Description
We investigate the influence of total cost of ownership (TCO) information on buyer–supplier negotiations in different
power settings. Based on social exchange theory and recent literature on information processing, we expect that buyers
with detailed TCO information and less power than their negotiation partners may try to (re)gain control over their
own outcomes by sharing information. The results of our experiment indicate that the performance disadvantage of less
powerful buyers is less pronounced when the buyer has detailed TCO information, whereas more powerful buyers do not
seem to be able to profit from TCO information
The e?ect of cost information on buyer–supplier negotiations
in di?erent power settings
Alexandra Van den Abbeele
a,
*
, Filip Roodhooft
a,b
, Luk Warlop
a
a
Katholieke Universiteit Leuven, Naamsestraat 69, 3000 Leuven, Belgium
b
Vlerick Leuven Gent Management School, Vlamingenstraat 83, 3000 Leuven, Belgium
Abstract
We investigate the in?uence of total cost of ownership (TCO) information on buyer–supplier negotiations in di?erent
power settings. Based on social exchange theory and recent literature on information processing, we expect that buyers
with detailed TCO information and less power than their negotiation partners may try to (re)gain control over their
own outcomes by sharing information. The results of our experiment indicate that the performance disadvantage of less
powerful buyers is less pronounced when the buyer has detailed TCO information, whereas more powerful buyers do not
seem to be able to pro?t from TCO information. These somewhat counterintuitive ?ndings are explained through detailed
analysis of the buyer’s negotiation behavior, which shows that less powerful buyers who have access to TCO data use prob-
lem solving techniques more frequently than powerful buyers, who tend to rely on distributive bargaining techniques
instead. We conclude that power can motivate a failure to share TCO information, resulting in less e?ective inter?rm nego-
tiation outcomes.
Ó 2008 Elsevier Ltd. All rights reserved.
Introduction
This study investigates whether buyers bene?t
from total cost of ownership (TCO) information
when negotiating with suppliers across di?erent
power settings. TCO can be seen as an application
of activity-based costing (ABC) to an inter?rm con-
text (Wouters, Anderson, & Wynstra, 2005). It
quanti?es the costs of the activities involved in
acquiring and using purchased goods or services.
As inter?rm relations have become increasingly
important and as purchasing professionals in many
companies still need to demonstrate the contribu-
tion they make to their ?rm, it is relevant for buyers
to identify the management accounting tools that
might contribute to their market success and pro?t-
ability. TCO is a tool that supports purchasing deci-
sions by focusing on all costs related to a purchase
rather than simply on price (Ellram, 1995). Prior lit-
erature suggests that TCO information can improve
inter?rm cooperation, as a buyer can use the quan-
ti?ed cost and cost driver information resulting
from a TCO analysis to communicate to the sup-
plier which activities cause higher costs on his side,
increasing his bargaining power and in turn supply
chain performance (e.g., Ellram, 1995; Roodhooft,
0361-3682/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.aos.2008.05.005
*
Corresponding author. Tel.: +32 16 32 69 36; fax: +32 16 32
67 32.
E-mail addresses: [email protected].
be (A. Van den Abbeele), [email protected]
(F. Roodhooft), [email protected] (L. Warlop).
Available online at www.sciencedirect.com
Accounting, Organizations and Society 34 (2009) 245–266
www.elsevier.com/locate/aos
Van den Abbeele, & Peeters, 2005). However, due to
buyers’ reluctance to share the information needed
for inter?rm cost minimization, ?rms may fail to
maximize possible gains from buyer–supplier nego-
tiations (Baiman & Rajan, 2002). Drake and Haka
(2008) show, for instance, that concerns about ineq-
uity may lead to reluctance to share detailed
accounting information. In this study, we investi-
gate whether bargaining power may prevent buyers
from sharing private TCO information and whether
this might result in less e?ective negotiation out-
comes between buyers and suppliers.
Power can be broadly de?ned as the capacity to
exert in?uence on other people (Kelley & Thibaut,
1978). Although power may derive from a variety
of ‘‘power bases”, it is the mutual dependence of
individuals that allows power di?erences to exist.
Accordingly, in inter?rm relations, power is evi-
denced as the availability of alternatives (e.g.,
Anderson & Dekker, 2005). Prior research indicates
that power di?erences in?uence information search
strategies and drive the processing of information
about other people (e.g., De Dreu & Van Kleef,
2004; Fiske, 1993). These studies suggest that power
may be pivotal in the interplay between, on the one
hand, the need to share information in order to
optimize activities across the supply chain and, on
the other hand, the reluctance to share private
information.
This paper reports the results of a buyer–supplier
negotiation experiment in which we manipulate
both the cost information and the power of the
buyer in order to assess whether access to detailed
TCO information bene?ts buyers across di?erent
power settings. Based on social exchange theory
(e.g., Emerson, 1976; Thibaut & Kelley, 1959) and
recent research on information processing in negoti-
ations (e.g., De Dreu & Carnevale, 2003), we expect
that less powerful buyers who have detailed TCO
information at their disposal may try to (re)gain
control over negotiation outcomes by sharing infor-
mation and creating integrative bargaining situa-
tions. Our results indicate that TCO information
reduces the performance disadvantage of less pow-
erful buyers, and further that this moderation e?ect
is mediated by buyers’ bargaining behavior; that is,
less powerful buyers with TCO information use
more problem solving techniques and fewer distrib-
utive bargaining techniques than more powerful
buyers or buyers that lack TCO information. These
results imply that less powerful buyers can compen-
sate for their power disadvantage by acquiring more
detailed TCO information. However, the results
also suggest that powerful buyers seem unable to
use TCO information to exploit their power advan-
tage in order to obtain even greater individual prof-
its. We therefore conclude that a position of power
may impede the sharing of TCO information, result-
ing in suboptimal negotiation outcomes between
buyers and suppliers.
This paper contributes to the existing literature
on buyer–supplier relations and the use of account-
ing information in three signi?cant ways. First, we
analyze empirically whether buyers pro?t from
TCO information when negotiating with suppliers.
The literature on TCO discusses its adoption and
its potential bene?ts (e.g., Carr & Ittner, 1992; Ell-
ram, 1995; Wouters et al., 2005). For example, these
studies have argued that TCO can be used not only
to compare and evaluate di?erent suppliers, but also
to provide decision-makers with an objective and
easily understood argument to support purchasing
decisions. It has also been suggested that a buyer
can use the cost and cost driver information result-
ing from a TCO analysis to communicate to the
supplier which activities cause higher costs on his
side, increasing his bargaining position (e.g., Ellram,
1995; Roodhooft, Hiel, Van den Abbeele, & van
Doveren, 2003, 2005). However, there is little empir-
ical research on the use of TCO during actual
buyer–supplier negotiations. To our knowledge, this
study is the ?rst to investigate whether buyers can
bene?t from TCO when negotiating with suppliers.
Our second contribution is to add to a small but
growing body of work on the role of accounting
information in negotiations. The existing research
on accounting negotiation has mainly focused on
three intra?rm issues: collective bargaining (e.g.,
Craft, 1981; Waterhouse, Gibbins, & Richardson,
1993), transfer pricing (e.g., Kachelmeier & Towry,
2002; Luft & Libby, 1997), and budgeting (e.g.,
Fisher, Frederickson, & Pe?er, 2000, 2006). The lit-
erature on inter?rm accounting negotiations has pri-
marily concentrated on auditor–client interactions
(e.g., Bame-Aldred & Kida, 2007; Gibbins, Salterio,
& Webb, 2001; Ng & Tan, 2003; Windsor & Ash-
kanasy, 1995). Although some of these studies con-
sider the importance of power in negotiations (e.g.,
Gibbins et al., 2001; Ng & Tan, 2003; Windsor &
Ashkanasy, 1995), they typically do not consider
the role of sharing private cost information. Fur-
thermore, auditor–client negotiations are often
viewed as mainly distributive. An important reason
is that audit clients seem to be inclined to negotiate
246 A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266
within a single-issue framework, which acts as a bar-
rier to ?nding integrative solutions (Gibbins, McC-
racken, & Salterio, 2005). Conversely, we expect
(and our results provide evidence) TCO to be a tool
that may support buyers in adopting a multiple-
issue framework during buyer–supplier negotia-
tions, which may enable buyers to create integrative
bargaining situations. Overall, accounting research
on inter?rm negotiations beyond the speci?c audit-
ing context has been very limited. This paper
answers recent calls for empirical research ‘‘with a
greater emphasis upon business processes and the
use of accounting in action/negotiation” (Tomkins,
2001, p. 164).
Our third contribution lies in the application of a
more sophisticated causal model to test for medi-
ated moderation e?ects of negotiation behavior on
negotiation outcomes. In an extensive review of the-
ory-consistent management accounting research,
Luft and Shields (2003) conclude that the majority
of management accounting studies employ additive
causal-model forms. Because such forms may limit
our understanding of management accounting by
representing causes and e?ects as universal rather
than conditional on other mediating or moderating
variables, Luft and Shields (2003) recommend the
use of more sophisticated causal models.
1
Muller,
Judd, and Yzerbyt (2005) observe that mediation
and moderation may be combined in informative
ways to better understand causal e?ects. However,
while well-developed and validated procedures exist
for examining whether treatment e?ects on an out-
come are mediated and/or moderated, very few
studies analyze mediation and moderation simulta-
neously. This study uses a rigorous three-step
approach to causal modeling in order to test for
mediated moderation e?ects (cf. Baron & Kenny,
1986). Our analyses indicate that TCO information
reduces the performance disadvantage of less pow-
erful buyers, and that this moderation e?ect is med-
iated by buyers’ bargaining behavior. Studying this
mediated moderation e?ect allows us to explain the
seemingly counterintuitive negotiation outcomes
that we ?nd.
The next section reviews the relevant literature
and develops our hypotheses. Section three presents
an outline of the experimental procedures and sec-
tion four reports the results. Section ?ve provides
discussion and suggestions for future research.
Theoretical background and hypotheses
Our hypotheses are based upon social exchange
theory
2
(e.g., Emerson, 1976; Thibaut & Kelley,
1959). Social exchange theory describes negotiation
as a process characterized by information exchange,
joint problem solving, and persuasion. In this
framework, negotiation outcomes (e.g., level of
buyer and/or supplier satisfaction, pro?ts, whether
or not agreement is reached) result from the com-
plex interaction of three constructs: negotiator char-
acteristics, situational characteristics, and the
characteristics of the negotiation process itself
(Campbell, Graham, Jolibert, & Meissner, 1988).
Negotiator characteristics and situational character-
istics are seen as a?ecting both process-related
behaviors and performance outcomes. In this study,
we control for negotiator characteristics and focus
on two situational characteristics: the availability
of TCO information and the relative power of the
buyer. Based on social exchange theory, we predict
that the e?ect of these situational characteristics on
negotiation outcomes is mediated by the negotiation
process.
The negotiation process may be thought of as a
series of interactions (Kelley & Thibaut, 1978).
According to social exchange theory, communica-
tion between parties during these interactions is very
1
More speci?cally, if the proposed causal model is additive,
Luft and Shields (2003, p. 198) recommend providing both the
reasons for assuming there are no important mediation or
moderation e?ects and the consequences of omitting these
relations if they exist. In other words, the default expectation is
that there are important mediation and moderation e?ects that
we should account for.
2
Although economic models have become the mainstream
explanatory models in many business disciplines, it has been
argued that advice to negotiators should depend on an under-
standing of the opponent’s actual decision process, rather than on
the assumption that the other party is fully rational (Bottom,
Holloway, Miller, Mislin, & Whitford, 2006). Indeed, recent
research suggests that negotiation behavior is seldom rational
(for a review, see Bazerman, Curhan, Moore, & Valley, 2000).
For example, evidence shows that despite the existence of an
agreement zone, deviations from rationality in individual deci-
sions lead to disagreements and Pareto-ine?cient agreements.
The problem with economic models is that researchers are subject
to the empirical indeterminacy of economic phenomena (Emer-
son, 1976). In most real-life situations, many di?erent outcomes
(from full cooperation to near-disastrous con?ict) are consistent
with ‘‘rationality”. Thus, game-theoretic models rarely provide
unique predictions of negotiation outcomes. Behavioral decision
models, such as exchange models, o?er a set of adjustments to
rational models (Murnigham & Bazerman, 1990).
A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266 247
important, as it shapes their perceptions of each
other and their interpretation of subsequent actions
(e.g., Bottom et al., 2006).
3
Communication in?u-
ences the development of a relationship, because it
provides an opportunity to express interest in the
exchange partner, to ascertain the exchange part-
ner’s openness to reciprocity and to in?uence the
exchange partner’s perceptions. Generally, two
basic types of negotiation behavior are distin-
guished within the negotiation process: problem
solving and distributive bargaining (Walton &
McKersie, 1966). Problem solving primarily
involves discovering ways to increase the bene?ts
available in the relationship (Walton & McKersie,
1966). In the context of purchasing, bargaining is
integrative to the extent that buyers actively seek
coordination with suppliers to develop alternative
purchasing arrangements that are likely to reduce
costs and/or increase performance. The intent is to
identify a solution via open and accurate informa-
tion exchange and to make trade-o?s based on pri-
orities and mutual respect for the other party’s
individual goals (Campbell et al., 1988). In contrast,
distributive bargaining is a process whereby each
party tries to maximize their private outcome (Wal-
ton & McKersie, 1966). Distributive bargaining is
characterized by the use of ?xed-sum or ‘‘win–lose”
tactics such as communicating threats, excessive
demands, positional commitments, and persuasive
arguments (Campbell et al., 1988). Prior studies
indicate that distributive bargaining is ‘‘natural”
or more salient to negotiators than problem solving
(e.g., Weingart, Hyder, & Prietula, 1996). Naive
negotiators are more familiar with distributive tac-
tics and therefore use these tactics as default
approaches. This ties in with their assumption that
the negotiation is a ?xed-sum game (Bazerman
et al., 2000).
The main e?ects of cost information and power on
negotiation behavior and outcomes
Cost information
Aprerequisite for any buyer–supplier relationship
is that cost information between exchange partners is
shared (e.g., Dekker, 2003, 2004; Tomkins, 2001).
Shared cost information can be used to analyze the
value chain in order to identify cost-reduction oppor-
tunities across the companies’ boundaries. Value
chain analysis (VCA), introduced by Porter (1985)
and further developed by Shank (1989) and Shank
and Govindarajan (1992), allows one to analyze,
coordinate, and optimize linkages between activities
in the value chain by focusing on the interdependence
between these activities (Dekker, 2003).
Traditionally, purchasing was simply a matter of
negotiating the best price. Accordingly, under tradi-
tional management accounting, purchasing decisions
tend to track only the purchase price associated with
a particular supplier, burying all the other costs the
supplier may introduce in the value chain of the pur-
chasing organization (Carr & Ittner, 1992; Degraeve
&Roodhooft, 1999). Traditional accounting systems
are therefore unable to adequately support a VCA
(Porter, 1985). In contrast, total cost of ownership
(TCO) systems, which attempt to quantify all the
costs associated with the purchase of a given product
or service, presume the existence of boundary-span-
ning activities (Wouters et al., 2005) and hence start
froma value chain perspective (Shank &Govindara-
jan, 1992). The cost and cost driver information
resulting from TCO analysis can be used to optimize
the performance of activities across the supply chain
(Porter, 1985). For instance, suppliers can be made
aware of the extra costs they generate and also of
ways to improve their competitive position by reduc-
ing these costs on the buyer’s side (e.g., Ellram, 1995;
Roodhooft et al., 2003, 2005). TCO information
therefore helps buyer–supplier partners detect
trade-o?s along the value chain and improve pro?t-
ability by modifying the way in which they do busi-
ness with each other (Wouters et al., 2005).
Accordingly, we expect that:
H1a: Buyers with TCO information achieve
higher individual pro?ts than buyers with
traditional cost information.
In order to reach integrative agreements that
lower the total costs over the value chain, buyers
need to have a good understanding of their own
3
Verbal communication between buyers and suppliers, a very
basic feature of inter?rm negotiations, has not been systemati-
cally examined in accounting research. One reason may be that,
according to economic models, such communications are irrel-
evant. The rationale is as follows: verbal claims made by either
party are not veri?able and thus not strategically credible, so they
will be ignored and should not in?uence the actions taken by
rational self-interested counterparts (Bottom et al., 2006). This
paper reports a detailed analysis of verbal communication of
accounting information during buyer–supplier negotiations. The
results provide clear evidence of the importance of this type of
communication.
248 A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266
priorities, communicate these to the supplier, and
integrate information about suppliers’ preferences
into their own understanding of the problem at
hand so that they can make trade-o?s between
important and unimportant issues (De Dreu, Bee-
rsma, Stroebe, & Euwema, 2006). Prior research
suggests that the availability of quanti?ed informa-
tion increases the likelihood that buyers engage in
such problem solving e?orts. For example, evidence
has been provided that ?nancial quanti?cations
carry more weight in decisions than non-?nancial
information (Ittner, Larcker, & Meyer, 2003), that
quanti?ed information enhances persuasion, espe-
cially if it is regarded as objective (Kadous, Koonce,
& Towry, 2005), and that relatively abstract costs
are less salient than outright losses (Northcraft &
Neale, 1986). Since TCO systems attempt to quan-
tify all relevant costs (as opposed to traditional cost
systems, which focus on price and do not quantify
other costs), and since quanti?ed costs carry more
weight in decisions, TCO systems provide better
insights into the relative importance of di?erent
costs. Hence, they provide relevant information that
enlarges the set of possible outcomes in a negotia-
tion, which in turn leads to better problem solving
(e.g., Kersten, 2001). As buyers have greater insight
into their own constraints and objectives, and as the
quanti?ed costs within TCO information are more
persuasive than traditional accounting information,
buyers are better-placed to cooperate with suppliers
on identifying and assessing alternative courses of
action. Thus, the availability of TCO information
should encourage buyers to engage in problem solv-
ing e?orts. Accordingly, we expect that:
H1b: Buyers with TCO information use more
problem solving techniques than buyers
with traditional cost information.
Power
Among the variables identi?ed as factors in?u-
encing negotiation processes and outcomes, power
is one of the most important (De Dreu & Van Kleef,
2004). Existing research on negotiations in account-
ing settings shows that various sources of power
may have an important impact on negotiation pro-
cesses and outcomes (e.g., Fisher et al., 2000; Ng &
Tan, 2003). In particular, the power that results
from the availability of alternative negotiation part-
ners is a key element of many negotiations: indeed,
if other negotiation partners are available, this
reduces a party’s dependence on the other side,
which weakens the other party’s power position
(Giebels, De Dreu, & Van de Vliert, 2000; Thibaut
& Kelley, 1959). Further, empirical research shows
that negotiators with a viable alternative achieve
higher personal outcomes than negotiators without
alternative options or with less attractive ones:
negotiators with less power tend to demand less,
make more concessions, and achieve less successful
outcomes than those with more power (Pinkley,
Neale, & Bennett, 1994).
4
We therefore expect that:
H2a: More powerful buyers achieve higher indi-
vidual pro?ts than less powerful buyers.
Turning again to negotiation behavior, a buyer’s
relative power is expected to a?ect his bargaining
aggressiveness. In particular, buyers’ use of distribu-
tive bargaining techniques, which include implicit or
explicit threats (e.g., warning the supplier that he is in
danger of losing the contract), are likely to rise with
supplier competition. First, such techniques are more
credible when several suppliers are interested in a
contract. Second, increased supplier competition
reduces the buyer’s need to solve problems with
any given supplier or group of suppliers (Perdue &
Summer, 1991). Consequently, we expect that:
H2b: More powerful buyers use fewer problem
solving techniques and more distributive
bargaining techniques than less powerful
buyers.
