, respectively, in round
13
28 participants were randomly assigned to condition P, and 27 each to C and A respectively. Participants who
submitted budget requests that were outside the range of actual costs were omitted from the data analysis and
hypothesis testing. Thus, the number of observations in various tests can differ from 82.
20 1; and 6.18, 2.81, and 3.52, respectively, in round 2. The mean of the extent to which superiors
are believed to be honest under P, A, and C are (reversed scaled) 2.56, 4.21, and 3.93,
respectively, in round 1; and 3.25, 4.30, 4.22, respectively, in round 2. Finally, the mean of the
extent to which superiors are trustworthy under P, A, and C are (reversed scaled) 2.33, 3.96, and
4.00, respectively, in round 1; and 3.18, 4.70, and 4.56, respectively, in round 2.
___________________
Insert Table 1
___________________
Built-In Budgetary Slack
Mean built-in slack in periods 1- 5 is lowest for P (23.40 percent) and highest for C
(45.97 percent); mean built-in slack for A is 35.45 percent (Appendix, Panel A). Participants’
budget requests contain no budgetary slack (i.e., %BUILT_IN_SLACK = 0%) more frequently
under P than under either A or C, regardless of experimental period.
Mean built-in slack in periods 6-10 has limited interpretive value without controlling for
the type of budgeting faced by participants in the first five periods. Recall that participants under
P in periods 6-10 are under A or C in periods 1-5. Therefore, participants under P in periods 6-10
have previously experienced a breach of their psychological contract. For completeness,
however, we report mean built-in slack in periods 6-10 in the Appendix (Panel B).
Test of Hypothesis 1
H1 predicts that participants under A or C compared to P will believe that their
psychological contracts of participation are breached to a greater extent. First, we separately test
H1 using each of the four measures of the dependent variable. The means of P compared to the
means of A or C are significantly (p’s < 0.01) higher for superiors’ fulfillment of their
obligations to the managers and for superiors keeping their promises to the managers. The last
21 two measures are reversed scaled. For superiors’ honesty, the means are lower (meaning higher
honesty) for P compared to A or C. The difference in the means is significant for questions
related to the round 1 supervisor (p’s < 0.01) and questions related to the round 2 supervisor (p’s
< 0.10). Finally, for superiors’ trustworthiness, the means for P compared to A or C are
significantly lower (p’s < 0.01) for each of the two round supervisors, which indicates that under
P the participants believe that their superiors are more trustworthy.
Next, we conduct an additional test of H1 based on a factor analysis of the four measures
of the dependent variable. For the first (second) round supervisor, our results reveal that these
four measures load on to a single factor with factor loadings of 0.85, 0.86, -0.76, and -0.88 (0.87,
0.85, -0.76, and -0.84), respectively, for each of the four questions in Panels A (B) of Table 1.
The single factor explains 71.2 percent (68.9 percent) of the variance. We construct an OLS
regression with the psychological contract factor as the dependent variable and budgeting type as
the two independent variables, which take the value of 1 when participants are assigned to a
budgeting condition and zero otherwise. A and C are included in the regression, while P is
omitted in order to prevent singularity. The results of the regression analysis are in Table 1
(Panels C and D). A and C have significant negative coefficients (p’s < .001), which indicates
that compared to P, participants under A and C are significantly more likely to believe that their
psychological contracts are breached to a greater extent.
Overall, these results provide support for H1 that the extent of breach of psychological
contract is larger for A or C compared to P. Participants under P believe that their superiors
fulfill their obligations, keep their promises, and are more honest and trustworthy compared to
participants under A or C. This evidence is not only consistent with H1, but also to our
knowledge is the first evidence in the accounting literature about whether budgeting establishes
22 psychological contracts of participation in managers and to what extent managers believe that
different types of budgeting breach their psychological contracts of participation.
Test of Hypothesis 2
H2 predicts that built-in budgetary slack will be lower under P than under A and C. We
test H2 using data from round 1, which provides evidence on participants’ responses to
psychological contract breach when it first occurs (i.e., when they have not previously
experienced psychological contract breach during the experiment). We use data for periods 2-5
to test H2. We exclude period 1 because when participants decide their budget request in period
1, they do not have budget feedback from their superiors and therefore cannot know whether or
to what extent their psychological contracts are breached.14 We combine data from A and C
because differences in participants’ beliefs about the extent to which their psychological
contracts are breached in these conditions are not statistically significant (p = .75). Importantly,
however, participants under A and C are significantly more likely to believe that their
psychological contracts are breached compared to participants under P (p = .003).