The mediated moderation e?ect of cost information
on the relation between power and individual pro?t
Prior research indicates that power not only in?u-
ences strategic decisions such as the use of distribu-
tive bargaining techniques during negotiation, but
also drives the processing of information (e.g., De
Dreu & Carnevale, 2003). For instance, compared
to individuals with a power advantage, those at a
power disadvantage are more motivated to under-
stand the other’s needs, desires, and possible actions
(Gelfand & Christakopoulou, 1999) and seek more
situational control by making a positive impression
(Goodwin, Gubin, Fiske, &Yzerbyt, 2000). Further,
negotiators with less power ask diagnostic rather
4
For a more extensive review of the e?ects of power, see Van
Kleef, De Dreu, Pietroni, and Manstead, 2006.
A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266 249
than leading questions and belief-congruent rather
than incongruent questions when facing a competi-
tive partner rather than a cooperative one (De Dreu
& Van Kleef, 2004). These ?ndings suggest that,
compared to more powerful individuals, less power-
ful individuals have a higher epistemic motivation,
i.e., a greater tendency to engage in information
acquisition and processing, and thus a greater desire
to develop and maintain a rich and accurate under-
standing of the opponent’s intentions and behavior
(Van Kleef et al., 2006). Because of their high episte-
mic motivation, dependent buyers may try to
(re)gain control over their own outcomes by paying
close attention to suppliers in order to accurately
predict their intentions and behaviors (De Dreu &
Van Kleef, 2004). In contrast, powerful individuals
are less likely to seek detailed information as they
have more resources available and/or are less depen-
dent on others, enabling them to act as they want
without serious consequences (Fiske, 1993; Keltner,
Gruenfeld, & Anderson, 2003; Van Kleef et al.,
2006). Because of their lower levels of epistemic
motivation, powerful buyers are more likely to solve
problems and to form an impression of their oppo-
nent through a quick, e?ortless, and heuristic pro-
cessing of information that rests on well-learned
prior associations (De Dreu & Van Kleef, 2004).
As distributive bargaining behavior is more salient
to negotiators than problem solving behavior (e.g.,
Weingart et al., 1996), we expect powerful buyers
to rely on distributive bargaining rather than on
problem solving techniques.
With regard to the implication of power across
TCO versus traditional information, it was already
pointed out that buyers with traditional informa-
tion, irrespective of whether they have power or
not, are encouraged to focus on price; these buyers
are less likely to take into account additional costs
that the supplier may cause. This limited scope of
information leaves traditional buyers (as opposed
to TCO buyers) with fewer opportunities to propose
a mutually acceptable solution, thus decreasing the
likelihood of problem solving e?orts.
Accordingly, we expect that less powerful buyers
with TCO information may try to (re)gain control
over their own outcomes by sharing information
5
and paying close attention to their opponent’s inten-
tions and behavior. That is, less powerful buyers
with TCO information are more likely to use inte-
grative bargaining techniques than either more pow-
erful buyers with TCO information, who are more
likely to using distributive bargaining techniques,
or buyers without TCO information, who will be
less able to create an integrative bargaining situa-
tion as their traditional cost information encourages
them to focus solely on price. We therefore expect
that:
H3a: Cost information moderates the e?ect of
power on buyers’ pro?ts, such that TCO
information reduces the performance disad-
vantage that less powerful buyers have com-
pared to more powerful buyers.
H3b: Cost information moderates the e?ect of
power on buyers’ bargaining behavior, such
that TCO information encourages less pow-
erful buyers to increase the use of problem
solving techniques and to decrease the use
of distributive bargaining techniques as
compared to more powerful buyers.
H3c: Buyers’ use of problem solving and distrib-
utive bargaining techniques mediate the
moderation e?ect of cost information on
the relation between power and buyers’
pro?ts.
To summarize, we expect that the overall e?ect of
power on the negotiation outcome will be moder-
ated by cost information and that this moderation
e?ect will be in?uenced by the e?ect of power and
cost information on negotiation behavior (the medi-
ator). These expectations are consistent with a med-
iated moderation hypothesis (Muller et al., 2005), as
represented in Fig. 1.
Research method
Experimental design
To test our hypotheses, we developed a 2 (TCO
information versus traditional cost informa-
tion) Â 2 (equal power versus low power) experi-
mental design, in which buyers and suppliers
negotiate a lease contract for a set of machines.
We manipulated the cost information and power
of the buyer in order to assess whether access to
detailed TCO information bene?ts buyers across
di?erent power settings. Cost information was
5
We do not expect that less powerful buyers will hoard their
private TCO information, as this would not be to their bene?t. In
order to bene?t from TCO information, this needs to be
exchanged so that trade-o?s between important and unimportant
issues can be made (Baiman & Rajan, 2002; De Dreu et al., 2006).
250 A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266
manipulated by providing the buyer with either
TCO information or traditional cost information
for the purchasing decision. The TCO information
quanti?ed all relevant costs (in this case: price, spare
parts, and maintenance) associated with the pur-
chase of a given quantity of products or services
from a given supplier. All costs were expressed in
the same currency as the price. Buyers with tradi-
tional information were given only an indication
of the costs and of the relative importance of each
of the issues to be negotiated (cf. Appendix A). This
corresponds to traditional management accounting
practices, which only track the purchase price and
‘‘bury” other purchasing costs in overhead accounts
or general expenses (Carr & Ittner, 1992). In addi-
tion, participants were informed that price is the
most expensive cost, followed by maintenance and
then spare parts. This information on the relative
importance of the di?erent issues provided buyers,
who had traditional cost information at their dis-
posal, with an understanding of their own prefer-
ences and priorities, which they could then
communicate to the suppliers in order to reach an
integrative solution.
We manipulated power through the availability
of an outside option (e.g., Giebels et al., 2000; Pink-
ley et al., 1994). In the equal power setting, the writ-
ten instructions for both the buyer and the supplier
contained a short paragraph about the presence of
an alternative negotiation partner. Buyers and sup-
pliers were equally powerful, in the sense that they
both had an outside option worth 1000 Euro. This
outside option was relatively unattractive, however,
as higher gains could be obtained by reaching an
agreement. In the low power setting, buyers had
no outside option and were fully dependent on
reaching an agreement with their partner to earn
any money.
6
Suppliers always had an outside option
of 1000 Euro and full cost and price information.
where
power
cost information
power*cost information
negotiation behavior
individual profit
= the manipulated independent variable
= the manipulated moderation variable
= the moderation of cost information on the effect of power on individual profit
= the mediation variable (in Tables 2, 3, and 4 measured by information
exchange, distributive behavior, and problem solving approach, respectively)
= the negotiation outcome variable
power
cost information
power*cost information
negotiation behavior
- information exchange
- distributive behavior
- problem solving approach
individual profit
Fig. 1. Regression models illustrating moderated mediation.
6
Consistent with prior research (for an overview, see Wolfe &
McGinn, 2005), the individual parties were aware of their own
alternatives but not the other party’s. If the buyer had no
alternative, the buyer was told that in the short run the only
supplier that could deliver the required machines was the supplier
they were negotiating with. This made the buyer highly depen-
dent on the supplier, but the supplier was unaware of this. This
situation allows us to fully focus on the e?ects of the buyer’s
situation (information and power) and how this a?ects negoti-
ation behavior and outcomes, without having to deal with the
complexity of the supplier’s perceptions of his or her power over
the buyer. This manipulation also has the additional advantage of
being a strong manipulation from the viewpoint of the buyer,
without having to deal with the ‘‘obvious” result of virtually all of
the pro?ts going to the supplier if the latter was aware of the
buyer’s complete dependence.
A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266 251
Subjects and procedures
Two hundred and eight participants were
recruited from a graduate management accounting
course that was part of a Master’s program in busi-
ness administration at a large West-European uni-
versity. The course had covered traditional
accounting methods, activity-based costing, TCO,
and supplier selection problems before the experi-
ment took place. The experiment was run in a com-
puter laboratory. Each session was restricted to a
maximum of one hour. The opportunity to earn a
performance-based cash payout was the only incen-
tive o?ered. Participants earned 0.5% of their com-
pany’s pro?ts (on average 5.74 Euro; min = 0
Euro; max = 15 Euro). The participants were ran-
domly assigned to one of the four experimental con-
ditions and to a buyer or a supplier role. Procedures
were identical for all treatments.
Buyers and suppliers sat in di?erent rooms so
that they were unable to identify their partner’s
identity; personality e?ects and collusion were there-
fore precluded. Participants read the instructions
describing their role and the nature of the bargain-
ing task and could play the game at their own pace.
The supplier started the game by making a ?rst
o?er. Participants could send messages along with
their o?ers and countero?ers if they desired to do
so. All messages were recorded. The game ended
when (i) an agreement was reached, (ii) a player
opted for the outside option, or (iii) after 10 rounds.
In the last case, participants were informed by the
computer program that time was running out. This
only happened in a few cases (4 out of 104).
The bargaining task
The bargaining task is based on a negotiation
game developed by Kelley (1966) and applied by
many other researchers (e.g., Campbell et al.,
1988; De Dreu & Van Kleef, 2004). The game was
adapted to suit a TCO setting. This means that
the payo? tables in Kelley’s game were replaced
by cost tables for the buyers and cost and income
tables for the suppliers. The tables were constructed
in such a manner that the minimum and maximum
pro?ts that could be earned were the same for buy-
ers and suppliers (cf. Appendix A). Buyers and sup-
pliers had to negotiate a lease contract for a set of
machines. The buyer could earn a ?xed income (of
6000 Euro) by selling end products to an end cus-
tomer. The instructions explained that maintenance
and spare parts were needed to run the machines
and to produce the end product. Consequently,
the game involved the simultaneous negotiation of
price, maintenance, and spare parts. For each of
these issues, nine di?erent contract terms were
possible.
Price was an income for the supplier and a cost
for the buyer. Accordingly, the price issue was dis-
tributive in nature. This issue was worth the same
for each negotiator, with preferences on the issue
going in opposite directions. Consequently, one
party’s gain was equal to the other party’s loss.
The task, however, provided an opportunity for
the parties to integrate their personal interests.
The buyer had a comparative advantage in taking
care of the spare parts and the supplier had a com-
parative advantage in maintaining the machines.
Since the issue that was most valuable to one party
was automatically less valuable to the other party,
it was possible for participants to trade-o? issues.
Such ‘‘logrolling”, giving up on less valuable issues
to maximize outcomes on the most valuable ones,
could yield a fully integrative solution or a Pareto-
optimal solution (Kersten, 2001). The Pareto-opti-
mal solution is the solution whereby neither dyad
member can improve his situation without the other
party being worse o?; in other words, no other com-
bination of contracts o?ers as much or more pro?t
to both parties. The Pareto-optimal joint outcome
could be reached when the buyer and supplier
agreed on contract 5AZ. In this agreement, the dis-
tributive issue (price) is set at the middle and the two
integrative issues (maintenance and spare parts) are
fully traded o?. In this situation, denoted with an
asterisk (
*
) in Appendix A, the maximum level of
joint pro?t was reached and was equally divided
between the buyer and the supplier.
7
As cost tables
were private, participants had to ?nd out the possi-
bility of a win–win solution through the process of
o?ers and countero?ers and by exchanging informa-
tion about their interests.
Before the negotiations started, participants were
informed that they should realize that their oppo-
nent’s cost structure did not mirror their own cost
structure. Participants were reminded that, just like
7
The price contract did not in?uence the joint pro?ts:
maximum joint pro?ts were reached as long as participants
agreed on contract AZ. Nine (1AZ to 9IZ) di?erent contract
combinations led to maximum joint pro?ts. However, only
contract 5AZ divided the maximum joint pro?t equally between
the buyer and the supplier.
252 A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266
in real life, companies may have di?erent competi-
tive advantages and that, as a result, their company
might be competitively better at maintenance than
their opponent, while their opponent might be more
competitive in spare parts (or vice versa).
8
Further-
more, the instructions informed participants that at
all times during the negotiation a table would be
shown on their screen with an overview of their
opponent’s proposals and their own counter-pro-
posals, so that they would have an overview of all
proposals already made.
9
Dependent variables
Negotiation outcomes
As we want to assess whether buyers can bene?t
from more re?ned TCO data in making purchasing
decisions, this study focuses on the buyer’s individual
pro?t. This is the amount of money earned by a
buyer at the conclusion of the negotiations; it is
measured as the buyer’s individual pro?t level asso-
ciated with ?nal agreement in the negotiation.
Although this is not central to the analysis, the
study also brie?y discusses the results relating to
the supplier’s individual pro?t and the dyad’s joint
pro?t, as these provide a deeper understanding of
the study’s main ?ndings.
Negotiation behavior
Negotiation behavior is derived from two sets of
measures: (1) a set based on interaction analysis,
and (2) a set based on participants’ responses to
the post-game questionnaire administered immedi-
ately following the negotiation exercise.
First, verbal behavior is coded by means of inter-
action analysis in order to examine categories and
meanings embedded in structural patterns of com-
munication (Putman & Fairhurst, 2001). The classi-
?cation scheme, which is included in Appendix B1,
is based on negotiation communication coding
schemes used in prior studies (e.g., Giebels et al.,
2000). Three judges, who were blind to conditions
or hypotheses, coded each negotiation indepen-
dently. Inter-rater agreements, expressed in Cohen’s
Kappa, varied between 0.75 and 0.95. After com-
pleting the coding, the coders compared their cod-
ing and resolved disagreements by jointly revisiting
the negotiation messages and producing a single
set of codes for each subject. Negotiation behavior
is determined by analyzing this last set of codes.
Messages sent by participants were coded for (i)
problem solving techniques, measured as informa-
tion exchange and as other integrative behaviors,
and (ii) distributive bargaining techniques, such as
sending threats or issuing warnings. Information
exchange re?ects the extent to which a participant
shared information about priorities. Information
exchange was coded ‘‘0” for participants not reveal-
ing any information about their cost structure, ‘‘1”
for participants revealing the relative importance
(but not the numerical values) of each of the three
issues under negotiation, and ‘‘2” for participants
revealing both the relative importance and the
numerical values of each of the three issues under
negotiation. Integrative behavior is based on the
average of two measures: the number of rewards
o?ered and the number of positive normative
appeals extended (Cronbach’s alpha = 0.70). Inte-
grative behavior is also assessed by counting the
number of explicit requests for cooperation. How-
ever, as the number of requests for cooperation
did not load with the rewards and the positive nor-
mative appeals on one factor, it is treated as a
8
This was illustrated in the instructions by an example relating
to maintenance. Buyers with TCO information were referred to
the cost table for cells 1 and 2 in Appendix A: if they agreed on
maintenance contract ‘‘A”, their company would have a relatively
low maintenance cost (€ 250), but the supplier would have to
provide the maximum level of maintenance. It was explained that
this did not mean that the supplier’s cost for providing this level
of maintenance would be € 2250, as this was the cost to the buyer
if he were to provide the maximum level of maintenance himself;
rather, this amount was di?erent from what the supplier would
have to pay in order to provide the same maintenance level, since
the cost structures of both companies were di?erent. The
instructions also explained that the same reasoning was to be
applied to spare parts. Buyers with traditional cost information
were referred to the cost table for cells 3 and 4 in Appendix A and
were provided with the same text, except that the cost ?gures were
replaced by the corresponding number of maintenance sessions
(as provided in the cost table for cells 3 and 4 in Appendix A).
9
The information in this table was essentially the same in all
experimental cells. The table provided an overview of all
proposals made by both players (e.g., proposal supplier: 6EV;
counter-proposal buyer: 4DW; counter-proposal supplier: 5CA;
etc.). Thus, at all times during the negotiation, participants could
check the history of proposals and counter-proposals made.
Participants with TCO information also received the correspond-
ing pro?t or loss ?gure for the di?erent o?ers. This ?gure
provided no new or extra information, as participants with TCO
information could easily compute this ?gure from the informa-
tion provided in their private cost tables. This information was
solely provided for the purpose of speeding up the game. The
pro?t/loss ?gure was not provided in experimental cells with
traditional information: participants in these experimental cells
could derive the corresponding number of maintenance sessions
and spare parts from their private cost tables, but not the exact
corresponding costs.
A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266 253
separate measure. Distributive behavior is a sum-
mated scale calculated by adding ?ve types of dis-
tributive bargaining techniques together and
determining the mean value. The ?ve behavior types
included are as follows: general threats, exit threats,
warnings, commitment and punishments (Cron-
bach’s alpha = 0.81).
Second, negotiation behavior is derived from
post-game questionnaires. Participants rated their
opponent’s bargaining behavior on four items that
measure the overall problem solving approach of
the buyers as assessed by the supplier (Cronbach’s
alpha = 0.88). The di?erent items are listed in
Appendix B2 and are based on prior studies (e.g.,
Campbell et al., 1988). A summated scale was
obtained by calculating the mean value across the
four items.
Results
Experimental checks
After completing the negotiation task, partici-
pants ?lled out a post-bargaining questionnaire on
a 1–5 scale, enabling us to check for motivation,
task understanding, and their usage of cost reports.
All checks on the experimental conditions (on cost
information relevance and power) are statistically
signi?cant and have means in the appropriate direc-
tion. More powerful buyers considered themselves
more powerful (F(1, 102) = 14.61, p < 0.01) and
buyers with TCO information judged the cost infor-
mation they received to be more relevant than buy-
ers with traditional cost information (F(1,
102) = 41.62, p < 0.01). Checks on procedures,
including the subject’s involvement in the task, their
understanding of instructions and payo? tables, and
whether they had su?cient time to complete the
exercise, show no di?erences between conditions
(p > 0.10), as is appropriate. Means on these ques-
tions indicate that subjects were highly involved
(mean = 4.33; std = 0.69), that they assessed the
exercise as ‘‘fun” (mean = 4.07; std = 0.70), that
they understood both the instructions (mean = 4.41;
std = 0.77) and the payo? tables (mean = 4.62;
std = 0.51), and that they had enough time to com-
plete the task (mean = 4.40; std = 0.98). Partici-
pants in di?erent experimental cells required an
equal amount of time to read the instructions and
familiarize themselves with the game before actually
starting the negotiation (p > 0.10). On average, par-
ticipants needed 559 s (9.3 min) to read the instruc-
tions. Participant gender produced neither main nor
interaction e?ects on negotiation process or out-
comes and is therefore excluded from further
analysis.
The e?ects of cost information and power on
negotiation outcomes
Table 1 presents the results.
10
A correlation
matrix of all variables included in the study is pro-
vided in Appendix C. As expected in H1a, buyers
with TCO information obtained signi?cantly higher
individual pro?ts than buyers with traditional cost
information (F(1, 100) = 52.90, p < 0.01). Consis-
tent with H2a, more powerful buyers generated
higher individual pro?ts than less powerful buyers
(F(1, 100) = 14.02, p < 0.01). The results also indi-
cate a moderation e?ect of cost information on
the overall e?ect of power on individual pro?t
(F(1, 100) = 4.58, p < 0.05). The performance de?cit
of less powerful buyers actually disappeared when
they had TCO information: the individual pro?t
made by buyers with TCO information and an out-
side option did not signi?cantly di?er from the
pro?t made by buyers who had access to the same
information but no outside option (buyer’s mean
individual pro?t = 1148.33 versus buyer’s mean
individual pro?t = 1007.69; F(1, 54) = 2.46,
p > 0.10). The performance de?cit of less powerful
buyers was thus less pronounced when buyers had
detailed TCO information, supporting H3a.
The e?ects of cost information and power on
negotiation behavior
To explain negotiation outcomes, we analyze
participants’ negotiation behavior. First, we test
H1b, which predicts that buyers with TCO informa-
tion use problem solving techniques more frequently
than buyers with traditional cost information. An
ANOVA on information exchange reveals a main
e?ect for cost information (F(1, 100) = 6.14,
10
Analyses are performed for all subjects in the study. However,
we obtain similar results when excluding the subjects that opted
for an outside option and thus did not reach agreement. Only 14
out of 208 subjects opted for this outside option, which may be
explained by the fact that the outside option was relatively
unattractive, as higher gains could be obtained by reaching an
agreement. Pairwise comparisons with Tukey tests indicate that
the number of dyads choosing for an outside option is not
signi?cantly di?erent across the four experimental conditions
(p > 0.10).
254 A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266
p < 0.05): buyers with TCO information disclosed
more information than buyers with traditional
information. An ANOVA on integrative behavior
indicates that the e?ect for cost information (F(1,
100) = 2.39, p = 0.12) is not signi?cant at the 10%-
level, but closely approaches signi?cance. An
ANOVA on the number of explicit requests for
cooperation shows a main e?ect for cost information
(F(1, 100) = 10.45, p < 0.01), indicating that buyers
with TCO information requested active cooperation
more frequently than buyers with traditional cost
information. Buyers’ negotiation behavior is further
examined by an ANOVA on the composite measure
problem solving approach. A signi?cant main e?ect
for cost information is found (F(1, 100) = 242.00,
p < 0.01), indicating that, according to the suppliers,
buyers with TCO information used problem solving
techniques more frequently than buyers with tradi-
tional information. Although the e?ect of cost infor-
mation on integrative behavior is somewhat weaker,
the e?ects of cost information on information
exchange, requests for cooperation and problem solv-
ing approach, provide strong support for H1b. The
ANOVA on distributive behavior reveals an unex-
pected main e?ect of cost information (F(1,
100) = 5.77, p < 0.05), indicating that TCO infor-
mation signi?cantly increased the use of distributive
bargaining techniques. Unreported Tukey tests indi-
cate that this e?ect is mainly driven by buyers with
TCO information and an outside option, as these
buyers used distributive bargaining techniques more
frequently than buyers in the other three conditions
(p < 0.01). The Tukey tests also indicate that buyers
with TCO information but no outside option did not
use distributive bargaining techniques more fre-
quently than buyers in the two conditions with tra-
ditional information (p > 0.10).
In order to test H2b, we analyze the e?ect of
power on the negotiation behavior variables. An
ANOVA on information exchange shows a main
Table 1
Analysis of negotiation outcomes and buyer’s behavior
Panel A: Summary statistics for negotiation outcomes and buyer’s behavior
a
TCO information Traditional cost information
Outside option No outside option Outside option No outside option
Buyer’s individual pro?t 1148.33(184.99) 1007.69(449.59) 698.08 (479.06) 181.81 (616.74)
Supplier’s individual pro?t 1188.33 (205.81) 1438.46 (349.66) 1421.15 (427.83) 1881.81 (680.61)
Joint pro?t 2336.67 (324.29) 2446.15 (506.15) 2119.23 (263.85) 2063.64 (405.38)
Information exchange 0.37 (0.49) 0.73 (0.78) 0.23 (0.43) 0.32 (0.48)
Integrative behavior 0.02 (0.09) 0.27 (0.55) 0.06 (0.16) 0.05 (0.15)
Requests for cooperation 0.17 (0.46) 0.69 (0.88) 0.08 (0.27) 0.09 (0.29)
Distributive behavior 0.57 (0.73) 0.06 (0.11) 0.18 (0.30) 0.05 (0.14)
Problem solving approach 2.98 (0.71) 3.62 (0.57) 1.47 (0.40) 1.70 (0.45)
Panel B: ANOVA for negotiation outcomes and buyer’s behavior
b
Cost information Power Cost information
*
power
Buyer’s individual pro?t 52.90 (
***
) 14.02 (
***
) 4.58 (
**
)
Supplier’s individual pro?t 15.80 (
***
) 17.46 (
***
) 1.53
Joint pro?t 15.78 (
***
) 0.13 1.20
Information exchange 6.14 (
**
) 4.16 (
**
) 1.56
Integrative behavior 2.39 4.13 (
**
) 5.02 (
**
)
Requests for cooperation 10.45 (
***
) 6.37 (
**
) 5.73 (
**
)
Distributive behavior 5.77 (
**
) 14.20 (
***
) 4.89 (
**
)
Problem solving approach 242.00 (
***
) 15.85 (
***
) 3.44 (
*
)
(
***
), (
**
), and (
*
) indicate signi?cance levels of 1%, 5%, and 10%.
a
Variable de?nitions in Appendix C. The table cells in Panel A contain, for each of the experimental cells, the means and standard
deviation for the variables. Standard deviations are in parentheses.
b
Panel B presents the results of eight ANOVA analyses. Cost information and power are the between-subject factors. The ANOVA on
the buyer’s individual pro?t tests for the main e?ect of cost information (H1a), the main e?ect of power (H2a), and the moderation e?ect
(H3a) on buyer’s individual pro?t. The ANOVA’s on the supplier’s individual pro?t and on the joint pro?t are supplementary analyses to
provide a deeper understanding of the study’s main ?ndings. The last ?ve ANOVA’s are tests for the main e?ect of cost information (H1b),
the main e?ect of power (H2b), and the moderation e?ect (H3b) on the buyer’s use of problem solving and distributive bargaining
techniques. The buyer’s use of problem solving techniques is assessed through the buyer’s information exchange, integrative behavior, and
requests for cooperation. The buyer’s use of distributive bargaining techniques is assessed through the buyer’s distributive behavior. The last
measure, problem solving approach, provides an overall assessment of the buyer’s negotiation behavior as assessed by the supplier.
Reported are the F-statistics.
A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266 255
e?ect for power (F(1, 100) = 4.16, p < 0.05): more
powerful buyers disclosed less information than less
powerful buyers. The main e?ects of power on inte-
grative behavior (F(1, 100) = 4.13, p < 0.05) and on
the number of explicit requests for cooperation
(F(1, 100) = 6.37, p < 0.05) indicate that more pow-
erful buyers used fewer integrative bargaining tech-
niques and formulated fewer requests for
cooperation than buyers without an outside option.
We also ?nd a strong main e?ect of power on dis-
tributive behavior (F(1, 100) = 14.20, p < 0.01), indi-
cating that power signi?cantly increased the use of
distributive bargaining techniques. Finally, buyers’
overall negotiation behavior is examined by means
of the composite measure problem solving approach.
A signi?cant main e?ect is found for power (F(1,
100) = 15.85, p < 0.01). Thus, according to the sup-
pliers, less powerful buyers adopted problem solving
strategies more frequently than buyers with an out-
side option. Overall, these results provide strong
support for H2b.
Finally, with respect to H3b, we examine whether
cost information moderates the overall e?ect of
power on negotiation behavior. Although the
ANOVA on information exchange indicates that
the moderation e?ect of cost information on power
is not signi?cant at the 10%-level, pairwise compar-
isons with Tukey tests indicate that buyers with
TCO information and no outside option exchanged
information more frequently than buyers with TCO
information and an outside option (p < 0.10). We
?nd that cost information has a signi?cant modera-
tion e?ect on the relationship between power and
integrative behavior (F(1, 100) = 5.02, p < 0.05).
Pairwise comparisons with Tukey tests indicate that
buyers with TCO information and no outside
option used integrative bargaining techniques more
frequently than buyers in the three other conditions
(p < 0.10). Cost information also moderates the
e?ect of power on the number of explicit requests
for cooperation (F(1, 100) = 5.73, p < 0.05). Pairwise
comparisons with Tukey tests indicate that buyers
with TCO information and no outside option
requested active cooperation more frequently than
buyers in the three other conditions (p < 0.01).
Additionally, we ?nd a moderation e?ect of cost
information on the overall e?ect of power on distrib-
utive behavior (F(1, 100) = 4.89, p < 0.05). Pairwise
comparisons with Tukey tests reveal that buyers
with TCO information and an outside option used
signi?cantly more distributive bargaining tech-
niques than buyers in the other three experimental
conditions (p < 0.01). Finally, we also ?nd a signi?-
cant moderation e?ect on problem solving approach
(F(1, 100) = 3.44, p < 0.10). Pairwise comparisons
with Tukey tests indicate that, according to the sup-
pliers, buyers with TCO information and no outside
option made more extensive use of problem solving
techniques than buyers in the other conditions
(p < 0.01). Overall, these results on the moderating
e?ect of cost information on the relation between
power and negotiation behavior, provide support
for H3b.
The mediated moderation e?ect of cost information
on the relation between power and individual pro?t
Based on exchange theory, we hypothesize in
H3c that the negotiation process has a mediation
e?ect on the negotiation outcome. More speci?cally,
we expect the nature of negotiation behavior to
mediate the moderation e?ect of cost information
on the overall e?ect that power has on individual
pro?t (cf. Fig. 1). Since the e?ect of power on nego-
tiation outcomes is moderated by cost information
and since we expect this moderation e?ect to be
mediated by negotiation behavior, we use Baron
and Kenny’s (1986) framework for combining mod-
eration and mediation. We perform the analyses for
the two main types of bargaining behavior: problem
solving behavior (operationalized as information
exchange) and distributive behavior.
11
In a sensitiv-
ity analysis, we also conduct the analysis for the
measure of the overall problem solving approach
of the buyers as assessed by the supplier.
11
Baron and Kenny (1986) procedure to test for mediation and
moderation is designed to test the e?ects of a single moderator
and a single mediator. In the present analysis, however, we are
interested in the mediating e?ects of problem solving as well as
distributive bargaining. Accordingly, we perform the procedures
for the di?erent mediating variables separately. Table 2 intro-
duces (in Step 3) information exchange and its interaction term
with cost information, whereas Table 3 introduces (again in Step
3) distributive behavior and its interaction e?ect with power. The
mediation–moderation procedure is not performed for the
combination of distributive behavior and cost information,
because we only expect an e?ect of TCO on problem solving
behavior and not on distributive behavior (cf. H1b). Because we
expect power to a?ect problem solving (cf. H2b), we also
performed the mediation–moderation procedure for the combi-
nation of information exchange and power. The results are not
signi?cant and are therefore not reported. To test for overall
mediating e?ects, we conduct the analyses reported in Table 5.
Although this overall test provides an interesting addition to the
three-step procedures listed in Tables 2–4, it is not su?cient in
itself to demonstrate mediated moderation (Muller et al., 2005).
256 A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266
First, we consider the e?ect of information
exchange. As shown in Table 2, this analysis pro-
ceeds in three steps. The ?rst step is a regression
analysis of power, cost information, and the mod-
eration e?ect of cost information on the e?ect of
power on individual pro?t. In the second step,
two equations are estimated: (1) information
exchange is regressed on power, on cost informa-
tion, and on the moderation e?ect of cost informa-
tion on the e?ect of power, and (2) individual
pro?t is regressed on power, on cost information,
on the moderation e?ect of cost information on
the e?ect of power, and on information exchange.
Eventually, in the third step, one equation is esti-
mated: individual pro?t is regressed on power, on
cost information, on the moderation e?ect of cost
information on the e?ect of power, on information
exchange, and on the moderation e?ect of cost
information on information exchange. This last
equation is identical to the second Step 2 equation,
except that the moderation term of cost informa-
tion on information exchange is now added. The
key question is the extent to which the moderation
e?ect of cost information on the relation between
power and individual pro?t is reduced in moving
from Step 2 to Step 3. Information exchange medi-
ates the moderation e?ect of cost information on
power if the following conditions are met (Baron
& Kenny, 1986, p. 1179): (1) the moderation e?ect
of cost information on power should a?ect individ-
ual pro?t less at Step 3 than at Step 2; (2) in Step
3, the moderation e?ect of cost information on
information exchange should a?ect individual
pro?t signi?cantly; (3) ?nally, in Step 2 cost infor-
mation should a?ect information exchange, which
results in a correlation between the moderation
e?ect of cost information on the e?ect of power
(i.e., power
*
cost information) and the moderation
e?ect of cost information on the e?ect of informa-
tion exchange (i.e., cost information
*
information
exchange). As can been seen in Table 2, all these
conditions are met. In Step 2, cost information sig-
ni?cantly explains information exchange (coe?-
cient = 0.41, p < 0.05), and in Step 3 the
moderation e?ect of cost information on informa-
tion exchange signi?cantly a?ects individual pro?t
(coe?cient = 651.51, p < 0.01). Furthermore, the
moderation e?ect of cost information on the e?ect
of power on individual pro?t is reduced in moving
from Step 2 to Step 3, where it has dropped to a
non-signi?cant level (coe?cient in Step
1 = À375.62, p < 0.05 and coe?cient in Step
3 = À250.73, p > 0.10), indicating ‘‘full” mediated
moderation. These results imply that information
exchange mediates the moderation e?ect of cost
information on power.
Table 2
Three-step regression procedure for testing mediation and moderation of information exchange
a
Step 1 Step 2 Step 3
Buyer’s individual
pro?t
Information
exchange
Buyer’s individual
pro?t
Buyer’s individual
pro?t
Constant 181.82 (
*
) 0.32 (
***
) 165.53 (
*
) 310.99 (
***
)
Cost information 825.87 (
***
) 0.41 (
**
) 804.75 (
***
) 517.27 (
***
)
Power 516.26 (
***
) À0.09 520.73 (
***
) 480.77 (
***
)
Power
*
cost information À375.62 (
**
) À0.28 À361.45 (
**
) À250.73
Information exchange 51.19 À405.96 (
***
)
Cost information
*
information
exchange
651.51 (
***
)
R
2
(F for R
2
) 0.41 (
***
) 0.11 (
***
) 0.41 (
***
) 0.49(
***
)
(
***
), (
**
), and (
*
) indicate signi?cance levels of 1%, 5%, and 10%.
a
Variable de?nitions in Appendix C. This table presents our results on Baron and Kenny’s (1986) three-step procedure to test whether
buyers’ use of problem solving techniques (operationalized as information exchange) mediates the moderation e?ect of cost information on
the relation between power and buyers’ pro?ts (H3c). (See Table 3 for our test for distributive bargaining techniques.) The table above
presents the results of four regression analyses. In Step 1, the dependent variable buyer’s individual pro?t is regressed on cost information,
on power, and on the moderation term power
*
cost information. In Step 2, the dependent variable information exchange is ?rst regressed
on cost information, on power, and on the moderation term power
*
cost information. Then, the dependent variable buyer’s individual
pro?t is regressed on cost information, on power, on the moderation term power
*
cost information, and on information exchange. In Step
3, the dependent variable buyer’s individual pro?t is regressed on cost information, on power, on the moderation term power
*
cost
information, on information exchange, and on the moderation term cost information
*
information exchange. Regression coe?cients are
reported.
A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266 257
The same three-step procedure is undertaken for
distributive bargaining behavior (cf. Table 3). As
shown by the results, neither cost information nor
power has a signi?cant e?ect on distributive behav-
ior in Step 2. However, mediated moderation is also
demonstrated (1) if the moderation e?ect of cost
information on power has less of an e?ect on indi-
vidual pro?t in Step 3 than in Step 2, (2) if the e?ect
of power on distributive behavior is moderated by
cost information in Step 2, and (3) if the e?ect of
distributive behavior on individual pro?t is non-
zero (Muller et al., 2005). As can been seen in Table
3, all these conditions are met. In Step 2 the moder-
ation e?ect signi?cantly explains distributive behav-
ior (coe?cient = 0.37, p < 0.05), and in Step 3
distributive behavior signi?cantly a?ects individual
pro?t (coe?cient = 942.23, p < 0.10). Furthermore,
the moderation e?ect of cost information on the
relation between power and individual pro?t is
reduced in moving from Step 2 to Step 3 (though
it does not drop to non-signi?cance, providing evi-
dence for ‘‘partial” mediated moderation). Finally,
in a sensitivity analysis, the three-step procedure is
also applied to determine the buyer’s overall prob-
lem solving approach as assessed by the supplier
(cf. Table 4). The results are very similar to those
in Table 2. Overall, we conclude that the results
reported in Tables 2 and 4 provide support for
our expectation that the moderation e?ect of cost
information on the relation between power and buy-
ers’ pro?ts is mediated by buyers’ problem solving
behavior. Table 3 provides support that the moder-
ation e?ect is partially mediated by the buyers’ dis-
tributive bargaining behavior. Together, these
results support H3c.
In a last set of analyses, we test for the overall
mediation e?ect of the negotiation process (cf. Table
5).
12,13
We conduct a hierarchical regression analy-
sis, in which dummy variables representing the
manipulations are entered in Step 1 and the pro-
posed sets of mediators are entered in Step 2 (cf.
Weingart et al., 1996). Results for regression (1)
are very similar to those obtained from the ANOVA
Table 3
Three-step regression procedure for testing mediation and moderation of distributive behavior
a
Step 1 Step 2 Step 3
Buyer’s individual pro?t Distributive behavior Buyer’s individual pro?t Buyer’s individual pro?t
Constant 181.82 (
*
) 0.05 179.73 (
*
) 138.99
Cost information 825.87 (
***
) 0.02 825.13 (
***
) 810.72 (
***
)
Power 516.26 (
***
) 0.13 510.21 (
***
) 557.21 (
***
)
Power
*
cost information À375.62 (
**
) 0.37 (
**
) À392.81 (
**
) À364.60 (
**
)
Distributive behavior 46.00 942.23 (
*
)
Power
*
distributive behavior À931.62 (
*
)
R
2
(F for R
2
) 0.41 (
***
) 0.21 (
***
) 0.41 (
***
) 0.42 (
***
)
(
***
), (
**
), and (
*
) indicate signi?cance levels of 1%, 5%, and 10%.
a
Variable de?nitions in Appendix C. This table presents our results on Baron and Kenny’s (1986) three-step procedure to test whether
buyers’ use of distributive bargaining techniques mediates the moderation e?ect of cost information on the relation between power and
buyers’ pro?ts (H3c). (See Table 2 for our tests for problem solving techniques.) The table above presents the results of four regression
analyses. In Step 1, the dependent variable buyer’s individual pro?t is regressed on cost information, on power, and on the moderation
term power
*
cost information. In Step 2, the dependent variable distributive behavior is ?rst regressed on cost information, on power, and
on the moderation term power
*
cost information. Then, the dependent variable buyer’s individual pro?t is regressed on cost information,
on power, on the moderation term power
*
cost information, and on distributive behavior. In Step 3, the dependent variable buyer’s
individual pro?t is regressed on cost information, on power, on the moderation term power
*
cost information, on distributive behavior,
and on the moderation term power
*
distributive behavior. Regression coe?cients are reported.
12
The process variables in Table 5 are information exchange
and distributive behavior (and their interaction with cost infor-
mation and power, respectively). Replacing the information
exchange variable by the problem solving behavior variable yields
similar results.