The test of H2 uses a repeated measures analysis of variance (ANOVA) with period as a
4-level within-subjects factor, budgeting as a 2-level between-subjects factor (A and C, P), and
built-in budgetary slack in periods 2 – 5 as the dependent variable. The mean built-in slack in
participants’ budget requests is lower under P (23.83 percent) than A and C (42.49 percent)
(Table 2, Panel A). The results of the ANOVA indicate a significant main effect of budgeting on
%BUILT_IN_SLACK (F = 6.37, p < 0.01) (Table 2, Panel B). These results support H2. Thus,
participants under authoritative and consultative relative to participative budgeting believe that
We do not exclude period 5 data from the test because participants are not aware of the number of periods in the
experiment or that they will experience a change in budgeting during the experiment until these events occur.
14
23 their psychological contracts of participation are breached to a greater extent and, as a result,
build in more budgetary slack.
Our results also indicate that there is neither a significant main effect of period on
%BUILT_IN_SLACK (F = 2.71, p > 0.10) nor a significant period-by-budgeting interaction (F =
0.66, p > 0.50), indicating that mean built-in slack does not significantly change over the
experimental periods in any of the three conditions. If, for example, there is a tendency for
participants to increase built-in slack over the experimental periods because of learning effects
(i.e., if participants learn over multiple periods that there are no repercussions for reporting
budgetary slack), then the period variable would be statistically significant. We also include
cumulative prior earnings as a covariate in the test of H2 and do not find that it has a significant
effect on built-in budgetary slack in subsequent periods (p > 0.60).
___________________
Insert Table 2
___________________
Test of Hypothesis 3
H3 predicts that built-in budgetary slack will be higher when managers have previously
experienced authoritative or consultative budgeting than when they have not. To test H3, we
compare mean built-in slack under P in round 1 and 2. Recall that all participants under P in
round 2 previously experienced a breach of their psychological contracts because they were
under either A or C in round 1. If the effects of psychological contract breach persist, then we
expect that mean built-in slack under P in round 2 will be higher than under P in round 1.
Our test of H3 uses a repeated measures ANOVA with period as a 4-level within-subjects
factor, budgeting as a 2-level between-subjects factor (P in round 1, P in round 2), and mean
built-in slack in periods 2-5 and 7-10 as the dependent variable. Recall from our test of H2 that it
24 is necessary to use data from periods 2-5 in order to test the effect of the extent to which
participants’ psychological contracts are breached on built-in budgetary slack because it is
necessary for participants to receive at least one round of budgetary feedback in order for them to
evaluate whether and to what extent their psychological contracts are breached. Using the same
logic, we use data for participants under P in periods 7 – 10 in the second round in our test of H3.
_________________
Table 3
________________
Mean built-in slack is higher when participants’ psychological contracts are previously
breached (36.77 percent) than when they are not (23.83 percent) (Table 3, Panel A). The results
of the ANOVA indicate a significant main effect of budgeting on mean %BUILT_IN_SLACK (F
= 9.60, p < 0.01) (Table 2, Panel B). These results support H3. Thus, when participants
experience a breach of their psychological contracts in round 1 of the experiment (i.e., when they
are under A or C in round 1), they continue to report budgetary slack in round 2, even though
they are under P in round 2 where their psychological contracts are breached to a lesser extent.
Results in Table 3 also indicate that there is neither a significant main effect of period on
%BUILT_IN_SLACK (F = 0.99, p = 0.32) nor a significant period-by-budgeting interaction (F =
2.17, p = 0.15). These results indicate that mean built-in slack does not change over these periods
between rounds one and two.
Supplemental Analyses
25 This section reports the results of three analyses: comparison of differences in built-in
budgetary slack under A and C, firm profit analysis under A, C, and P, and analysis of
participants’ donations to charity under A, C, and P.
Comparison of Authoritative and Consultative Budgeting
Recall that managers under A have no influence on their approved budgets while
managers under C have some influence on their approved budgets. In consequence, the ratio of
benefits promised to benefits received is likely to be higher under A than C and, as a result, the
theory of psychological contracts predicts that psychological contracts are breached to a greater
extent under A than C. If the extent to which psychological contracts are breached is positively
associated with built-in budgetary slack, then built-in budgetary slack should be higher under A
than under C.
We conjecture, however, that C relative to A will lead to higher built-in budgetary slack
because it provides managers with the opportunity for redress against their superiors
economically as well as psychologically. Managers under A can only seek redress
psychologically for their superiors’ breach of their psychological contracts. This is because, on
average, subordinates under C can benefit monetarily by building in budgetary slack because
they have some influence on their approved budgets, while subordinates under A cannot because
they have no influence on their approved budgets.
Managers’ decisions to build in budgetary slack under A therefore are likely to be
influenced by two factors. First, consistent with psychology theory, subordinates managers can
seek psychological redress by building in budgetary slack as an emotional response (release,
venting) to psychological contract breach, even though there are no economic implications for
either them or their superiors by doing so (Morrison and Robinson 1997; Dabos and Rousseau
26 2004). Second, consistent with economic theory, managers can minimize their cost of their effort
by reducing their involvement in budgeting and simply requesting their actual costs (i.e.,
submitting budget requests with no budgetary slack).