13
Finally, we perform several sensitivity analyses to test the
robustness of the results against di?erent ways of measuring
negotiation behavior. First, similar results are obtained when
information exchange is coded ‘‘0” for participants not revealing
any information about their cost structure, and ‘‘1” for partic-
ipants revealing the relative importance of each of the three issues
(with or without the numerical values). Second, we analyze
participants’ ratings of their own bargaining strategies. The self-
assessed problem solving behavior of the buyers correlates well
with suppliers’ assessment of buyers’ problem solving behavior
(Pearson correlation = 0.41, p < 0.01). However, as one of the
four items (i.e., the question whether the participant was honest
or deceptive) assessed by the buyers does not load with the other
items on one factor, the problem solving construct is based on the
supplier’s assessment of the buyer’s behavior. Similar results are
obtained and hence the conclusions remain the same when the
problem solving approach construct combines the items from
both the suppliers’ and the buyers’ questionnaires.
258 A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266
analysis: both main e?ects and the moderation e?ect
are signi?cant. The R
2
of the model is 0.41 and the
F-test is highly signi?cant (p < 0.01). When the pro-
cess variables are added to the equation, the moder-
ation e?ect of cost information on the e?ect of
power on individual pro?t drops to a non-signi?-
cant level (coe?cient = À233.69, p > 0.10) and all
process variables reach signi?cance. Information
exchange has a negative impact on individual pro?t
(coe?cient = À467.70, p < 0.01). However, the
positive and signi?cant moderation e?ect of cost
information on the e?ect of information exchange
Table 4
Three-step regression procedure for testing mediation and moderation of problem solving approach
a
Step 1 Step 2 Step 3
Buyer’s
individual pro?t
Problem
solving approach
Buyer’s
individual pro?t
Buyer’s
individual pro?t
Constant 181.82 (
*
) 1.71 (
***
) 174.60 1012.78 (
***
)
Cost information 825.87 (
***
) 1.91 (
***
) 817.78 (
***
) À666.23
Power 516.26 (
***
) À1.45 517.25 (
***
) 402.48 (
***
)
Power
*
cost information À375.62 (
**
) À1.85 (
*
) À373.89 (
**
) À144.73
Problem solving approach 4.24 À487.50 (
***
)
Cost information
*
problem solving approach 670.37 (
***
)
R
2
(F for R
2
) 0.41 (
***
) 0.72 (
***
) 0.41 (
***
) 0.49 (
***
)
(
***
), (
**
), and (
*
) indicate signi?cance levels of 1%, 5%, and 10%.
a
Variable de?nitions in Appendix C. This table reports a sensitivity analysis using Baron and Kenny’s (1986) three-step procedure to
test whether buyers’ overall problem solving approach (as assessed by the suppliers) mediates the moderation e?ect of cost information on
the relation between power and buyers’ pro?ts (H3c). The table presents the results of four regression analyses. In Step 1, the dependent
variable buyer’s individual pro?t is regressed on cost information, on power, and on the moderation term power
*
cost information. In Step
2, the dependent variable problem-solving approach is ?rst regressed on cost information, on power, and on the moderation term
power
*
cost information. Then, the dependent variable buyer’s individual pro?t is regressed on cost information, on power, on the
moderation term power
*
cost information, and on problem solving approach. In Step 3, the dependent variable buyer’s individual pro?t is
regressed on cost information, on power, on the moderation term power
*
cost information, on problem solving approach, and on the
moderation term cost information
*
problem solving approach. Regression coe?cients are reported.
Table 5
Hierarchical regression results for buyer’s individual pro?t
a
Equation 1: situational
characteristics
Equations 2: situational characteristics
and negotiation process
Constant 181.82 (
*
) 276.58 (
***
)
Cost information 825.87 (
***
) 488.09 (
***
)
Power 516.26 (
***
) 528.19 (
***
)
Power
*
cost information À375.62 (
**
) À233.69
Information exchange À467.67 (
***
)
Cost information
*
information exchange 700.11 (
***
)
Distributive behavior 1188.89 (
**
)
Power
*
distributive behavior À1181.94 (
**
)
R
2
0.41 0.52
F for R
2
22.66 (
***
) 14.84 (
***
)
Change in R
2
0.11
F for change in R
2
5.76 (
***
)
(
***
), (
**
), and (
*
) indicate signi?cance levels of 1%, 5%, and 10%.
a
Variable de?nitions in Appendix C. This table reports our results of a test for the overall mediating e?ects of negotiation behavior on
buyer’s individual pro?t. Although this overall test provides an interesting addition to the three-step procedures listed in Tables 2–4, it is
not su?cient in itself to demonstrate mediated moderation (Muller et al., 2005). The table presents the results of two regression analyses.
First, the dependent variable buyer’s individual pro?t is regressed on cost information, on power, and on the moderation term power
*
cost
information. Second, the dependent variable buyer’s individual pro?t is regressed on cost information, on power, on the moderation term
power
*
cost information, on information exchange, on the moderation term cost information
*
information exchange, on distributive
behavior, and on the moderation term power
*
distributive behavior. Regression coe?cients are reported.
A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266 259
(coe?cient = 700.11, p < 0.01) implies that this neg-
ative relation between information exchange and
individual pro?t is only true for buyers lacking
TCO information. Buyers who have access to
TCO information and who exchange this informa-
tion earn signi?cantly higher individual pro?ts. Dis-
tributive behavior has a positive e?ect on individual
pro?t (coe?cient = 1188.89, p < 0.05), but this is
not the case for buyers with power (coe?-
cient = À1181.94, p < 0.05). In addition, the mod-
el’s R
2
increases signi?cantly, from 0.41 to 0.52 (F
for change in R
2
= 5.76, p < 0.01). Together, these
results provide support for our expectation that
the moderation e?ect on the outcome variable is
mediated by the process variables. As shown by
the signi?cant moderation e?ect in equation 1, the
performance disadvantage of less powerful buyers
is less pronounced when the buyer has detailed
TCO information. We also expect this moderation
e?ect to be mediated by the process variables: med-
iated moderation implies that the overall modera-
tion e?ect is reduced once the mediating process is
taken into account (Muller et al., 2005). Equation
2 clearly indicates that the moderation e?ect of cost
information and power becomes insigni?cant when
the process variables are added.
Supplementary analyses of the supplier’s negotiation
outcome and behavior
Table 1 also reports the results of the analyses of
the supplier’s individual pro?t and of the joint
pro?t. An ANOVA on the supplier’s individual pro?t
reveals main e?ects for power and cost information.
Suppliers facing less powerful buyers earned higher
individual pro?ts (F(1, 100) = 17.46, p < 0.01).
However, supplier pro?ts were lower when they
negotiated with buyers with TCO information
(F(1, 100) = 15.80, p < 0.01). Joint pro?ts were sig-
ni?cantly higher when the buyer possessed TCO
information (F(1, 100) = 15.78, p < 0.01). Together,
these results indicate that buyers with TCO infor-
mation and no outside option obtained high indi-
vidual pro?ts not because they knew ‘‘how to
fool” the (more powerful) suppliers, but because
these dyads realized higher joint pro?ts than dyads
in which the buyer had an outside option (mean
joint pro?t = 2446.15 versus mean joint
pro?t = 2336.67). Suppliers facing buyers without
an outside option earned more than their less pow-
erful opponents (buyer’s mean individual
pro?t = 1007.69 versus supplier’s mean individual
pro?t = 1438.46), re?ecting the power imbalance.
However, less powerful buyers with TCO informa-
tion were able to earn similar individual pro?ts as
buyers with an outside option and TCO informa-
tion, a result that we explained through the detailed
analysis of the buyers’ negotiation behavior.
Finally, in order to assess the e?ects of buyers’
behavior on suppliers’ behavior, we present results
(not tabulated) on the suppliers’ bargaining behav-
ior. Recall that the experimental manipulation for
the suppliers was the same in each of the experimen-
tal conditions: suppliers always had an outside
option and full cost information. Accordingly, the
di?erences in suppliers’ bargaining behavior across
the di?erent experimental cells can be explained by
their interaction with buyers. An ANOVA on the
suppliers’ information exchange shows a main e?ect
for cost information (F(1, 100) = 6.92, p < 0.01):
suppliers facing buyers with TCO information dis-
closed more information than suppliers facing buy-
ers with traditional information. We ?nd no e?ect of
the manipulations with respect to the buyers’ condi-
tions on the suppliers’ integrative behavior, nor on
the number of explicit requests for cooperation
issued by suppliers. Furthermore, we observe that
suppliers facing a powerful buyer used distributive
behavior more frequently than suppliers in the other
experimental conditions (F(1, 100) = 0.58, p < 0.10).
More speci?cally, we ?nd that suppliers facing a
powerful buyer with TCO information used more
distributive threats and referred to their outside
option more often than suppliers in the other exper-
imental conditions. Suppliers’ negotiation behavior
is also assessed by the composite measure problem
solving approach (based on the post-bargaining
questionnaires of the buyers). Main e?ects are
found for cost information (F(1, 100) = 4.83,
p < 0.05) and power (F(1, 100) = 3.18, p < 0.10).
Thus, compared to buyers with traditional cost
information, buyers with TCO information
reported that suppliers used more problem solving
techniques and fewer distributive bargaining tech-
niques. Furthermore, less powerful buyers felt that
their opponent employed more problem solving
techniques than buyers with an outside option.
Overall, these results provide evidence that buyers’
behavior strongly a?ects suppliers’ behavior.
Discussion and conclusion
In this paper, we examine the moderation e?ect
of cost information (TCO information versus
260 A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266
traditional cost information) on a buyer’s individual
pro?t at both high and low levels of buyer power.
Our results indicate that the availability of detailed
TCO information may alleviate the disadvantage
that dependent buyers face vis-a`-vis a more power-
ful supplier. This ?nding has important managerial
implications. On the one hand, it implies that less
powerful buyers can compensate for their power
disadvantage by gathering more detailed cost infor-
mation. On the other hand, powerful buyers do not
seem to be able to use this more detailed cost infor-
mation to enhance their power advantage in order
to obtain even higher individual pro?ts.
We explore the driver behind this result by exam-
ining the negotiation process. Consistent with
exchange theory and recent literature on informa-
tion processing, we expect that buyers with detailed
cost information and less power than their oppo-
nent try to (re)gain control over their own outcomes
by sharing information in an e?ort to create integra-
tive solutions. Because of their higher epistemic
motivation, less powerful buyers have a greater
incentive to engage in information acquisition and
processing (e.g., Fiske, 1993; Van Kleef et al.,
2006). Accordingly, we expect that less powerful
buyers with TCO information seek a more integra-
tive bargaining situation than powerful buyers,
who are more prone to using distributive bargaining
techniques because of their low epistemic motiva-
tion (e.g., De Dreu & Van Kleef, 2004; Fiske,
1993; Weingart et al., 1996).
Support for these conjectures is found in our fol-
low-up analyses, in which we examine whether the
moderation e?ect of TCO information on the rela-
tion between power and individual pro?t can be
explained by the choice of negotiation strategy.
Using Baron and Kenny’s (1986) framework for
combining mediation and moderation, we ?nd that
the moderation e?ect of cost information on the
relation between power and individual pro?t is med-
iated by buyers’ bargaining behavior. Analysis of
the buyers’ communications suggests that buyers
with TCO information disclose more information
than buyers with traditional cost information. Con-
sistent with prior research (e.g., Perdue & Summer,
1991), we ?nd that more powerful buyers disclose
less information and use more distributive bargain-
ing techniques than less powerful buyers. The
results also reveal an unexpected positive e?ect of
TCO information on distributive behavior. How-
ever, this e?ect can be explained by the ?nding that
buyers with TCO information and power tend to
use distributive techniques. Furthermore, a limited
analysis of suppliers’ bargaining behavior revealed
its reciprocal nature. Suppliers facing buyers with
TCO information disclosed more information than
suppliers facing buyers with traditional information.
It also emerged that suppliers facing powerful buy-
ers with TCO information used more distributive
bargaining techniques than suppliers in the other
experimental conditions. As the experimental
manipulation for the suppliers is the same in each
of the experimental conditions, the di?erences in
suppliers’ bargaining behavior across the di?erent
experimental cells can only be explained by their
interaction with the buyers. These ?ndings provide
support for our conjecture that less powerful buyers
are able to create a cooperative relationship, into
which the supplier is willing to enter. This resulted
not only in higher individual pro?ts, but also in
higher joint pro?ts.
From these results, we conclude that the manip-
ulation of power and cost information resulted in
buyers using di?erent negotiation techniques. Less
powerful buyers who have access to TCO data
are more likely to resort to problem solving tech-
niques, whereas powerful buyers tend to rely on
distributive bargaining techniques. Particularly
interesting is that the problem solving strategy
adopted by less powerful buyers with TCO infor-
mation seems to be e?ective, whereas the distribu-
tive bargaining strategy adopted by more
powerful buyers with TCO results in less informa-
tion sharing and in less e?ective negotiation out-
comes. Less powerful buyers seem to be
motivated to create a cooperative and coordinated
relationship, in which the supplier is willing to con-
sider the buyer’s objectives. When a buyer shares
information about needs and preferences and/or
makes concessions that facilitate the development
of a solution, the supplier is likely to reciprocate
(e.g., Campbell et al., 1988; Dekker, 2004). This
may explain the higher individual pro?ts for less
powerful buyers with TCO information. Less pow-
erful buyers lacking TCO information are less able
or willing to communicate the right information
and therefore to create integrative situations. More
powerful buyers, on the other hand, believe to have
power and choose a distributive bargaining strategy
that, in a tit-for-tat move, the supplier responds to
by also resorting to a distributive bargaining strat-
egy. Thus, a powerful buyer’s distributive bargain-
ing strategy may not be e?ective when he is facing
an equally powerful supplier.
A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266 261
Overall, our results suggest that powerful buyers
with re?ned TCO information may not maximize
possible bene?ts from buyer–supplier interactions
due to their bargaining strategy. Their (false) feeling
of power causes an increase in distributive bargain-
ing techniques and a reluctance to share the infor-
mation necessary for inter?rm cost minimization.
This implies that powerful buyers may only bene?t
from more re?ned accounting information systems
if these ?rms undertake e?orts that encourage infor-
mation sharing and discourage distributive bargain-
ing strategies. Future research should examine
whether, and under what conditions, inter?rm con-
trol systems and incentive systems may motivate
buyers and suppliers to share cost information in
such a way that supply chain performance is
improved.
We conclude with remarks on several limitations
of this study and further directions for future
research. First, while the experimental context
induced by a simple negotiation game allows us to
maintain control over exogenous variables, the
scope for generalizing the results is somewhat lim-
ited. Other factors, such as the incentive system,
past negotiation history, and future negotiation
probabilities, have been shown to impact negotiated
outcomes but are not included in the present analy-
sis. Future research is needed to determine the sen-
sitivity of our results to several parameters excluded
in the current study.
Second, this study manipulates the experimental
conditions of the buyers, but not the suppliers, that
is, the suppliers’ experimental conditions were held
constant. In particular, suppliers were always fully
informed and powerful. As a consequence, our con-
clusions do not generalize beyond negotiation set-
tings in which the supplier is always at least as
powerful as the buyer. Further research can modify
these experimental conditions and examine the role
of TCO information from both buyer and supplier
perspectives.
Third, although optimal joint outcomes are
introduced, our study focuses primarily on buyer
outcomes. This focus underplays the cost/bene?t
trade-o? of obtaining the additional information
needed for TCO. Since TCO information is not
costless, it would be interesting to consider whether
buyers would be willing to incur the cost of obtain-
ing TCO information if it is e?cient to do so.
Fourth, our manipulation of TCO information is
obviously a simpli?cation of reality, providing many
avenues for further research. In this paper, we take a
deterministic approach to TCO information, which
allows us to study the impact of cost information
on negotiation behavior and outcomes. Our aim
was not to describe a value chain and ABC analysis,
nor to discuss data collection mechanisms and chal-
lenges. Neither did we seek to identify and categorize
important TCO cost drivers. However, more
research is needed on these issues to e?ectively iden-
tify critical cost drivers for estimating TCO. Future
research may consider aspects such as the manner
in which buyers collect TCOdata (e.g., fromthe sup-
plier, benchmarking, past experience), the type of
TCOsystems implemented (e.g., formal vs. informal,
standardized vs. unique models), and how these
design aspects impact buyer–supplier relations and
negotiations. In addition, examining the e?ects of
imperfect TCO information would also be interest-
ing. In this study, perfect TCO information was pro-
vided to decision-makers. In reality, TCO
information is characterized by mistakes and simpli-
?cations, which may have important implications on
the negotiation process and its outcomes. Similarly,
it would be interesting to consider how costs that are
di?cult to measure (e.g., opportunity costs of
reduced sales, productivity losses due to downtime)
can be included in TCO systems. Because of the dif-
?culty of measuring these costs reliably, they are typ-
ically not recorded in a buyer’s accounting system.
However, they may represent a substantial cost in
total costs. An investigation of how the negotiation
process and outcomes are a?ected when less objec-
tive or less reliable cost estimates are explicitly
included in the TCO system would add useful
insights to the literature. Finally, further research
could consider whether audited TCO information
(either by a third party or by the partner) would be
considered more reliable than unaudited TCO infor-
mation, and if so, what the implications for the
buyer–supplier relation would be.
Acknowledgements
The comments of Shannon Anderson, Eddy Car-
dinaels, Henri Dekker, Luc Sels, the participants at
the BAA conference in Edinburgh, the EAA confer-
ence in Gothenburg, the AOS-Bocconi conference
in Milan, the MAS conference in Clearwater Beach,
and two anonymous reviewers are gratefully
acknowledged, as is the research funding provided
by the Fund for Scienti?c Research – Flanders (Pro-
ject G.0261.00) and by the National Bank of Bel-
gium (Project NB/05/01).
262 A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266
Appendix A. Experimental cells and corresponding cost tables
Power
Low power (no outside option) Equal power (outside option)
Cost information
TCO information Cell 1 (n = 30) Cell 2 (n = 26)
Traditional cost information Cell 3 (n = 26) Cell 4 (n = 22)
Cost table for the supplier (in cell 1, 2, 3 and 4)
a
Price (=income) Maintenance Spare parts
Contract 1 600 Contract A
*
1350 Contract R 2250
Contract 2 1200 Contract B 1200 Contract S 2000
Contract 3 1800 Contract C 1050 Contract T 1750
Contract 4 2400 Contract D 900 Contract U 1500
Contract 5
*
3000 Contract E 750 Contract V 1250
Contract 6 3600 Contract F 600 Contract W 1000
Contract 7 4200 Contract G 450 Contract X 750
Contract 8 4800 Contract H 300 Contract Y 500
Contract 9 5400 Contract I 150 Contract Z
*
250
a
The tables were constructed such that the minimum (À3000 Euro) and maximum pro?ts (5000 Euro) that buyers and suppliers could
earn were the same. Pareto-optimal joint outcomes correspond to the buyer and the supplier agreeing to contract 5AZ. This situation is
denoted with an asterisk (
*
).
Cost table for the buyer with TCO information (in cell 1 and 2):
Income = 6000
Price (=cost) Maintenance Spare parts
Contract 1 600 Contract A
*
250 Contract R 150
Contract 2 1200 Contract B 500 Contract S 300
Contract 3 1800 Contract C 750 Contract T 450
Contract 4 2400 Contract D 1000 Contract U 600
Contract 5
*
3000 Contract E 1250 Contract V 750
Contract 6 3600 Contract F 1500 Contract W 900
Contract 7 4200 Contract G 1750 Contract X 1050
Contract 8 4800 Contract H 2000 Contract Y 1200
Contract 9 5400 Contract I 2250 Contract Z
*
1350
Cost table for the buyer with traditional cost information (in cell 3 and 4):
Income = 6000
Price (=cost) Maintenance
a
Spare parts
b
Contract 1 600 Contract A
*
1 Contract R 3
Contract 2 1200 Contract B 2 Contract S 6
Contract 3 1800 Contract C 3 Contract T 9
Contract 4 2400 Contract D 4 Contract U 12
Contract 5
*
3000 Contract E 5 Contract V 15
Contract 6 3600 Contract F 6 Contract W 18
Contract 7 4200 Contract G 7 Contract X 21
Contract 8 4800 Contract H 8 Contract Y 24
Contract 9 5400 Contract I 9 Contract Z
*
27
a
Number of maintenance sessions performed by the buyer each month.
b
Spare parts procured by the buyer from a third party each month.
A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266 263
Appendix B. Measuring negotiation behavior
B.1. Behavioral coding categories
Category Examples
Information exchange – Maintenance is more expensive for my company than spare parts
– Maintenance contracts start at €250 (=contract A) and increase by €250 until €2250 (=contract I); con-
tracts for spare parts start at €150 (= contract R) and increase by €150 until €1350 (=contract Z)
Rewards – I am pleased with the concessions made thus far
Positive normative appeals – Your o?ers have been fair and equitable
Request for cooperation – Let us cooperate
General threats – Make a concession or you will be in trouble
Exit threats – Respond with a concession or I will call another supplier
Punishment – This negotiation is going nowhere
Warnings – My company has a policy against uncooperative suppliers
Positional commitment – I refuse to concede any further
– I refuse to drop my price below the present level
B.2. Buyers’ problem solving approach
Observed ratings from suppliers’ questionnaires (items were reverse coded)
Do you feel that the person with whom you were paired was more interested in solving your mutual
problem, or more self-interested?
1 2 3 4 5
Solving a mutual problem Self-interested
Rate your partner’s bargaining strategies on the following scales:
1 2 3 4 5
Accommodating Exploitative
1 2 3 4 5
Honest Deceptive
1 2 3 4 5
Information sharing Withholding
information
Appendix C. Bivariate correlation matrix (n = 104)
1 2 3 4 5 6 7 8 9 10
1 Buyer’s individual pro?t
a
1
2 Supplier’s individual pro?t
b
À0.71 (
**
) 1
3 Joint pro?t
c
0.54 (
**
) 0.21 (
*
) 1
4 Information exchange
d
0.14 0.19 0.41 (
**
) 1
5 Integrative behavior
e
0.06 À0.02 0.05 0.02 1
6 Requests for cooperation
f
0.15 0.05 0.27 (
**
) 0.30 (
**
) 0.36 (
**
) 1
7 Distributive behavior
g
0.22 (
*
) À0.20 (
*
) 0.07 0.09 À0.11 À0.11 1
8 Problem solving approach
h
0.41 (
**
) À0.10 0.44 (
**
) 0.35 (
**
) 0.16 0.38 (
**
) 0.09 1
9 Cost information
i
0.55 (
**
) À0.34 (
**
) 0.36 (
**
) 0.23 (
*
) 0.13 0.23 (
**
) 0.23 (
*
) 0.82 (
**
) 1
10 Power
j
0.27 (
**
) À0.35 (
**
) À0.04 À0.20 (
*
) À0.21 (
*
) À0.25 (
**
) 0.35 (
**
) À0.23 (
*
) 0.00 1
(
*
) Correlation is signi?cant at the 0.05 level (2-tailed).
(
**
) Correlation is signi?cant at the 0.01 level (2-tailed).
a
Buyer’s individual pro?t at the end of the game (based on the cost table provided in Appendix A).
b
Supplier’s individual pro?t at the end of the game (based on the cost table provided in Appendix A).
c
Joint pro?t is the sum of buyer’s individual pro?t and supplier’s individual pro?t at the end of the game.
d
Information exchange is coded ‘‘0” for participants not revealing any information about their cost structure, ‘‘1” for participants
revealing the relative importance (but not the numerical values) of each of the three issues under negotiation, and ‘‘2” for participants
revealing the relative importance and the numerical values of each of the three issues under negotiation.
264 A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266
e
Coded from the messages sent by buyers (based on the behavioral coding scheme in Appendix B1). Summated scale determined by
adding up two integrative behavior types (rewards and positive normative appeals) and solving for the mean value. Cronbach’s
alpha = 0.70.
f
Coded from the messages sent by buyers (based on the behavioral coding scheme in Appendix B1). Single scale item (requests for
cooperation).
g
Coded from the messages sent by buyers (based on the behavioral coding scheme in Appendix B1). Summated scale determined by
adding ?ve distributive behavior types (general threats, exit threats, positional commitment, punishments, warnings) together and solving
for the mean value. Cronbach’s alpha = 0.81.
h
Buyer’s problem solving approach based on observed ratings from supplier’s questionnaire. Summated scale determined by adding the
four post-bargaining questionnaire items included in Appendix B2 together and solving for the mean value. Cronbach’s alpha = 0.88.
i
Dummy variable, experimental manipulation: 0 for traditional cost information, 1 for TCO information.
j
Dummy variable, experimental manipulation: 0 for buyers with no outside option, 1 for buyers with an outside option.
References
Anderson, S. W., & Dekker, H. C. (2005). Management
control for market transactions: The relation between
transaction characteristics, incomplete contract design and
subsequent performance. Management Science, 51,
1734–1752.
Baiman, S., & Rajan, M. V. (2002). Incentive issues in inter-?rm
relationships. Accounting, Organizations and Society, 27,
213–238.
Bame-Aldred, C. W., & Kida, T. (2007). A comparison of auditor
and client initial negotiation positions and tactics. Accounting,
Organizations and Society, 32, 497–511.
Baron, R., & Kenny, D. (1986). The moderator–mediator
variable distinction in social psychological research: Concep-
tual, strategic, and statistical considerations. Journal of
Personality and Social Psychology, 51, 1173–1182.
Bazerman, M. H., Curhan, J. R., Moore, D. A., & Valley, K. L.
(2000). Negotiation. Annual Review of Psychology, 51,
279–314.
Bottom, W. P., Holloway, J., Miller, G. J., Mislin, A., &
Whitford, A. (2006). Building a pathway to cooperation:
Negotiation and social exchange between principal and agent.
Administrative Science Quarterly, 51, 29–58.
Campbell, N. C. G., Graham, J. L., Jolibert, A., & Meissner, H.
G. (1988). Marketing negotiations in France, Germany, the
United Kingdom, and the United States. Journal of Market-
ing, 52, 49–62.
Carr, L. P., & Ittner, C. D. (1992). Measuring the cost of
ownership. Journal of Cost Management, 6, 7–13.
Craft, J. A. (1981). Information disclosure and the role of the
accountant in collective bargaining. Accounting, Organiza-
tions and Society, 6, 97–107.
De Dreu, C. K. W., Beersma, B., Stroebe, K., & Euwema, M. C.
(2006). Motivated information processing, strategic choice,
and the quality of negotiated agreement. Journal of Person-
ality and Social Psychology, 90, 927–943.
De Dreu, C. K. W., & Carnevale, P. (2003). Motivational bases
of information processing and strategy in con?ict and
negotiation. Advances in Experimental Social Psychology,
35, 235–291.
De Dreu, C. K. W., & Van Kleef, G. A. (2004). The in?uence of
power on the information search, impression formation, and
demands in negotiation. Journal of Experimental Social
Psychology, 40, 303–319.
Degraeve, Z., & Roodhooft, F. (1999). E?ectively selecting
suppliers using total cost of ownership. Journal of Supply
Chain Management, 35, 5–10.
Dekker, H. C. (2003). Value chain analysis in inter?rm relation-
ships: A ?eld study. Management Accounting Research, 14,
1–23.
Dekker, H. C. (2004). Control of inter-organizational relation-
ships: Evidence on appropriation concerns and coordination
requirements. Accounting, Organizations and Society, 29,
27–49.
Drake, A., & Haka, S. (2008). Does ABC exacerbate the hold-up
problem between buyers and suppliers?. The Accounting
Review, 83, 29–60.
Ellram, L. M. (1995). Activity based costing and total cost of
ownership: A critical linkage. Journal of Cost Management, 9,
22–30.
Emerson, R. M. (1976). Social exchange theory. Annual Review of
Sociology, 2, 335–362.
Fisher, J., Frederickson, J. R., & Pe?er, S. A. (2000). Budgeting:
An experimental investigation of the e?ects of negotiation.
The Accounting Review, 75, 93–114.
Fisher, J., Frederickson, J. R., & Pe?er, S. A. (2006). Budget
negotiations in multi-period settings. Accounting, Organiza-
tions and Society, 31, 511–528.
Fiske, S. T. (1993). Controlling other people: The impact of
power on stereotyping. American Psychologist, 48, 621–628.
Gelfand, M., & Christakopoulou, S. (1999). Culture and nego-
tiation cognition: Judgment accuracy and negotiation pro-
cesses in individualistic and collectivistic cultures.
Organizational Behavior and Human Decision Processes, 79,
248–269.
Gibbins, M., McCracken, S. A., & Salterio, S. E. (2005).
Negotiations over accounting issues: The congruency of audit
partner and chief ?nancial o?cer recalls. Auditing: A Journal
of Practice and Theory, 24, 171–193.
Gibbins, M., Salterio, S. E., & Webb, A. (2001). Evidence about
auditor–client management negotiation concerning client’s
?nancial reporting. Journal of Accounting Research, 39,
553–563.
Giebels, E., De Dreu, C. K. W., & Van de Vliert, E. (2000).
Interdependence in negotiation: E?ects of exit options and
social motive on distributive and integrative negotiation.
European Journal of Social Psychology, 30, 255–272.
Goodwin, S. A., Gubin, A., Fiske, S. T., & Yzerbyt, V. Y. (2000).
Power can bias impression processes: Stereotyping subordi-
A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266 265
nates by default and by design. Group Processes and
Intergroup Relations, 3, 227–256.
Ittner, C. D., Larcker, D. F., & Meyer, M. W. (2003). Subjectivity
and the weighting of performance measures: Evidence from
the balanced scorecard. The Accounting Review, 78, 725–758.
Kachelmeier, S. J., & Towry, K. L. (2002). Negotiated transfer
pricing: Is fairness easier said than done?. The Accounting
Review, 77, 571–593.
Kadous, K., Koonce, L., & Towry, K. L. (2005). Quanti?cation
and persuasion in managerial judgment. Contemporary
Accounting Research, 22, 643–686.
Kelley, H. H. (1966). A classroom study of the dilemmas in
interpersonal negotiation. In K. Archibald (Ed.), Strategic
interaction and con?ict. Berkerley, California: Institute of
International Studies.
Kelley, H. H., & Thibaut, J. (1978). Interpersonal relations: A
theory of interdependence. New York: Wiley.
Keltner, D., Gruenfeld, D. H., & Anderson, C. (2003). Power,
approach, and inhibition. Psychological Review, 110, 265–284.
Kersten, G. (2001). Modeling distributive and integrative nego-
tiations: Reviewed and revised characterization. Group Deci-
sion and Negotiation, 10, 493–514.
Luft, J., & Libby, R. (1997). Pro?t comparisons, market prices
and managers’ judgments about negotiated transfer prices.
The Accounting Review, 72, 217–229.
Luft, J., & Shields, M. D. (2003). Mapping management
accounting: Graphics and guidelines for theory-consistent
empirical research. Accounting, Organizations and Society, 28,
169–249.
Muller, D., Judd, C. M., & Yzerbyt, V. Y. (2005). When
moderation is mediated and mediation is moderated. Journal
of Personality and Social Psychology, 89, 852–863.
Murnigham, J., & Bazerman, M. (1990). A perspective on
negotiation research in accounting and auditing. The Account-
ing Review, 65, 642–657.
Ng, T. B-P., & Tan, H-T. (2003). E?ects of authoritative guidance
availability and audit committee e?ectiveness on auditors’
judgment in an auditor–client negotiation context. The
Accounting Review, 78, 801–818.
Northcraft, G. B., & Neale, M. A. (1986). Opportunity costs and
the framing of resource allocation decisions. Organizational
Behavior and Human Decision Processes, 37, 348–356.
Perdue, B. C., & Summer, J. O. (1991). Purchasing agents’ use of
negotiation strategies. Journal of Marketing Research, 28,
175–189.
Pinkley, R. L., Neale, M. A., & Bennett, R. J. (1994). The impact
of alternatives to settlement in dyadic negotiation. Organiza-
tional Behavior and Human Decision Processes, 57, 97–116.
Porter, M. E. (1985). Competitive advantage: Creating and
sustaining superior performance. New York: The Free Press.
Putman, L. L., & Fairhurst, G. T. (2001). Discourse analysis in
organization. In F. Jablin & L. Putman (Eds.), The New
Handbook of Organizational Communication. Thousand Oaks,
CA: Sage.
Roodhooft, F., Hiel, G., Van den Abbeele, A., & van Doveren, B.
(2003). Optimised sourcing strategies using total cost of
ownership. Cost Management, 17, 28–35.
Roodhooft, F., Van den Abbeele, A., & Peeters, F. (2005).
Calculating the total cost of ownership of utilities: A case of
inter-?rm cost management. Cost Management, 19, 13–23.
Shank, J. K. (1989). Strategic cost management: New wine, or
just new bottles?. Management Accounting Research 1, 47–65.
Shank, J. K., & Govindarajan, V. (1992). Strategic cost manage-
ment: The value chain perspective. Journal of Management
Accounting Research, 4, 179–197.
Thibaut, J. W., & Kelley, H. H. (1959). The social psychology of
groups. NY: John Wiley & Sons.
Tomkins, C. (2001). Interdependencies, trust and information in
relationships, alliances and networks. Accounting, Organiza-
tions and Society, 26, 161–191.
Van Kleef, G. A., De Dreu, C. W., Pietroni, D., & Manstead, A.
S. R. (2006). Power and emotion in negotiation: Power
moderates the interpersonal e?ects of anger and happiness
and concession making. European Journal of Social Psychol-
ogy, 36, 557–581.
Walton, R. E., & McKersie, R. B. (1966). Behavioral dilemmas in
mixed-motive decision making. Behavioral Science, 11,
370–384.
Waterhouse, J., Gibbins, M., & Richardson, A. J. (1993).
Strategic ?nancial disclosure: Evidence from labor negotia-
tions. Contemporary Accounting Research, 9, 526–550.
Weingart, L. R., Hyder, E. B., & Prietula, M. J. (1996).
Knowledge matters: The e?ect of tactical descriptions on
negotiation behavior and outcome. Journal of Personality and
Social Psychology, 70, 1204–1217.
Windsor, C. A., & Ashkanasy, N. M. (1995). The e?ect of client
management bargaining power, moral reasoning develop-
ment, and belief in a just world on auditor independence.
Accounting, Organizations and Society, 20, 701–720.
Wolfe, R., & McGinn, K. L. (2005). Perceived, relative power
and its in?uence on negotiations. Group Decisions and
Negotiation, 14, 3–20.
Wouters, M., Anderson, J. C., & Wynstra, F. (2005). The
adoption of total cost of ownership for sourcing decisions – A
structural equations analysis. Accounting, Organizations and
Society, 30, 167–191.
266 A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266
doc_110177264.pdf
We investigate the influence of total cost of ownership (TCO) information on buyer–supplier negotiations in different
power settings. Based on social exchange theory and recent literature on information processing, we expect that buyers
with detailed TCO information and less power than their negotiation partners may try to (re)gain control over their
own outcomes by sharing information. The results of our experiment indicate that the performance disadvantage of less
powerful buyers is less pronounced when the buyer has detailed TCO information, whereas more powerful buyers do not
seem to be able to profit from TCO information
The e?ect of cost information on buyer–supplier negotiations
in di?erent power settings
Alexandra Van den Abbeele
a,
*
, Filip Roodhooft
a,b
, Luk Warlop
a
a
Katholieke Universiteit Leuven, Naamsestraat 69, 3000 Leuven, Belgium
b
Vlerick Leuven Gent Management School, Vlamingenstraat 83, 3000 Leuven, Belgium
Abstract
We investigate the in?uence of total cost of ownership (TCO) information on buyer–supplier negotiations in di?erent
power settings. Based on social exchange theory and recent literature on information processing, we expect that buyers
with detailed TCO information and less power than their negotiation partners may try to (re)gain control over their
own outcomes by sharing information. The results of our experiment indicate that the performance disadvantage of less
powerful buyers is less pronounced when the buyer has detailed TCO information, whereas more powerful buyers do not
seem to be able to pro?t from TCO information. These somewhat counterintuitive ?ndings are explained through detailed
analysis of the buyer’s negotiation behavior, which shows that less powerful buyers who have access to TCO data use prob-
lem solving techniques more frequently than powerful buyers, who tend to rely on distributive bargaining techniques
instead. We conclude that power can motivate a failure to share TCO information, resulting in less e?ective inter?rm nego-
tiation outcomes.
Ó 2008 Elsevier Ltd. All rights reserved.
Introduction
This study investigates whether buyers bene?t
from total cost of ownership (TCO) information
when negotiating with suppliers across di?erent
power settings. TCO can be seen as an application
of activity-based costing (ABC) to an inter?rm con-
text (Wouters, Anderson, & Wynstra, 2005). It
quanti?es the costs of the activities involved in
acquiring and using purchased goods or services.
As inter?rm relations have become increasingly
important and as purchasing professionals in many
companies still need to demonstrate the contribu-
tion they make to their ?rm, it is relevant for buyers
to identify the management accounting tools that
might contribute to their market success and pro?t-
ability. TCO is a tool that supports purchasing deci-
sions by focusing on all costs related to a purchase
rather than simply on price (Ellram, 1995). Prior lit-
erature suggests that TCO information can improve
inter?rm cooperation, as a buyer can use the quan-
ti?ed cost and cost driver information resulting
from a TCO analysis to communicate to the sup-
plier which activities cause higher costs on his side,
increasing his bargaining power and in turn supply
chain performance (e.g., Ellram, 1995; Roodhooft,
0361-3682/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.aos.2008.05.005
*
Corresponding author. Tel.: +32 16 32 69 36; fax: +32 16 32
67 32.
E-mail addresses: [email protected].
be (A. Van den Abbeele), [email protected]
(F. Roodhooft), [email protected] (L. Warlop).
Available online at www.sciencedirect.com
Accounting, Organizations and Society 34 (2009) 245–266
www.elsevier.com/locate/aos
Van den Abbeele, & Peeters, 2005). However, due to
buyers’ reluctance to share the information needed
for inter?rm cost minimization, ?rms may fail to
maximize possible gains from buyer–supplier nego-
tiations (Baiman & Rajan, 2002). Drake and Haka
(2008) show, for instance, that concerns about ineq-
uity may lead to reluctance to share detailed
accounting information. In this study, we investi-
gate whether bargaining power may prevent buyers
from sharing private TCO information and whether
this might result in less e?ective negotiation out-
comes between buyers and suppliers.
Power can be broadly de?ned as the capacity to
exert in?uence on other people (Kelley & Thibaut,
1978). Although power may derive from a variety
of ‘‘power bases”, it is the mutual dependence of
individuals that allows power di?erences to exist.
Accordingly, in inter?rm relations, power is evi-
denced as the availability of alternatives (e.g.,
Anderson & Dekker, 2005). Prior research indicates
that power di?erences in?uence information search
strategies and drive the processing of information
about other people (e.g., De Dreu & Van Kleef,
2004; Fiske, 1993). These studies suggest that power
may be pivotal in the interplay between, on the one
hand, the need to share information in order to
optimize activities across the supply chain and, on
the other hand, the reluctance to share private
information.