We compare built-in slack under A and C using data from the periods in which
participants are able to distinguish between whether they are under A or C. We exclude period 1
because, as previously stated, participants cannot know the type of budgeting they are under or
the extent to which their psychological contracts are breached in period 1 because they have not
yet received budgetary feedback. We exclude periods 2 and 3 because participants cannot know
whether they are under A (when their approved budget is always $5) or C (when their approved
budget depends on their budget request but is always an amount lower than their budget request).
We therefore assume that three periods of feedback are required before participants can know
whether they are under A or C and use data from the final two periods of round one (i.e., periods
4 and 5) for this analysis.
We use repeated measures ANOVA with period as a 2-level within-subjects factor,
budgeting as a 2-level between-subjects factor (A, C), and built-in slack in periods 4 and 5 as the
dependent variable. Mean built-in slack is higher under C (54.28 percent) than A (36.74 percent)
(Table 4, Panel A). The results of the ANOVA indicate a significant main effect of budgeting on
%BUILT_IN_SLACK (F = 3.49, p < 0.07) (Table 4, Panel B). Consistent with our conjecture,
these results provide evidence that participants’ built-in slack under C compared to A is higher.
Furthermore, we find neither a significant main effect of period on %BUILT_IN_SLACK (F =
1.19, p = 0.28) nor a significant period-by-budgeting interaction (F = 0.01, p = 0.95), indicating
that mean built-in slack does not change over these two periods in any of the conditions.15
15
We provide evidence that psychological contract breach is an underlying causal mechanism that drives differences
in built-in slack between A and C. Comparing built-in slack between C and P provides further evidence that higher
27 ___________________
Insert Table 4
___________________
Firm Profit under Authoritative and Participative Budgeting
We analyze whether the lower built-in slack under P translates into higher firm profit
relative to the type of budgeting that relies on the economically optimal hurdle cost (A), which is
associated with breach of managers’ psychological contracts and higher built-in budgetary slack.
Recall that the firm’s actual production cost in each period was distributed uniformly over the
range of $2 to $8. For the purpose of this analysis, we assume a selling price of $8. If managers’
budget requests are aimed at wealth maximization, then expected firm profit is $0 in each period
under P because the managers would request $8 in each period and the superiors would approve
a budget of $8 in each period. Managers’ expected slack under P in each period is $3.16 Under A,
expected firm profit is $1.50 and managers’ expected slack is $0.75 in each period.17 Although
built-in slack under C results from psychological contract breach (i.e., psychological motives) rather than economic
motives. Participants under C compared to P build in significantly higher slack (mean value of difference in built in
slack in rounds 3-5 between C and P = 28.98 percent, t = 18.95, p < 0.01, two-sided). If economic motives were the
only driver of built-in slack under C, then participants under P compared to C should have built in more slack
because it provides stronger economic incentives to build in budgetary slack.
16
Under P, expected cost is $5, which is the mean of a uniform distribution of [2, 8]. A wealth-maximizing manager
will always report $8. If the selling price is $8, then expected firm profit is $0, and the manager’s expected slack is
$3, computed as $8-$5.
17
Under A, if the budget request is less than or equal to $5 (the economic optimal hurdle), then managers are
approved a budget of $5 and production occurs. Expected firm profit is $3 (i.e., selling price of $8 minus the
approved budget of $5). The managers’ expected slack is $1.50 (i.e., the approved budget of $5 minus the expected
actual cost of $3.50, which is the mean of the uniform cost distribution with a range of $2 to $5.) If the budget
request is greater than $5, then the approved budget is $0, there is no production, expected firm profit is $0, and the
managers’ expected slack is $0. For a uniform distribution with a range of $2 to $8, there is a 50 percent probability
that actual cost will be greater than $5, and a 50 percent probability that actual cost will be less than $5 in any
period. Therefore, expected firm profit is ($3*50%) + ($0*50%) = $1.50, and the managers’ expected slack is
($1.50*50%) + ($0*50%) = $0.75.
28 not the focus of this analysis, note that under C, expected firm profit is $0.83 and managers’
expected slack is $2.08 in each period.18
Mean realized firm profit in periods 2 – 5 is $2.21 under P, $1.49 under A, and $1.05
under C. Mean firm profit under P is significantly higher than the mean profit under A (p <
0.01). Furthermore, mean profit under P is significantly higher than the economic prediction of
zero (p < 0.01). These results provide evidence that the type of budgeting that is based on the
economic contract that assumes that subordinates will report to maximize their monetary wealth
and disregards their psychological contracts of participation (e.g., A) does not perform as well in
terms of firm profit as the type of budgeting that takes into consideration subordinates’
psychological contracts (e.g., P).