This paper reports the results of a buyer–supplier
negotiation experiment in which we manipulate
both the cost information and the power of the
buyer in order to assess whether access to detailed
TCO information bene?ts buyers across di?erent
power settings. Based on social exchange theory
(e.g., Emerson, 1976; Thibaut & Kelley, 1959) and
recent research on information processing in negoti-
ations (e.g., De Dreu & Carnevale, 2003), we expect
that less powerful buyers who have detailed TCO
information at their disposal may try to (re)gain
control over negotiation outcomes by sharing infor-
mation and creating integrative bargaining situa-
tions. Our results indicate that TCO information
reduces the performance disadvantage of less pow-
erful buyers, and further that this moderation e?ect
is mediated by buyers’ bargaining behavior; that is,
less powerful buyers with TCO information use
more problem solving techniques and fewer distrib-
utive bargaining techniques than more powerful
buyers or buyers that lack TCO information. These
results imply that less powerful buyers can compen-
sate for their power disadvantage by acquiring more
detailed TCO information. However, the results
also suggest that powerful buyers seem unable to
use TCO information to exploit their power advan-
tage in order to obtain even greater individual prof-
its. We therefore conclude that a position of power
may impede the sharing of TCO information, result-
ing in suboptimal negotiation outcomes between
buyers and suppliers.
This paper contributes to the existing literature
on buyer–supplier relations and the use of account-
ing information in three signi?cant ways. First, we
analyze empirically whether buyers pro?t from
TCO information when negotiating with suppliers.
The literature on TCO discusses its adoption and
its potential bene?ts (e.g., Carr & Ittner, 1992; Ell-
ram, 1995; Wouters et al., 2005). For example, these
studies have argued that TCO can be used not only
to compare and evaluate di?erent suppliers, but also
to provide decision-makers with an objective and
easily understood argument to support purchasing
decisions. It has also been suggested that a buyer
can use the cost and cost driver information result-
ing from a TCO analysis to communicate to the
supplier which activities cause higher costs on his
side, increasing his bargaining position (e.g., Ellram,
1995; Roodhooft, Hiel, Van den Abbeele, & van
Doveren, 2003, 2005). However, there is little empir-
ical research on the use of TCO during actual
buyer–supplier negotiations. To our knowledge, this
study is the ?rst to investigate whether buyers can
bene?t from TCO when negotiating with suppliers.
Our second contribution is to add to a small but
growing body of work on the role of accounting
information in negotiations. The existing research
on accounting negotiation has mainly focused on
three intra?rm issues: collective bargaining (e.g.,
Craft, 1981; Waterhouse, Gibbins, & Richardson,
1993), transfer pricing (e.g., Kachelmeier & Towry,
2002; Luft & Libby, 1997), and budgeting (e.g.,
Fisher, Frederickson, & Pe?er, 2000, 2006). The lit-
erature on inter?rm accounting negotiations has pri-
marily concentrated on auditor–client interactions
(e.g., Bame-Aldred & Kida, 2007; Gibbins, Salterio,
& Webb, 2001; Ng & Tan, 2003; Windsor & Ash-
kanasy, 1995). Although some of these studies con-
sider the importance of power in negotiations (e.g.,
Gibbins et al., 2001; Ng & Tan, 2003; Windsor &
Ashkanasy, 1995), they typically do not consider
the role of sharing private cost information. Fur-
thermore, auditor–client negotiations are often
viewed as mainly distributive. An important reason
is that audit clients seem to be inclined to negotiate
246 A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266
within a single-issue framework, which acts as a bar-
rier to ?nding integrative solutions (Gibbins, McC-
racken, & Salterio, 2005). Conversely, we expect
(and our results provide evidence) TCO to be a tool
that may support buyers in adopting a multiple-
issue framework during buyer–supplier negotia-
tions, which may enable buyers to create integrative
bargaining situations. Overall, accounting research
on inter?rm negotiations beyond the speci?c audit-
ing context has been very limited. This paper
answers recent calls for empirical research ‘‘with a
greater emphasis upon business processes and the
use of accounting in action/negotiation” (Tomkins,
2001, p. 164).
Our third contribution lies in the application of a
more sophisticated causal model to test for medi-
ated moderation e?ects of negotiation behavior on
negotiation outcomes. In an extensive review of the-
ory-consistent management accounting research,
Luft and Shields (2003) conclude that the majority
of management accounting studies employ additive
causal-model forms. Because such forms may limit
our understanding of management accounting by
representing causes and e?ects as universal rather
than conditional on other mediating or moderating
variables, Luft and Shields (2003) recommend the
use of more sophisticated causal models.
1
Muller,
Judd, and Yzerbyt (2005) observe that mediation
and moderation may be combined in informative
ways to better understand causal e?ects. However,
while well-developed and validated procedures exist
for examining whether treatment e?ects on an out-
come are mediated and/or moderated, very few
studies analyze mediation and moderation simulta-
neously. This study uses a rigorous three-step
approach to causal modeling in order to test for
mediated moderation e?ects (cf. Baron & Kenny,
1986). Our analyses indicate that TCO information
reduces the performance disadvantage of less pow-
erful buyers, and that this moderation e?ect is med-
iated by buyers’ bargaining behavior. Studying this
mediated moderation e?ect allows us to explain the
seemingly counterintuitive negotiation outcomes
that we ?nd.
The next section reviews the relevant literature
and develops our hypotheses. Section three presents
an outline of the experimental procedures and sec-
tion four reports the results. Section ?ve provides
discussion and suggestions for future research.
Theoretical background and hypotheses
Our hypotheses are based upon social exchange
theory
2
(e.g., Emerson, 1976; Thibaut & Kelley,
1959). Social exchange theory describes negotiation
as a process characterized by information exchange,
joint problem solving, and persuasion. In this
framework, negotiation outcomes (e.g., level of
buyer and/or supplier satisfaction, pro?ts, whether
or not agreement is reached) result from the com-
plex interaction of three constructs: negotiator char-
acteristics, situational characteristics, and the
characteristics of the negotiation process itself
(Campbell, Graham, Jolibert, & Meissner, 1988).
Negotiator characteristics and situational character-
istics are seen as a?ecting both process-related
behaviors and performance outcomes. In this study,
we control for negotiator characteristics and focus
on two situational characteristics: the availability
of TCO information and the relative power of the
buyer. Based on social exchange theory, we predict
that the e?ect of these situational characteristics on
negotiation outcomes is mediated by the negotiation
process.
The negotiation process may be thought of as a
series of interactions (Kelley & Thibaut, 1978).
According to social exchange theory, communica-
tion between parties during these interactions is very
1
More speci?cally, if the proposed causal model is additive,
Luft and Shields (2003, p. 198) recommend providing both the
reasons for assuming there are no important mediation or
moderation e?ects and the consequences of omitting these
relations if they exist. In other words, the default expectation is
that there are important mediation and moderation e?ects that
we should account for.
2
Although economic models have become the mainstream
explanatory models in many business disciplines, it has been
argued that advice to negotiators should depend on an under-
standing of the opponent’s actual decision process, rather than on
the assumption that the other party is fully rational (Bottom,
Holloway, Miller, Mislin, & Whitford, 2006). Indeed, recent
research suggests that negotiation behavior is seldom rational
(for a review, see Bazerman, Curhan, Moore, & Valley, 2000).
For example, evidence shows that despite the existence of an
agreement zone, deviations from rationality in individual deci-
sions lead to disagreements and Pareto-ine?cient agreements.
The problem with economic models is that researchers are subject
to the empirical indeterminacy of economic phenomena (Emer-
son, 1976). In most real-life situations, many di?erent outcomes
(from full cooperation to near-disastrous con?ict) are consistent
with ‘‘rationality”. Thus, game-theoretic models rarely provide
unique predictions of negotiation outcomes. Behavioral decision
models, such as exchange models, o?er a set of adjustments to
rational models (Murnigham & Bazerman, 1990).
A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266 247
important, as it shapes their perceptions of each
other and their interpretation of subsequent actions
(e.g., Bottom et al., 2006).
3
Communication in?u-
ences the development of a relationship, because it
provides an opportunity to express interest in the
exchange partner, to ascertain the exchange part-
ner’s openness to reciprocity and to in?uence the
exchange partner’s perceptions. Generally, two
basic types of negotiation behavior are distin-
guished within the negotiation process: problem
solving and distributive bargaining (Walton &
McKersie, 1966). Problem solving primarily
involves discovering ways to increase the bene?ts
available in the relationship (Walton & McKersie,
1966). In the context of purchasing, bargaining is
integrative to the extent that buyers actively seek
coordination with suppliers to develop alternative
purchasing arrangements that are likely to reduce
costs and/or increase performance. The intent is to
identify a solution via open and accurate informa-
tion exchange and to make trade-o?s based on pri-
orities and mutual respect for the other party’s
individual goals (Campbell et al., 1988). In contrast,
distributive bargaining is a process whereby each
party tries to maximize their private outcome (Wal-
ton & McKersie, 1966). Distributive bargaining is
characterized by the use of ?xed-sum or ‘‘win–lose”
tactics such as communicating threats, excessive
demands, positional commitments, and persuasive
arguments (Campbell et al., 1988). Prior studies
indicate that distributive bargaining is ‘‘natural”
or more salient to negotiators than problem solving
(e.g., Weingart, Hyder, & Prietula, 1996). Naive
negotiators are more familiar with distributive tac-
tics and therefore use these tactics as default
approaches. This ties in with their assumption that
the negotiation is a ?xed-sum game (Bazerman
et al., 2000).
The main e?ects of cost information and power on
negotiation behavior and outcomes
Cost information
Aprerequisite for any buyer–supplier relationship
is that cost information between exchange partners is
shared (e.g., Dekker, 2003, 2004; Tomkins, 2001).
Shared cost information can be used to analyze the
value chain in order to identify cost-reduction oppor-
tunities across the companies’ boundaries. Value
chain analysis (VCA), introduced by Porter (1985)
and further developed by Shank (1989) and Shank
and Govindarajan (1992), allows one to analyze,
coordinate, and optimize linkages between activities
in the value chain by focusing on the interdependence
between these activities (Dekker, 2003).
Traditionally, purchasing was simply a matter of
negotiating the best price. Accordingly, under tradi-
tional management accounting, purchasing decisions
tend to track only the purchase price associated with
a particular supplier, burying all the other costs the
supplier may introduce in the value chain of the pur-
chasing organization (Carr & Ittner, 1992; Degraeve
&Roodhooft, 1999). Traditional accounting systems
are therefore unable to adequately support a VCA
(Porter, 1985). In contrast, total cost of ownership
(TCO) systems, which attempt to quantify all the
costs associated with the purchase of a given product
or service, presume the existence of boundary-span-
ning activities (Wouters et al., 2005) and hence start
froma value chain perspective (Shank &Govindara-
jan, 1992). The cost and cost driver information
resulting from TCO analysis can be used to optimize
the performance of activities across the supply chain
(Porter, 1985). For instance, suppliers can be made
aware of the extra costs they generate and also of
ways to improve their competitive position by reduc-
ing these costs on the buyer’s side (e.g., Ellram, 1995;
Roodhooft et al., 2003, 2005). TCO information
therefore helps buyer–supplier partners detect
trade-o?s along the value chain and improve pro?t-
ability by modifying the way in which they do busi-
ness with each other (Wouters et al., 2005).
Accordingly, we expect that:
H1a: Buyers with TCO information achieve
higher individual pro?ts than buyers with
traditional cost information.
In order to reach integrative agreements that
lower the total costs over the value chain, buyers
need to have a good understanding of their own
3
Verbal communication between buyers and suppliers, a very
basic feature of inter?rm negotiations, has not been systemati-
cally examined in accounting research. One reason may be that,
according to economic models, such communications are irrel-
evant. The rationale is as follows: verbal claims made by either
party are not veri?able and thus not strategically credible, so they
will be ignored and should not in?uence the actions taken by
rational self-interested counterparts (Bottom et al., 2006). This
paper reports a detailed analysis of verbal communication of
accounting information during buyer–supplier negotiations. The
results provide clear evidence of the importance of this type of
communication.
248 A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266
priorities, communicate these to the supplier, and
integrate information about suppliers’ preferences
into their own understanding of the problem at
hand so that they can make trade-o?s between
important and unimportant issues (De Dreu, Bee-
rsma, Stroebe, & Euwema, 2006). Prior research
suggests that the availability of quanti?ed informa-
tion increases the likelihood that buyers engage in
such problem solving e?orts. For example, evidence
has been provided that ?nancial quanti?cations
carry more weight in decisions than non-?nancial
information (Ittner, Larcker, & Meyer, 2003), that
quanti?ed information enhances persuasion, espe-
cially if it is regarded as objective (Kadous, Koonce,
& Towry, 2005), and that relatively abstract costs
are less salient than outright losses (Northcraft &
Neale, 1986). Since TCO systems attempt to quan-
tify all relevant costs (as opposed to traditional cost
systems, which focus on price and do not quantify
other costs), and since quanti?ed costs carry more
weight in decisions, TCO systems provide better
insights into the relative importance of di?erent
costs. Hence, they provide relevant information that
enlarges the set of possible outcomes in a negotia-
tion, which in turn leads to better problem solving
(e.g., Kersten, 2001). As buyers have greater insight
into their own constraints and objectives, and as the
quanti?ed costs within TCO information are more
persuasive than traditional accounting information,
buyers are better-placed to cooperate with suppliers
on identifying and assessing alternative courses of
action. Thus, the availability of TCO information
should encourage buyers to engage in problem solv-
ing e?orts. Accordingly, we expect that:
H1b: Buyers with TCO information use more
problem solving techniques than buyers
with traditional cost information.
Power
Among the variables identi?ed as factors in?u-
encing negotiation processes and outcomes, power
is one of the most important (De Dreu & Van Kleef,
2004). Existing research on negotiations in account-
ing settings shows that various sources of power
may have an important impact on negotiation pro-
cesses and outcomes (e.g., Fisher et al., 2000; Ng &
Tan, 2003). In particular, the power that results
from the availability of alternative negotiation part-
ners is a key element of many negotiations: indeed,
if other negotiation partners are available, this
reduces a party’s dependence on the other side,
which weakens the other party’s power position
(Giebels, De Dreu, & Van de Vliert, 2000; Thibaut
& Kelley, 1959). Further, empirical research shows
that negotiators with a viable alternative achieve
higher personal outcomes than negotiators without
alternative options or with less attractive ones:
negotiators with less power tend to demand less,
make more concessions, and achieve less successful
outcomes than those with more power (Pinkley,
Neale, & Bennett, 1994).
4
We therefore expect that:
H2a: More powerful buyers achieve higher indi-
vidual pro?ts than less powerful buyers.
Turning again to negotiation behavior, a buyer’s
relative power is expected to a?ect his bargaining
aggressiveness. In particular, buyers’ use of distribu-
tive bargaining techniques, which include implicit or
explicit threats (e.g., warning the supplier that he is in
danger of losing the contract), are likely to rise with
supplier competition. First, such techniques are more
credible when several suppliers are interested in a
contract. Second, increased supplier competition
reduces the buyer’s need to solve problems with
any given supplier or group of suppliers (Perdue &
Summer, 1991). Consequently, we expect that:
H2b: More powerful buyers use fewer problem
solving techniques and more distributive
bargaining techniques than less powerful
buyers.
The mediated moderation e?ect of cost information
on the relation between power and individual pro?t
Prior research indicates that power not only in?u-
ences strategic decisions such as the use of distribu-
tive bargaining techniques during negotiation, but
also drives the processing of information (e.g., De
Dreu & Carnevale, 2003). For instance, compared
to individuals with a power advantage, those at a
power disadvantage are more motivated to under-
stand the other’s needs, desires, and possible actions
(Gelfand & Christakopoulou, 1999) and seek more
situational control by making a positive impression
(Goodwin, Gubin, Fiske, &Yzerbyt, 2000). Further,
negotiators with less power ask diagnostic rather
4
For a more extensive review of the e?ects of power, see Van
Kleef, De Dreu, Pietroni, and Manstead, 2006.
A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266 249
than leading questions and belief-congruent rather
than incongruent questions when facing a competi-
tive partner rather than a cooperative one (De Dreu
& Van Kleef, 2004). These ?ndings suggest that,
compared to more powerful individuals, less power-
ful individuals have a higher epistemic motivation,
i.e., a greater tendency to engage in information
acquisition and processing, and thus a greater desire
to develop and maintain a rich and accurate under-
standing of the opponent’s intentions and behavior
(Van Kleef et al., 2006). Because of their high episte-
mic motivation, dependent buyers may try to
(re)gain control over their own outcomes by paying
close attention to suppliers in order to accurately
predict their intentions and behaviors (De Dreu &
Van Kleef, 2004). In contrast, powerful individuals
are less likely to seek detailed information as they
have more resources available and/or are less depen-
dent on others, enabling them to act as they want
without serious consequences (Fiske, 1993; Keltner,
Gruenfeld, & Anderson, 2003; Van Kleef et al.,
2006). Because of their lower levels of epistemic
motivation, powerful buyers are more likely to solve
problems and to form an impression of their oppo-
nent through a quick, e?ortless, and heuristic pro-
cessing of information that rests on well-learned
prior associations (De Dreu & Van Kleef, 2004).
As distributive bargaining behavior is more salient
to negotiators than problem solving behavior (e.g.,
Weingart et al., 1996), we expect powerful buyers
to rely on distributive bargaining rather than on
problem solving techniques.
With regard to the implication of power across
TCO versus traditional information, it was already
pointed out that buyers with traditional informa-
tion, irrespective of whether they have power or
not, are encouraged to focus on price; these buyers
are less likely to take into account additional costs
that the supplier may cause. This limited scope of
information leaves traditional buyers (as opposed
to TCO buyers) with fewer opportunities to propose
a mutually acceptable solution, thus decreasing the
likelihood of problem solving e?orts.
Accordingly, we expect that less powerful buyers
with TCO information may try to (re)gain control
over their own outcomes by sharing information
5
and paying close attention to their opponent’s inten-
tions and behavior. That is, less powerful buyers
with TCO information are more likely to use inte-
grative bargaining techniques than either more pow-
erful buyers with TCO information, who are more
likely to using distributive bargaining techniques,
or buyers without TCO information, who will be
less able to create an integrative bargaining situa-
tion as their traditional cost information encourages
them to focus solely on price. We therefore expect
that:
H3a: Cost information moderates the e?ect of
power on buyers’ pro?ts, such that TCO
information reduces the performance disad-
vantage that less powerful buyers have com-
pared to more powerful buyers.
H3b: Cost information moderates the e?ect of
power on buyers’ bargaining behavior, such
that TCO information encourages less pow-
erful buyers to increase the use of problem
solving techniques and to decrease the use
of distributive bargaining techniques as
compared to more powerful buyers.
H3c: Buyers’ use of problem solving and distrib-
utive bargaining techniques mediate the
moderation e?ect of cost information on
the relation between power and buyers’
pro?ts.
To summarize, we expect that the overall e?ect of
power on the negotiation outcome will be moder-
ated by cost information and that this moderation
e?ect will be in?uenced by the e?ect of power and
cost information on negotiation behavior (the medi-
ator). These expectations are consistent with a med-
iated moderation hypothesis (Muller et al., 2005), as
represented in Fig. 1.
Research method
Experimental design
To test our hypotheses, we developed a 2 (TCO
information versus traditional cost informa-
tion) Â 2 (equal power versus low power) experi-
mental design, in which buyers and suppliers
negotiate a lease contract for a set of machines.
We manipulated the cost information and power
of the buyer in order to assess whether access to
detailed TCO information bene?ts buyers across
di?erent power settings. Cost information was
5
We do not expect that less powerful buyers will hoard their
private TCO information, as this would not be to their bene?t. In
order to bene?t from TCO information, this needs to be
exchanged so that trade-o?s between important and unimportant
issues can be made (Baiman & Rajan, 2002; De Dreu et al., 2006).
250 A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266
manipulated by providing the buyer with either
TCO information or traditional cost information
for the purchasing decision. The TCO information
quanti?ed all relevant costs (in this case: price, spare
parts, and maintenance) associated with the pur-
chase of a given quantity of products or services
from a given supplier. All costs were expressed in
the same currency as the price. Buyers with tradi-
tional information were given only an indication
of the costs and of the relative importance of each
of the issues to be negotiated (cf. Appendix A). This
corresponds to traditional management accounting
practices, which only track the purchase price and
‘‘bury” other purchasing costs in overhead accounts
or general expenses (Carr & Ittner, 1992). In addi-
tion, participants were informed that price is the
most expensive cost, followed by maintenance and
then spare parts. This information on the relative
importance of the di?erent issues provided buyers,
who had traditional cost information at their dis-
posal, with an understanding of their own prefer-
ences and priorities, which they could then
communicate to the suppliers in order to reach an
integrative solution.
We manipulated power through the availability
of an outside option (e.g., Giebels et al., 2000; Pink-
ley et al., 1994). In the equal power setting, the writ-
ten instructions for both the buyer and the supplier
contained a short paragraph about the presence of
an alternative negotiation partner. Buyers and sup-
pliers were equally powerful, in the sense that they
both had an outside option worth 1000 Euro. This
outside option was relatively unattractive, however,
as higher gains could be obtained by reaching an
agreement. In the low power setting, buyers had
no outside option and were fully dependent on
reaching an agreement with their partner to earn
any money.
6
Suppliers always had an outside option
of 1000 Euro and full cost and price information.
where
power
cost information
power*cost information
negotiation behavior
individual profit
= the manipulated independent variable
= the manipulated moderation variable
= the moderation of cost information on the effect of power on individual profit
= the mediation variable (in Tables 2, 3, and 4 measured by information
exchange, distributive behavior, and problem solving approach, respectively)
= the negotiation outcome variable
power
cost information
power*cost information
negotiation behavior
- information exchange
- distributive behavior
- problem solving approach
individual profit
Fig. 1. Regression models illustrating moderated mediation.
6
Consistent with prior research (for an overview, see Wolfe &
McGinn, 2005), the individual parties were aware of their own
alternatives but not the other party’s. If the buyer had no
alternative, the buyer was told that in the short run the only
supplier that could deliver the required machines was the supplier
they were negotiating with. This made the buyer highly depen-
dent on the supplier, but the supplier was unaware of this. This
situation allows us to fully focus on the e?ects of the buyer’s
situation (information and power) and how this a?ects negoti-
ation behavior and outcomes, without having to deal with the
complexity of the supplier’s perceptions of his or her power over
the buyer. This manipulation also has the additional advantage of
being a strong manipulation from the viewpoint of the buyer,
without having to deal with the ‘‘obvious” result of virtually all of
the pro?ts going to the supplier if the latter was aware of the
buyer’s complete dependence.
A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266 251
Subjects and procedures
Two hundred and eight participants were
recruited from a graduate management accounting
course that was part of a Master’s program in busi-
ness administration at a large West-European uni-
versity. The course had covered traditional
accounting methods, activity-based costing, TCO,
and supplier selection problems before the experi-
ment took place. The experiment was run in a com-
puter laboratory. Each session was restricted to a
maximum of one hour. The opportunity to earn a
performance-based cash payout was the only incen-
tive o?ered. Participants earned 0.5% of their com-
pany’s pro?ts (on average 5.74 Euro; min = 0
Euro; max = 15 Euro). The participants were ran-
domly assigned to one of the four experimental con-
ditions and to a buyer or a supplier role. Procedures
were identical for all treatments.
Buyers and suppliers sat in di?erent rooms so
that they were unable to identify their partner’s
identity; personality e?ects and collusion were there-
fore precluded. Participants read the instructions
describing their role and the nature of the bargain-
ing task and could play the game at their own pace.
The supplier started the game by making a ?rst
o?er. Participants could send messages along with
their o?ers and countero?ers if they desired to do
so. All messages were recorded. The game ended
when (i) an agreement was reached, (ii) a player
opted for the outside option, or (iii) after 10 rounds.
In the last case, participants were informed by the
computer program that time was running out. This
only happened in a few cases (4 out of 104).
The bargaining task
The bargaining task is based on a negotiation
game developed by Kelley (1966) and applied by
many other researchers (e.g., Campbell et al.,
1988; De Dreu & Van Kleef, 2004). The game was
adapted to suit a TCO setting. This means that
the payo? tables in Kelley’s game were replaced
by cost tables for the buyers and cost and income
tables for the suppliers. The tables were constructed
in such a manner that the minimum and maximum
pro?ts that could be earned were the same for buy-
ers and suppliers (cf. Appendix A). Buyers and sup-
pliers had to negotiate a lease contract for a set of
machines. The buyer could earn a ?xed income (of
6000 Euro) by selling end products to an end cus-
tomer. The instructions explained that maintenance
and spare parts were needed to run the machines
and to produce the end product. Consequently,
the game involved the simultaneous negotiation of
price, maintenance, and spare parts. For each of
these issues, nine di?erent contract terms were
possible.
Price was an income for the supplier and a cost
for the buyer. Accordingly, the price issue was dis-
tributive in nature. This issue was worth the same
for each negotiator, with preferences on the issue
going in opposite directions. Consequently, one
party’s gain was equal to the other party’s loss.
The task, however, provided an opportunity for
the parties to integrate their personal interests.
The buyer had a comparative advantage in taking
care of the spare parts and the supplier had a com-
parative advantage in maintaining the machines.
Since the issue that was most valuable to one party
was automatically less valuable to the other party,
it was possible for participants to trade-o? issues.
Such ‘‘logrolling”, giving up on less valuable issues
to maximize outcomes on the most valuable ones,
could yield a fully integrative solution or a Pareto-
optimal solution (Kersten, 2001). The Pareto-opti-
mal solution is the solution whereby neither dyad
member can improve his situation without the other
party being worse o?; in other words, no other com-
bination of contracts o?ers as much or more pro?t
to both parties. The Pareto-optimal joint outcome
could be reached when the buyer and supplier
agreed on contract 5AZ. In this agreement, the dis-
tributive issue (price) is set at the middle and the two
integrative issues (maintenance and spare parts) are
fully traded o?. In this situation, denoted with an
asterisk (
*
) in Appendix A, the maximum level of
joint pro?t was reached and was equally divided
between the buyer and the supplier.
7
As cost tables
were private, participants had to ?nd out the possi-
bility of a win–win solution through the process of
o?ers and countero?ers and by exchanging informa-
tion about their interests.
Before the negotiations started, participants were
informed that they should realize that their oppo-
nent’s cost structure did not mirror their own cost
structure. Participants were reminded that, just like
7
The price contract did not in?uence the joint pro?ts:
maximum joint pro?ts were reached as long as participants
agreed on contract AZ. Nine (1AZ to 9IZ) di?erent contract
combinations led to maximum joint pro?ts. However, only
contract 5AZ divided the maximum joint pro?t equally between
the buyer and the supplier.
252 A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266
in real life, companies may have di?erent competi-
tive advantages and that, as a result, their company
might be competitively better at maintenance than
their opponent, while their opponent might be more
competitive in spare parts (or vice versa).
8
Further-
more, the instructions informed participants that at
all times during the negotiation a table would be
shown on their screen with an overview of their
opponent’s proposals and their own counter-pro-
posals, so that they would have an overview of all
proposals already made.
9
Dependent variables
Negotiation outcomes
As we want to assess whether buyers can bene?t
from more re?ned TCO data in making purchasing
decisions, this study focuses on the buyer’s individual
pro?t. This is the amount of money earned by a
buyer at the conclusion of the negotiations; it is
measured as the buyer’s individual pro?t level asso-
ciated with ?nal agreement in the negotiation.
Although this is not central to the analysis, the
study also brie?y discusses the results relating to
the supplier’s individual pro?t and the dyad’s joint
pro?t, as these provide a deeper understanding of
the study’s main ?ndings.
Negotiation behavior
Negotiation behavior is derived from two sets of
measures: (1) a set based on interaction analysis,
and (2) a set based on participants’ responses to
the post-game questionnaire administered immedi-
ately following the negotiation exercise.
First, verbal behavior is coded by means of inter-
action analysis in order to examine categories and
meanings embedded in structural patterns of com-
munication (Putman & Fairhurst, 2001). The classi-
?cation scheme, which is included in Appendix B1,
is based on negotiation communication coding
schemes used in prior studies (e.g., Giebels et al.,
2000). Three judges, who were blind to conditions
or hypotheses, coded each negotiation indepen-
dently. Inter-rater agreements, expressed in Cohen’s
Kappa, varied between 0.75 and 0.95. After com-
pleting the coding, the coders compared their cod-
ing and resolved disagreements by jointly revisiting
the negotiation messages and producing a single
set of codes for each subject. Negotiation behavior
is determined by analyzing this last set of codes.
Messages sent by participants were coded for (i)
problem solving techniques, measured as informa-
tion exchange and as other integrative behaviors,
and (ii) distributive bargaining techniques, such as
sending threats or issuing warnings. Information
exchange re?ects the extent to which a participant
shared information about priorities. Information
exchange was coded ‘‘0” for participants not reveal-
ing any information about their cost structure, ‘‘1”
for participants revealing the relative importance
(but not the numerical values) of each of the three
issues under negotiation, and ‘‘2” for participants
revealing both the relative importance and the
numerical values of each of the three issues under
negotiation. Integrative behavior is based on the
average of two measures: the number of rewards
o?ered and the number of positive normative
appeals extended (Cronbach’s alpha = 0.70). Inte-
grative behavior is also assessed by counting the
number of explicit requests for cooperation. How-
ever, as the number of requests for cooperation
did not load with the rewards and the positive nor-
mative appeals on one factor, it is treated as a
8
This was illustrated in the instructions by an example relating
to maintenance. Buyers with TCO information were referred to
the cost table for cells 1 and 2 in Appendix A: if they agreed on
maintenance contract ‘‘A”, their company would have a relatively
low maintenance cost (€ 250), but the supplier would have to
provide the maximum level of maintenance. It was explained that
this did not mean that the supplier’s cost for providing this level
of maintenance would be € 2250, as this was the cost to the buyer
if he were to provide the maximum level of maintenance himself;
rather, this amount was di?erent from what the supplier would
have to pay in order to provide the same maintenance level, since
the cost structures of both companies were di?erent. The
instructions also explained that the same reasoning was to be
applied to spare parts. Buyers with traditional cost information
were referred to the cost table for cells 3 and 4 in Appendix A and
were provided with the same text, except that the cost ?gures were
replaced by the corresponding number of maintenance sessions
(as provided in the cost table for cells 3 and 4 in Appendix A).
9
The information in this table was essentially the same in all
experimental cells. The table provided an overview of all
proposals made by both players (e.g., proposal supplier: 6EV;
counter-proposal buyer: 4DW; counter-proposal supplier: 5CA;
etc.). Thus, at all times during the negotiation, participants could
check the history of proposals and counter-proposals made.
Participants with TCO information also received the correspond-
ing pro?t or loss ?gure for the di?erent o?ers. This ?gure
provided no new or extra information, as participants with TCO
information could easily compute this ?gure from the informa-
tion provided in their private cost tables. This information was
solely provided for the purpose of speeding up the game. The
pro?t/loss ?gure was not provided in experimental cells with
traditional information: participants in these experimental cells
could derive the corresponding number of maintenance sessions
and spare parts from their private cost tables, but not the exact
corresponding costs.
A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266 253
separate measure. Distributive behavior is a sum-
mated scale calculated by adding ?ve types of dis-
tributive bargaining techniques together and
determining the mean value. The ?ve behavior types
included are as follows: general threats, exit threats,
warnings, commitment and punishments (Cron-
bach’s alpha = 0.81).
Second, negotiation behavior is derived from
post-game questionnaires. Participants rated their
opponent’s bargaining behavior on four items that
measure the overall problem solving approach of
the buyers as assessed by the supplier (Cronbach’s
alpha = 0.88). The di?erent items are listed in
Appendix B2 and are based on prior studies (e.g.,
Campbell et al., 1988). A summated scale was
obtained by calculating the mean value across the
four items.
Results
Experimental checks
After completing the negotiation task, partici-
pants ?lled out a post-bargaining questionnaire on
a 1–5 scale, enabling us to check for motivation,
task understanding, and their usage of cost reports.
All checks on the experimental conditions (on cost
information relevance and power) are statistically
signi?cant and have means in the appropriate direc-
tion. More powerful buyers considered themselves
more powerful (F(1, 102) = 14.61, p < 0.01) and
buyers with TCO information judged the cost infor-
mation they received to be more relevant than buy-
ers with traditional cost information (F(1,
102) = 41.62, p < 0.01). Checks on procedures,
including the subject’s involvement in the task, their
understanding of instructions and payo? tables, and
whether they had su?cient time to complete the
exercise, show no di?erences between conditions
(p > 0.10), as is appropriate. Means on these ques-
tions indicate that subjects were highly involved
(mean = 4.33; std = 0.69), that they assessed the
exercise as ‘‘fun” (mean = 4.07; std = 0.70), that
they understood both the instructions (mean = 4.41;
std = 0.77) and the payo? tables (mean = 4.62;
std = 0.51), and that they had enough time to com-
plete the task (mean = 4.40; std = 0.98). Partici-
pants in di?erent experimental cells required an
equal amount of time to read the instructions and
familiarize themselves with the game before actually
starting the negotiation (p > 0.10). On average, par-
ticipants needed 559 s (9.3 min) to read the instruc-
tions. Participant gender produced neither main nor
interaction e?ects on negotiation process or out-
comes and is therefore excluded from further
analysis.
The e?ects of cost information and power on
negotiation outcomes
Table 1 presents the results.
10
A correlation
matrix of all variables included in the study is pro-
vided in Appendix C. As expected in H1a, buyers
with TCO information obtained signi?cantly higher
individual pro?ts than buyers with traditional cost
information (F(1, 100) = 52.90, p < 0.01). Consis-
tent with H2a, more powerful buyers generated
higher individual pro?ts than less powerful buyers
(F(1, 100) = 14.02, p < 0.01). The results also indi-
cate a moderation e?ect of cost information on
the overall e?ect of power on individual pro?t
(F(1, 100) = 4.58, p < 0.05). The performance de?cit
of less powerful buyers actually disappeared when
they had TCO information: the individual pro?t
made by buyers with TCO information and an out-
side option did not signi?cantly di?er from the
pro?t made by buyers who had access to the same
information but no outside option (buyer’s mean
individual pro?t = 1148.33 versus buyer’s mean
individual pro?t = 1007.69; F(1, 54) = 2.46,
p > 0.10). The performance de?cit of less powerful
buyers was thus less pronounced when buyers had
detailed TCO information, supporting H3a.
The e?ects of cost information and power on
negotiation behavior
To explain negotiation outcomes, we analyze
participants’ negotiation behavior. First, we test
H1b, which predicts that buyers with TCO informa-
tion use problem solving techniques more frequently
than buyers with traditional cost information. An
ANOVA on information exchange reveals a main
e?ect for cost information (F(1, 100) = 6.14,
10
Analyses are performed for all subjects in the study. However,
we obtain similar results when excluding the subjects that opted
for an outside option and thus did not reach agreement. Only 14
out of 208 subjects opted for this outside option, which may be
explained by the fact that the outside option was relatively
unattractive, as higher gains could be obtained by reaching an
agreement. Pairwise comparisons with Tukey tests indicate that
the number of dyads choosing for an outside option is not
signi?cantly di?erent across the four experimental conditions
(p > 0.10).
254 A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266
p < 0.05): buyers with TCO information disclosed
more information than buyers with traditional
information. An ANOVA on integrative behavior
indicates that the e?ect for cost information (F(1,
100) = 2.39, p = 0.12) is not signi?cant at the 10%-
level, but closely approaches signi?cance. An
ANOVA on the number of explicit requests for
cooperation shows a main e?ect for cost information
(F(1, 100) = 10.45, p < 0.01), indicating that buyers
with TCO information requested active cooperation
more frequently than buyers with traditional cost
information. Buyers’ negotiation behavior is further
examined by an ANOVA on the composite measure
problem solving approach. A signi?cant main e?ect
for cost information is found (F(1, 100) = 242.00,
p < 0.01), indicating that, according to the suppliers,
buyers with TCO information used problem solving
techniques more frequently than buyers with tradi-
tional information. Although the e?ect of cost infor-
mation on integrative behavior is somewhat weaker,
the e?ects of cost information on information
exchange, requests for cooperation and problem solv-
ing approach, provide strong support for H1b. The
ANOVA on distributive behavior reveals an unex-
pected main e?ect of cost information (F(1,
100) = 5.77, p < 0.05), indicating that TCO infor-
mation signi?cantly increased the use of distributive
bargaining techniques. Unreported Tukey tests indi-
cate that this e?ect is mainly driven by buyers with
TCO information and an outside option, as these
buyers used distributive bargaining techniques more
frequently than buyers in the other three conditions
(p < 0.01). The Tukey tests also indicate that buyers
with TCO information but no outside option did not
use distributive bargaining techniques more fre-
quently than buyers in the two conditions with tra-
ditional information (p > 0.10).
In order to test H2b, we analyze the e?ect of
power on the negotiation behavior variables. An
ANOVA on information exchange shows a main
Table 1
Analysis of negotiation outcomes and buyer’s behavior
Panel A: Summary statistics for negotiation outcomes and buyer’s behavior
a
TCO information Traditional cost information
Outside option No outside option Outside option No outside option
Buyer’s individual pro?t 1148.33(184.99) 1007.69(449.59) 698.08 (479.06) 181.81 (616.74)
Supplier’s individual pro?t 1188.33 (205.81) 1438.46 (349.66) 1421.15 (427.83) 1881.81 (680.61)
Joint pro?t 2336.67 (324.29) 2446.15 (506.15) 2119.23 (263.85) 2063.64 (405.38)
Information exchange 0.37 (0.49) 0.73 (0.78) 0.23 (0.43) 0.32 (0.48)
Integrative behavior 0.02 (0.09) 0.27 (0.55) 0.06 (0.16) 0.05 (0.15)
Requests for cooperation 0.17 (0.46) 0.69 (0.88) 0.08 (0.27) 0.09 (0.29)
Distributive behavior 0.57 (0.73) 0.06 (0.11) 0.18 (0.30) 0.05 (0.14)
Problem solving approach 2.98 (0.71) 3.62 (0.57) 1.47 (0.40) 1.70 (0.45)
Panel B: ANOVA for negotiation outcomes and buyer’s behavior
b
Cost information Power Cost information
*
power
Buyer’s individual pro?t 52.90 (
***
) 14.02 (
***
) 4.58 (
**
)
Supplier’s individual pro?t 15.80 (
***
) 17.46 (
***
) 1.53
Joint pro?t 15.78 (
***
) 0.13 1.20
Information exchange 6.14 (
**
) 4.16 (
**
) 1.56
Integrative behavior 2.39 4.13 (
**
) 5.02 (
**
)
Requests for cooperation 10.45 (
***
) 6.37 (
**
) 5.73 (
**
)
Distributive behavior 5.77 (
**
) 14.20 (
***
) 4.89 (
**
)
Problem solving approach 242.00 (
***
) 15.85 (
***
) 3.44 (
*
)
(
***
), (
**
), and (
*
) indicate signi?cance levels of 1%, 5%, and 10%.
a
Variable de?nitions in Appendix C. The table cells in Panel A contain, for each of the experimental cells, the means and standard
deviation for the variables. Standard deviations are in parentheses.
b
Panel B presents the results of eight ANOVA analyses. Cost information and power are the between-subject factors. The ANOVA on
the buyer’s individual pro?t tests for the main e?ect of cost information (H1a), the main e?ect of power (H2a), and the moderation e?ect
(H3a) on buyer’s individual pro?t. The ANOVA’s on the supplier’s individual pro?t and on the joint pro?t are supplementary analyses to
provide a deeper understanding of the study’s main ?ndings. The last ?ve ANOVA’s are tests for the main e?ect of cost information (H1b),
the main e?ect of power (H2b), and the moderation e?ect (H3b) on the buyer’s use of problem solving and distributive bargaining
techniques. The buyer’s use of problem solving techniques is assessed through the buyer’s information exchange, integrative behavior, and
requests for cooperation. The buyer’s use of distributive bargaining techniques is assessed through the buyer’s distributive behavior. The last
measure, problem solving approach, provides an overall assessment of the buyer’s negotiation behavior as assessed by the supplier.