Donations
Participants who have budgetary slack in any period have the option to anonymously
consume all or part of their slack as compensation or donate all or part of their slack to charity at
the end of the period when they. The option to donate to charity is provided for three reasons.
First, participants who believe that their superiors have breached their psychological contracts
may want to respond by building in slack into their budget request, but not want to benefit
18
Under C, superiors always approve budget that are randomly less than the managers’ budget request by a mean of
$1, uniformly distributed between $0.80 and $1.20. If actual cost is greater than the approved budget, then there is
no production. Assuming wealth-maximizing managers, their budget requests will always be $8. The mean
approved budget would be $7 (i.e., the request of $8 less the reduction with a mean of $1). One-sixth of the time the
actual cost will be greater than $7 and the approved budget will be insufficient for production to occur; five-sixths of
the time the actual cost will be less than $7 and the approved budget will be sufficient for production to occur. The
expected firm profit during the five-sixths of the time that production occurs is $1 (i.e., selling price of $8 less
budget allocation of $7) and the managers’ expected slack is $2.50 (i.e., the mean approved budget of $7 less $4.50,
which is the mean of a uniform cost distribution with the range of $2 to $7.) The expected firm profit is $0 for the
one-sixth of the time when there is no production and the expected slack of the managers is $0. Therefore, the firm’s
expected profit in this condition is ($1*5/6) + ($0*1/6) = $0.83 and the managers’ expected slack is ($2.50*5/6) +
($0*1/6) = $2.08. 29 monetarily by doing so. Second, built-in slack under C could simply be driven by participants’
desire to have sufficient funds for production. Third, participants under A could receive slack
even when they did not want it. In order to investigate participants’ donations to charity, we
analyze their donations as a percentage of available budgetary slack (DONATION) for all periods
in which participants have budgetary slack. The mean DONATION is 18.55 percent under P,
27.14 percent under C, and 40.71 percent under A. Pair-wise contrast analyses indicate that mean
DONATION is higher under A than P (p = .03). There is no statistical difference in donations to
charity under C and P (p > .30).
Budgetary slack built-in by participants under P cannot be attributed to breach of their
psychological contracts because budgetary slack under P is received entirely at the discretion of
participants. Under P, approved budgets are always equal to budget requests. Participants under
A, however, sometimes had slack even when they submitted budget requests with no slack. For
example, built-in slack would have been imposed on participants under A when their actual costs
and budget requests were less than or equal to $5 because their approved budgets always equaled
the hurdle cost of $5. Interestingly, however, participants under A donate more of their
budgetary slack to charity than participants under P, suggesting that some participants under A
prefer not to consume slack. This indicates that some individuals are unwilling to benefit from
slack, even when it is imposed on them.
5. Conclusion
Although participative budgeting can improve organizational outcomes (e.g., profits) in
the presence of ex post information asymmetries, it can also establish psychological contracts of
participation in managers, who expect their involvement in budgeting to influence budgeting
30 outcomes. We provide evidence that three types of budgeting commonly implemented in practice
have different effects on the extent to which managers’ believe that their psychological contracts
of participation have been breached and, as a result, the extent to which they build in budgetary
slack to their budget requests. Specifically, we provide evidence that when organizations
communicate to managers that they will participate in budgeting, the managers believe that their
superiors have promised them that their involvement will influence their approved budgets.
When these managers subsequently experience budgeting in which their influence is less than
their expected influence, they experience breach of their psychological contracts of participation,
even when the terms of their economic contracts with their organizations are fulfilled. We also
provide evidence that the extent to which managers believe that their psychological contracts
have been breached is positively associated with built-in budgetary slack. Finally, we provide
evidence that the effects of breach of managers’ psychological contracts of participation persist
in the future when the managers no longer believe that their psychological contracts are
breached.
Interestingly, our results also indicate that built-in budgetary slack is highest when
budgeting is consultative (C): in this case, managers have some influence but not as much
influence on their approved budgets as when budgeting is participative (P). Furthermore,
managers under C compared to P are significantly less likely to believe that their superiors’
obligations and promises to them are fulfilled. These results suggest that managers are likely to
misinterpret the meaning of C as meant by their organizations (i.e., they are likely to expect more
influence than their organizations intended). A practical implication of this result is that
organizations should give managers either high influence on their approved budgets like that
which occurs under P or no influence on their approved budgets which occurs under A. A
31 second implication of this result is that organizations should provide managers with accurate and
truthful information about the type of budgeting in these organizations. Ambiguity about the type
of budgeting provides managers with the opportunity for different interpretations of the influence
that they can expect to have on their approved budgets. When managers experience a type of
budgeting that is inconsistent with their expectations about their influence on their approved
budgets, there can be negative economic and psychological consequences for them, superiors,
and organizations.
Although we provide evidence that budgeting is the basis of psychological contracts that
influence managers’ slack-building behavior during budgeting, several questions remain
unanswered. First, do managers revise their psychological contracts in response to changes in
their economic contracts? If so, then under what conditions do they do so? If they do not revise
their psychological contracts in response to changes in their economic contracts, then what is the
effect of their failure to do so on their slack-building behavior and on outcomes such as firm
profits?
Economic and psychological contracts influence and are influenced by accounting and by
the preferences, beliefs, and behaviors of superiors and managers in exchange relationships.
When organizations enter into economic contracts with their managers, it is important for them
to consider whether and how these contracts also create psychological contracts in their
managers because managers’ psychological contracts influence their budgeting-related behavior
and affect individual and organizational outcomes.
32 References
Antle, R., and G. D. Eppen. 1985. Capital rationing and organizational slack in capital
Budgeting. Management Science 31: 163-174.
Antle, R., and J. Fellingham. 1995. Information rents and preferences among information
systems in a model of resource allocation. Journal of Accounting Research 33: 41-58.
Argyris, C. 1952. The Impact of Budgets on People. New York, NY: Controllership Foundation.
Argyris, C. 1953. Human problems with budgets. Harvard Business Review 31: 97-110.
Baiman, S. 1990. Agency research in accounting: a second look. Accounting, Organizations and
Society 15: 341-371.
Baiman, S., and J. H. Evans III. 1983. Pre-decision information and participative management
control systems. Journal of Accounting Research 22: 371–395.
Baiman, S., D. Larcker, and M. Rajan. 1995. Organizational design for business units. Journal of
Accounting Research 33: 205–229.
Baiman, S. and B. L. Lewis. 1989. An experiment testing the behavioral equivalence of
strategically equivalent employment contracts. Journal of Accounting Research 27: 1-20.
Baiman, S., and M. Rajan. 2002. Incentive issues in inter-firm relationships. Accounting,
Organizations and Society 27: 213-238.
Blau, P. M. 1964. Exchange and Power in Social Life. New York, NY: John Wiley & Sons,.
Cotton, J. L., D. A. Vollrath, K. L. Froggatt, M. L. Lengnick-Hall, and K. R. Jennings. 1988.
Employee participation: diverse forms and different outcomes. Academy of Management Review
13: 8-22.
Covaleski, M., J. H. Evans III, J. Luft, and M. D. Shields. 2003. Budgeting research: three
theoretical perspectives for selective integration. Journal of Management Accounting Research
15: 3-49.
Dabos, G., and D. Rousseau. 2004. Mutuality and reciprocity in the psychological contracts of
employees and employers. Journal of Applied Psychology 89: 52-72.
Dulac, T., J. Coyle-Shapiro, D. J. Henderson, and S. J. Wayne. 2008. Not all responses to breach
are the same: the interconnection of social exchange and psychological contract processes in
organizations. Academy of Management Journal, 51(6): 1079-1098.
Evans III, J. H., R .L. Hannan, R. Krishnan, and D. Moser. 2001. Honesty in managerial
reporting. The Accounting Review 76: 537–559.
33 Fehr, E., and Gachter, S. 2000. Fairness and retaliation: the economics of reciprocity. Journal of
Economic Perspectives 14: 159-181.
Feldman, J. 2012. Same old budget battle. InformationWeek. October 29: 35-40.
Hannan, R.L., J. H. Kagel, and D. V. Moser. 2002. Partial gift exchange in an experimental labor
market: impact of subject population differences, productivity differences, and effort requests on
behavior. Journal of Labor Economics 20(41): 923-951.
Hannan, R. L., F. W. Rankin, and K. L. Towry. 2010. Flattening the organization: the
effect of organizational reporting structure on budgeting effectiveness. Review of
Accounting Studies 15(3): 503-536.
Kuang, X., and D. V. Moser. 2009. Reciprocity and the effectiveness of optimal agency
contracts. The Accounting Review 84: 1671–1694.
Lambert, R. 2001. Contracting theory and accounting. Journal of Accounting and Economics 32:
3-87.
Lambert, R. 2007. Agency theory and management accounting. In Handbook of Management of
Accounting, Vol. 1, edited by C. Chapman, A. Hopwood, and M. D. Shields. Oxford: Elsevier
Press.
Libby, T. 1999. The influence of voice and explanation on performance in a participative
budgeting setting. Accounting, Organizations and Society 24: 125-137.
Locke, E. and D. Schweiger. 1979. Participation in decision-making: one more look. In
Research in Organizational Behavior, edited by B. M. Staw. Greenwich: Conn., JAI Press: 265339.
McClenahen, J. S. 1995. Generally accepted practice? Industry Week November 6: 35.
Morrisson, E., and S. Robinson. 1997. When employees feel betrayed: a model of how
psychological contract breach develops. The Academy of Management Review 22: 226-256.
New York Times. 2013. Jackie Calmes and Jonathan Weisman. Trying to revive talks, Obama
goes around G.O.P. leaders. March 6: A1.
Penno, M. 1984. Asymmetry of pre-decision information and managerial accounting. Journal of
Accounting Research 22: 177-191.
Rabin, M. 1993. Incorporating fairness into game theory and economics. American Economic
Review 83: 1281–1302.
Rajan, M., and S. Reichelstein. 2009. Objective versus subjective indicators of managerial
performance. The Accounting Review 84: 209-237.
34 Rousseau, D. M. 1989. Psychological and implied contracts in organizations. Employee
Responsibilities and Rights Journal 2: 121-139.
Rousseau, D. M. 1995. Psychological Contracts in Organizations: Understanding Written and
Unwritten Agreements. Newbury Park, CA: Sage.
Rousseau, D. M. 2001. Schema, promise and mutuality: the building blocks of the psychological
contract. Journal of Occupational and Organizational Psychology 74(4): 511–541.
Rupp, D., and R. Cropanzano. 2002. The mediating effects of social exchange relationships in
predicting workplace outcomes from multi-foci organizational justice. Organizational Behavior
and Human Decision Processes 89: 925-946.
Shields, J. F., and M. D. Shields. 1998. Antecedents of participative budgeting. Accounting,
Organizations and Society 23: 49-76.
Shields, M. D., and S. M. Young. 1993. Antecedents and consequences of participative
budgeting: evidence on the effects of asymmetrical information. Journal of Management
Accounting Research 5: 265-280.
Thompson, J. A., and J. S. Bunderson. 2003. Violations of principle: ideological currency in the
psychological contract. Academy of Management Review 28(4): 571-586
Wagner, J.A. 1994. Participation's effects on performance and satisfaction: a reconsideration of
research evidence. Academy of Management Review 19: 312-330.
Wagner, J., C. Leana, E. Locke, and D. Schweiger. 1997. Cognitive and motivational
frameworks in U.S. research on participation: a meta-analysis of primary effects. Journal of
Organizational Behavior 18: 49-65.
35 TABLE 1
Effect of Budgeting on Extent of Psychological Contract Breach
Panel A: Mean (standard deviation) response to post-experimental questions about the firstround superior for each type of budgeting (n=82)
To what extent did your Division Manager
fulfill his or her obligations to you?
(1=Not at all,7=Completely)
To what extent did your Division Manager
keep his or her promises to you? (1=Not at
all, 7=Completely)
To what extent do you believe that your
Division Manager was honest?
(1=Completely Honest, 7=Completely
Dishonest)
To what extent did you trust your Division
Manager? (1=Completely Trust,
7=Completely Distrust)
Authoritative
(A)
3.00
(1.56)
Consultative
(C)
4.00
(1.69)
Participative
(P)
6.04
(1.25)
3.36
(1.62)
3.59
(1.89)
6.11
(1.34)
4.21
(1.32)
3.93
(1.49)
2.56
(1.72)
3.61
(1.69)
4.00
(1.44)
2.33
(1.57)
Panel B: Mean (standard deviation) response to post-experimental questions about the secondround superior for each type of budgeting (n=82)
To what extent did your Division Manager fulfill
his or her obligations to you? (1=Not at
all,7=Completely)
To what extent did your Division Manager keep
his or her promises to you? (1=Not at all,
7=Completely)
To what extent do you believe that your Division
Manager was honest? (1=Completely Honest,
7=Completely Dishonest)
To what extent did you trust your Division
Manager? (1=Completely Trust, 7=Completely
Distrust)
Authoritative Consultative Participative
(A)
(C)
(P)
2.93
4.56
5.89
(1.64)
(2.03)
(1.17)
2.81
(1.55)
3.52
(1.81)
6.18
(1.12)
4.30
(1.46)
4.22
(1.45)
3.25
(1.88)
4.70
(1.62)
4.56
(1.55)
3.18
(1.66)
36 Panel C: Regression analysis of psychological contract breach as a function of type of budgeting
for the first-round superior (n=82)
Authoritative (A)
Consultative (C)
Constant
Adjusted-R2
F-Value of Regression
Coefficient
(Std. Error)
-1.46
(0.20)
-1.23
(0.21)
0.91
(0.15)
t
p (two-tailed)
-7.01
< 0.001
-5.86
< 0.001
6.08
0.42
28.152 (p < 0.001)
< 0.001
Panel D: Regression analysis of psychological contract breach as a function of type of budgeting
for the second-round superior (n=82)
Authoritative (A)
Consultative (C)
Constant
Adjusted-R2
F-Value of Regression
Coefficient
(Std. Error)
-1.41
(0.21)
-1.20
(0.20)
0.86
(0.15)
t
p (two-tailed)
-6.63
< 0.001
-5.65
< 0.001
5.77
0.38
25.77 (p < 0.001)
< 0.001
Panels A and B report the mean responses to the post-experimental questions for each type of
budgeting and each round of superior. Panels C and D report the results of regression analysis
where the dependent variable is a factor score from a factor analysis of the four postexperimental questions related to psychological contract breach. The dependent variable in
Panels C and D thus indicates extent of psychological contract breach. The dropped variable in
the regressions is participative budgeting (P), which takes the value of one if subjects are under P
and zero otherwise. Consultative (C) and authoritative (A) budgeting are similarly defined and
take the value of one if subjects are in each respective condition and zero otherwise.
37 TABLE 2
Effect of Psychological Contract Breach on Built-in Budgetary Slack
Panel A: Mean (standard deviation) built-in slack by type of budgeting for periods 2 – 5
Authoritative (A) &
Consultative (C)
Participative (P)
Overall
Period 2
34.98%
(31.91%)
n = 49
Period 3
46.07%
(37.30%)
n = 49
Period 4
42.29%
(34.32%)
n =49
Period 5
46.62%
(38.56%)
n = 49
Overall
42.49%
(35.52%)
n = 196
21.98%
(26.03%)
21.03%
(23.16%)
21.22%
(25.03%)
31.08%
(33.04%)
23.83%
(26.81%)
n = 26
n = 26
n = 26
n = 26
n = 104
28.48%
(28.97%)
n = 75
33.55%
(30.23%)
n = 75
31.76%
(29.67%)
n = 75
38.85%
(35.80%)
n = 75
33.16%
(31.16%)
n = 300
Panel B: Repeated measures analysis of variance
Sum of
Factor
df
Squares
Between-subjects
Budgeting
2
3.37
Error
72
Within-subjects
Period
Period x Budgeting
Error
1
2
72
0.460
0.222
F
p (two-tailed)
6.372
0.003
2.710
0.655
0.104
0.522
Panel A reports the mean built-in budgetary slack under authoritative (A), consultative (C), and
participative (P) budgeting. Panel B reports the significance of differences in mean built-in slack
among the three types of budgeting. H2 predicts that mean built-in slack will be higher under A
and C than P. Tests of statistical significance in Panel B are from a repeated measures analysis of
variance in which budgeting type is a between-subjects factor with 3 levels (A, C, P), period is a
within-subjects factor with 4 levels (periods 2 – 5), and %BUILT_IN_SLACK in periods 2 – 5 is
the dependent variable. The dependent variable is the percentage of slack available to
participants that is built in to their budget requests and is calculated as (∑ built
inbudgetaryslack ÷ ∑ maximumavailablebudgetaryslack). All p-values are two-tailed.
38 TABLE 3
Effect of Prior Psychological Contract Breach on Built-in Budgetary Slack
Panel A: Mean (standard deviation) built-in slack by type of budgeting for periods 2 - 5 and
periods 7 – 10
Participative (P) (periods 2 - 5)
23.83%
(26.83%)
n = 26
Participative (P) (periods 7 -10)
36.77%
(51.11%)
n = 27
Panel B: Repeated measures analysis of variance
Factor
Between-subjects
Budgeting
Error
Within-subjects
Period
Period x Budgeting
Error
Sum of
Squares
F
p (two-tailed)
1
51
3.328
9.602
0.003
3
3
51
0.133
0.292
0.991
2.165
0.324
0.147
df
This table reports the mean built-in budgetary slack when psychological contracts have not been
previously breached (P in periods 2 – 5) and when they have been previously breached (P in
periods 7 – 10) (Panel A) and the significance of the differences in mean built-in slack between
these two conditions (Panel B). H3 predicts that built-in slack will be higher when managers
have previously experienced authoritative or consultative budgeting (P in periods 7 - 10) than
when they have not (P in periods 2 - 5). Tests for statistical significance in Panel B are from a
repeated measures analysis of variance in which budgeting type is a between-subjects factor with
2 levels (P in periods 2 – 5 and P in periods 7 – 10), period is a within-subjects factor with 4
levels (the second, third, fourth, and fifth periods of rounds 1 and 2). The dependent variable is
the percentage of slack that is available to participants that is built-in to their budget requests and
is calculated as (∑ built inbudgetaryslack ÷ ∑ maximumavailablebudgetaryslack).
All p-values are two-tailed.
39 TABLE 4
Effect of Authoritative and Consultative Budgeting on Built-in Budgetary Slack
Panel A: Mean (standard deviation) built-in slack by type of budgeting for periods 4 – 5
Period 4
Period 5
Overall
Authoritative (A)
34.60%
38.87%
36.74%
(31.04%)
(42.65%)
(36.84%)
n = 27
n = 27
n = 54
Consultative (C)
51.85%
56.70%
54.28%
(37.23%)
(36.18%)
(36.71%)
n = 25
n = 25
n = 50
Overall
43.23%
47.79%
45.51%
(34.13%)
(39.41%)
(36.77%)
n = 52
n = 52
n = 104
Panel B: Repeated measures analysis of variance
Factor
Between-subjects
Budgeting
Error
Within-subjects
Period
Period x Budgeting
Error
Sum of
Squares
F
1
50
0.799
3.491
0.068
1
1
50
0.054
0.000
1.194
0.005
0.280
0.945
df
p (two-tailed)
This table reports mean built-in budgetary slack under authoritative (A) and consultative (C)
budgeting (Panel A) and the significance of differences in mean built-in budgetary slack between
these two conditions (Panel B). The tests indicate whether built-in slack is lower under A than C
because there are no opportunities to obtain monetary payoffs by building in slack under A. Tests
for statistical significance in Panel B are from a repeated measures analysis of variance in which
budgeting type is a between-subjects factor with 2 levels (A, C), period is a within-subjects
factor with 2 levels (periods 4 and 5), and %BUILT_IN_SLACK in periods 4 and 5 is the
dependent variable. The dependent variable is the percentage of slack that is available to
participants that they build into their budget requests and is calculated as (∑ built
inbudgetaryslack ÷ ∑ maximumavailablebudgetaryslack). All p-values are two-tailed.
40 Appendix
Mean Built-In Budgetary Slack (Standard Deviation) by Period by Budgeting Type
Panel A: Round 1 (Periods 1 - 5)
A
32.13%
(35.24)
n = 27
Period 1
C
29.01%
(26.89)
n = 27
A
36.84%
(32.69)
n = 28
Period 4
C
53.20%
(37.12)
n = 26
P
18.68%
(24.07)
n = 26
P
23.45%
(27.15)
n = 27
A
27.13%
(30.19)
n = 27
Period 2
C
39.38%
(33.78)
n = 26
P
22.78%
(25.86)
n = 27
A
38.87%
(42.65)
n = 27
Period 5
C
56.70%
(36.19)
n = 25
P
31.08%
(33.04)
n = 26
A
42.22%
(37.57)
n = 27
Period 3
C
52.58%
(38.15)
n = 26
P
21.03%
(23.16)
n = 26
(CONTINUED)
41 Appendix (continued)
Panel B: Round 2 (Periods 6 - 10)
Period 6
PA
PC
AP
AC
22.95% 36.49% 55.84% 27.62%
(27.58) (33.47) (36.80) (29.01)
n = 14
n = 13
n = 14
n = 13
PA
25.61%
(32.01)
n = 14
PA
21.18%
(27.89)
n = 14
CA
35.61%
(30.68)
n = 12
CP
43.87%
(37.92)
n = 14
PC
40.37%
(31.01)
n = 13
Period 8
AP
AC
56.93% 58.48%
(37.43) (47.13)
n = 14
n = 14
CA
43.83%
(35.26)
n = 11
CP
50.57%
(42.93)
n = 14
PC
40.80%
(36.17)
n = 13
Period 10
AP
AC
48.61% 45.80%
(40.64) (37.73)
n = 14
n = 14
CA
58.66%
(38.04)
n = 12
CP
42.82%
(42.41)
n = 13
PA
44.39%
(40.99)
n = 14
PA
16.82%
(16.24)
n = 13
PC
46.33%
(43.14)
n = 13
Period 7
AP
AC
53.24% 38.89%
(35.26) (35.12)
n = 14
n = 13
CA
33.55%
(32.65)
n = 12
CP
46.11%
(41.30)
n = 14
PC
36.34%
(30.61)
n = 12
Period 9
AP
AC
44.39% 44.59%
(42.39) (36.97)
n = 14
n = 13
CA
47.94%
(42.34)
n = 13
CP
41.68%
(46.03)
n = 14
This table reports the mean built-in budgetary slack by type of budgeting in rounds 1 (periods 1 – 5) and 2 (periods 6 –
10) of the experiment. There are three budgeting conditions to which participants are assigned in round 1 (Panel A):
authoritative (A), consultative (C), and participative (P). There are six budgeting conditions to which participants are
assigned in round 2 of the experiment (Panel B) because participants experience different budgeting conditions in
rounds 1 and 2. Each condition in round 2 is denoted by two letters. The first (second) letter in Panel B indicates the
condition to which participants are randomly assigned in round 1 (round 2) of the experiment. For example,
participants in the “PA” condition in Panel B are randomly assigned to the participative budgeting condition in round 1
of the experiment and to the authoritative budgeting condition in round 2.
42 
Download