Reported are the F-statistics.
A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266 255
e?ect for power (F(1, 100) = 4.16, p < 0.05): more
powerful buyers disclosed less information than less
powerful buyers. The main e?ects of power on inte-
grative behavior (F(1, 100) = 4.13, p < 0.05) and on
the number of explicit requests for cooperation
(F(1, 100) = 6.37, p < 0.05) indicate that more pow-
erful buyers used fewer integrative bargaining tech-
niques and formulated fewer requests for
cooperation than buyers without an outside option.
We also ?nd a strong main e?ect of power on dis-
tributive behavior (F(1, 100) = 14.20, p < 0.01), indi-
cating that power signi?cantly increased the use of
distributive bargaining techniques. Finally, buyers’
overall negotiation behavior is examined by means
of the composite measure problem solving approach.
A signi?cant main e?ect is found for power (F(1,
100) = 15.85, p < 0.01). Thus, according to the sup-
pliers, less powerful buyers adopted problem solving
strategies more frequently than buyers with an out-
side option. Overall, these results provide strong
support for H2b.
Finally, with respect to H3b, we examine whether
cost information moderates the overall e?ect of
power on negotiation behavior. Although the
ANOVA on information exchange indicates that
the moderation e?ect of cost information on power
is not signi?cant at the 10%-level, pairwise compar-
isons with Tukey tests indicate that buyers with
TCO information and no outside option exchanged
information more frequently than buyers with TCO
information and an outside option (p < 0.10). We
?nd that cost information has a signi?cant modera-
tion e?ect on the relationship between power and
integrative behavior (F(1, 100) = 5.02, p < 0.05).
Pairwise comparisons with Tukey tests indicate that
buyers with TCO information and no outside
option used integrative bargaining techniques more
frequently than buyers in the three other conditions
(p < 0.10). Cost information also moderates the
e?ect of power on the number of explicit requests
for cooperation (F(1, 100) = 5.73, p < 0.05). Pairwise
comparisons with Tukey tests indicate that buyers
with TCO information and no outside option
requested active cooperation more frequently than
buyers in the three other conditions (p < 0.01).
Additionally, we ?nd a moderation e?ect of cost
information on the overall e?ect of power on distrib-
utive behavior (F(1, 100) = 4.89, p < 0.05). Pairwise
comparisons with Tukey tests reveal that buyers
with TCO information and an outside option used
signi?cantly more distributive bargaining tech-
niques than buyers in the other three experimental
conditions (p < 0.01). Finally, we also ?nd a signi?-
cant moderation e?ect on problem solving approach
(F(1, 100) = 3.44, p < 0.10). Pairwise comparisons
with Tukey tests indicate that, according to the sup-
pliers, buyers with TCO information and no outside
option made more extensive use of problem solving
techniques than buyers in the other conditions
(p < 0.01). Overall, these results on the moderating
e?ect of cost information on the relation between
power and negotiation behavior, provide support
for H3b.
The mediated moderation e?ect of cost information
on the relation between power and individual pro?t
Based on exchange theory, we hypothesize in
H3c that the negotiation process has a mediation
e?ect on the negotiation outcome. More speci?cally,
we expect the nature of negotiation behavior to
mediate the moderation e?ect of cost information
on the overall e?ect that power has on individual
pro?t (cf. Fig. 1). Since the e?ect of power on nego-
tiation outcomes is moderated by cost information
and since we expect this moderation e?ect to be
mediated by negotiation behavior, we use Baron
and Kenny’s (1986) framework for combining mod-
eration and mediation. We perform the analyses for
the two main types of bargaining behavior: problem
solving behavior (operationalized as information
exchange) and distributive behavior.
11
In a sensitiv-
ity analysis, we also conduct the analysis for the
measure of the overall problem solving approach
of the buyers as assessed by the supplier.
11
Baron and Kenny (1986) procedure to test for mediation and
moderation is designed to test the e?ects of a single moderator
and a single mediator. In the present analysis, however, we are
interested in the mediating e?ects of problem solving as well as
distributive bargaining. Accordingly, we perform the procedures
for the di?erent mediating variables separately. Table 2 intro-
duces (in Step 3) information exchange and its interaction term
with cost information, whereas Table 3 introduces (again in Step
3) distributive behavior and its interaction e?ect with power. The
mediation–moderation procedure is not performed for the
combination of distributive behavior and cost information,
because we only expect an e?ect of TCO on problem solving
behavior and not on distributive behavior (cf. H1b). Because we
expect power to a?ect problem solving (cf. H2b), we also
performed the mediation–moderation procedure for the combi-
nation of information exchange and power. The results are not
signi?cant and are therefore not reported. To test for overall
mediating e?ects, we conduct the analyses reported in Table 5.
Although this overall test provides an interesting addition to the
three-step procedures listed in Tables 2–4, it is not su?cient in
itself to demonstrate mediated moderation (Muller et al., 2005).
256 A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266
First, we consider the e?ect of information
exchange. As shown in Table 2, this analysis pro-
ceeds in three steps. The ?rst step is a regression
analysis of power, cost information, and the mod-
eration e?ect of cost information on the e?ect of
power on individual pro?t. In the second step,
two equations are estimated: (1) information
exchange is regressed on power, on cost informa-
tion, and on the moderation e?ect of cost informa-
tion on the e?ect of power, and (2) individual
pro?t is regressed on power, on cost information,
on the moderation e?ect of cost information on
the e?ect of power, and on information exchange.
Eventually, in the third step, one equation is esti-
mated: individual pro?t is regressed on power, on
cost information, on the moderation e?ect of cost
information on the e?ect of power, on information
exchange, and on the moderation e?ect of cost
information on information exchange. This last
equation is identical to the second Step 2 equation,
except that the moderation term of cost informa-
tion on information exchange is now added. The
key question is the extent to which the moderation
e?ect of cost information on the relation between
power and individual pro?t is reduced in moving
from Step 2 to Step 3. Information exchange medi-
ates the moderation e?ect of cost information on
power if the following conditions are met (Baron
& Kenny, 1986, p. 1179): (1) the moderation e?ect
of cost information on power should a?ect individ-
ual pro?t less at Step 3 than at Step 2; (2) in Step
3, the moderation e?ect of cost information on
information exchange should a?ect individual
pro?t signi?cantly; (3) ?nally, in Step 2 cost infor-
mation should a?ect information exchange, which
results in a correlation between the moderation
e?ect of cost information on the e?ect of power
(i.e., power
*
cost information) and the moderation
e?ect of cost information on the e?ect of informa-
tion exchange (i.e., cost information
*
information
exchange). As can been seen in Table 2, all these
conditions are met. In Step 2, cost information sig-
ni?cantly explains information exchange (coe?-
cient = 0.41, p < 0.05), and in Step 3 the
moderation e?ect of cost information on informa-
tion exchange signi?cantly a?ects individual pro?t
(coe?cient = 651.51, p < 0.01). Furthermore, the
moderation e?ect of cost information on the e?ect
of power on individual pro?t is reduced in moving
from Step 2 to Step 3, where it has dropped to a
non-signi?cant level (coe?cient in Step
1 = À375.62, p < 0.05 and coe?cient in Step
3 = À250.73, p > 0.10), indicating ‘‘full” mediated
moderation. These results imply that information
exchange mediates the moderation e?ect of cost
information on power.
Table 2
Three-step regression procedure for testing mediation and moderation of information exchange
a
Step 1 Step 2 Step 3
Buyer’s individual
pro?t
Information
exchange
Buyer’s individual
pro?t
Buyer’s individual
pro?t
Constant 181.82 (
*
) 0.32 (
***
) 165.53 (
*
) 310.99 (
***
)
Cost information 825.87 (
***
) 0.41 (
**
) 804.75 (
***
) 517.27 (
***
)
Power 516.26 (
***
) À0.09 520.73 (
***
) 480.77 (
***
)
Power
*
cost information À375.62 (
**
) À0.28 À361.45 (
**
) À250.73
Information exchange 51.19 À405.96 (
***
)
Cost information
*
information
exchange
651.51 (
***
)
R
2
(F for R
2
) 0.41 (
***
) 0.11 (
***
) 0.41 (
***
) 0.49(
***
)
(
***
), (
**
), and (
*
) indicate signi?cance levels of 1%, 5%, and 10%.
a
Variable de?nitions in Appendix C. This table presents our results on Baron and Kenny’s (1986) three-step procedure to test whether
buyers’ use of problem solving techniques (operationalized as information exchange) mediates the moderation e?ect of cost information on
the relation between power and buyers’ pro?ts (H3c). (See Table 3 for our test for distributive bargaining techniques.) The table above
presents the results of four regression analyses. In Step 1, the dependent variable buyer’s individual pro?t is regressed on cost information,
on power, and on the moderation term power
*
cost information. In Step 2, the dependent variable information exchange is ?rst regressed
on cost information, on power, and on the moderation term power
*
cost information. Then, the dependent variable buyer’s individual
pro?t is regressed on cost information, on power, on the moderation term power
*
cost information, and on information exchange. In Step
3, the dependent variable buyer’s individual pro?t is regressed on cost information, on power, on the moderation term power
*
cost
information, on information exchange, and on the moderation term cost information
*
information exchange. Regression coe?cients are
reported.
A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266 257
The same three-step procedure is undertaken for
distributive bargaining behavior (cf. Table 3). As
shown by the results, neither cost information nor
power has a signi?cant e?ect on distributive behav-
ior in Step 2. However, mediated moderation is also
demonstrated (1) if the moderation e?ect of cost
information on power has less of an e?ect on indi-
vidual pro?t in Step 3 than in Step 2, (2) if the e?ect
of power on distributive behavior is moderated by
cost information in Step 2, and (3) if the e?ect of
distributive behavior on individual pro?t is non-
zero (Muller et al., 2005). As can been seen in Table
3, all these conditions are met. In Step 2 the moder-
ation e?ect signi?cantly explains distributive behav-
ior (coe?cient = 0.37, p < 0.05), and in Step 3
distributive behavior signi?cantly a?ects individual
pro?t (coe?cient = 942.23, p < 0.10). Furthermore,
the moderation e?ect of cost information on the
relation between power and individual pro?t is
reduced in moving from Step 2 to Step 3 (though
it does not drop to non-signi?cance, providing evi-
dence for ‘‘partial” mediated moderation). Finally,
in a sensitivity analysis, the three-step procedure is
also applied to determine the buyer’s overall prob-
lem solving approach as assessed by the supplier
(cf. Table 4). The results are very similar to those
in Table 2. Overall, we conclude that the results
reported in Tables 2 and 4 provide support for
our expectation that the moderation e?ect of cost
information on the relation between power and buy-
ers’ pro?ts is mediated by buyers’ problem solving
behavior. Table 3 provides support that the moder-
ation e?ect is partially mediated by the buyers’ dis-
tributive bargaining behavior. Together, these
results support H3c.
In a last set of analyses, we test for the overall
mediation e?ect of the negotiation process (cf. Table
5).
12,13
We conduct a hierarchical regression analy-
sis, in which dummy variables representing the
manipulations are entered in Step 1 and the pro-
posed sets of mediators are entered in Step 2 (cf.
Weingart et al., 1996). Results for regression (1)
are very similar to those obtained from the ANOVA
Table 3
Three-step regression procedure for testing mediation and moderation of distributive behavior
a
Step 1 Step 2 Step 3
Buyer’s individual pro?t Distributive behavior Buyer’s individual pro?t Buyer’s individual pro?t
Constant 181.82 (
*
) 0.05 179.73 (
*
) 138.99
Cost information 825.87 (
***
) 0.02 825.13 (
***
) 810.72 (
***
)
Power 516.26 (
***
) 0.13 510.21 (
***
) 557.21 (
***
)
Power
*
cost information À375.62 (
**
) 0.37 (
**
) À392.81 (
**
) À364.60 (
**
)
Distributive behavior 46.00 942.23 (
*
)
Power
*
distributive behavior À931.62 (
*
)
R
2
(F for R
2
) 0.41 (
***
) 0.21 (
***
) 0.41 (
***
) 0.42 (
***
)
(
***
), (
**
), and (
*
) indicate signi?cance levels of 1%, 5%, and 10%.
a
Variable de?nitions in Appendix C. This table presents our results on Baron and Kenny’s (1986) three-step procedure to test whether
buyers’ use of distributive bargaining techniques mediates the moderation e?ect of cost information on the relation between power and
buyers’ pro?ts (H3c). (See Table 2 for our tests for problem solving techniques.) The table above presents the results of four regression
analyses. In Step 1, the dependent variable buyer’s individual pro?t is regressed on cost information, on power, and on the moderation
term power
*
cost information. In Step 2, the dependent variable distributive behavior is ?rst regressed on cost information, on power, and
on the moderation term power
*
cost information. Then, the dependent variable buyer’s individual pro?t is regressed on cost information,
on power, on the moderation term power
*
cost information, and on distributive behavior. In Step 3, the dependent variable buyer’s
individual pro?t is regressed on cost information, on power, on the moderation term power
*
cost information, on distributive behavior,
and on the moderation term power
*
distributive behavior. Regression coe?cients are reported.
12
The process variables in Table 5 are information exchange
and distributive behavior (and their interaction with cost infor-
mation and power, respectively). Replacing the information
exchange variable by the problem solving behavior variable yields
similar results.
13
Finally, we perform several sensitivity analyses to test the
robustness of the results against di?erent ways of measuring
negotiation behavior. First, similar results are obtained when
information exchange is coded ‘‘0” for participants not revealing
any information about their cost structure, and ‘‘1” for partic-
ipants revealing the relative importance of each of the three issues
(with or without the numerical values). Second, we analyze
participants’ ratings of their own bargaining strategies. The self-
assessed problem solving behavior of the buyers correlates well
with suppliers’ assessment of buyers’ problem solving behavior
(Pearson correlation = 0.41, p < 0.01). However, as one of the
four items (i.e., the question whether the participant was honest
or deceptive) assessed by the buyers does not load with the other
items on one factor, the problem solving construct is based on the
supplier’s assessment of the buyer’s behavior. Similar results are
obtained and hence the conclusions remain the same when the
problem solving approach construct combines the items from
both the suppliers’ and the buyers’ questionnaires.
258 A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266
analysis: both main e?ects and the moderation e?ect
are signi?cant. The R
2
of the model is 0.41 and the
F-test is highly signi?cant (p < 0.01). When the pro-
cess variables are added to the equation, the moder-
ation e?ect of cost information on the e?ect of
power on individual pro?t drops to a non-signi?-
cant level (coe?cient = À233.69, p > 0.10) and all
process variables reach signi?cance. Information
exchange has a negative impact on individual pro?t
(coe?cient = À467.70, p < 0.01). However, the
positive and signi?cant moderation e?ect of cost
information on the e?ect of information exchange
Table 4
Three-step regression procedure for testing mediation and moderation of problem solving approach
a
Step 1 Step 2 Step 3
Buyer’s
individual pro?t
Problem
solving approach
Buyer’s
individual pro?t
Buyer’s
individual pro?t
Constant 181.82 (
*
) 1.71 (
***
) 174.60 1012.78 (
***
)
Cost information 825.87 (
***
) 1.91 (
***
) 817.78 (
***
) À666.23
Power 516.26 (
***
) À1.45 517.25 (
***
) 402.48 (
***
)
Power
*
cost information À375.62 (
**
) À1.85 (
*
) À373.89 (
**
) À144.73
Problem solving approach 4.24 À487.50 (
***
)
Cost information
*
problem solving approach 670.37 (
***
)
R
2
(F for R
2
) 0.41 (
***
) 0.72 (
***
) 0.41 (
***
) 0.49 (
***
)
(
***
), (
**
), and (
*
) indicate signi?cance levels of 1%, 5%, and 10%.
a
Variable de?nitions in Appendix C. This table reports a sensitivity analysis using Baron and Kenny’s (1986) three-step procedure to
test whether buyers’ overall problem solving approach (as assessed by the suppliers) mediates the moderation e?ect of cost information on
the relation between power and buyers’ pro?ts (H3c). The table presents the results of four regression analyses. In Step 1, the dependent
variable buyer’s individual pro?t is regressed on cost information, on power, and on the moderation term power
*
cost information. In Step
2, the dependent variable problem-solving approach is ?rst regressed on cost information, on power, and on the moderation term
power
*
cost information. Then, the dependent variable buyer’s individual pro?t is regressed on cost information, on power, on the
moderation term power
*
cost information, and on problem solving approach. In Step 3, the dependent variable buyer’s individual pro?t is
regressed on cost information, on power, on the moderation term power
*
cost information, on problem solving approach, and on the
moderation term cost information
*
problem solving approach. Regression coe?cients are reported.
Table 5
Hierarchical regression results for buyer’s individual pro?t
a
Equation 1: situational
characteristics
Equations 2: situational characteristics
and negotiation process
Constant 181.82 (
*
) 276.58 (
***
)
Cost information 825.87 (
***
) 488.09 (
***
)
Power 516.26 (
***
) 528.19 (
***
)
Power
*
cost information À375.62 (
**
) À233.69
Information exchange À467.67 (
***
)
Cost information
*
information exchange 700.11 (
***
)
Distributive behavior 1188.89 (
**
)
Power
*
distributive behavior À1181.94 (
**
)
R
2
0.41 0.52
F for R
2
22.66 (
***
) 14.84 (
***
)
Change in R
2
0.11
F for change in R
2
5.76 (
***
)
(
***
), (
**
), and (
*
) indicate signi?cance levels of 1%, 5%, and 10%.
a
Variable de?nitions in Appendix C. This table reports our results of a test for the overall mediating e?ects of negotiation behavior on
buyer’s individual pro?t. Although this overall test provides an interesting addition to the three-step procedures listed in Tables 2–4, it is
not su?cient in itself to demonstrate mediated moderation (Muller et al., 2005). The table presents the results of two regression analyses.
First, the dependent variable buyer’s individual pro?t is regressed on cost information, on power, and on the moderation term power
*
cost
information. Second, the dependent variable buyer’s individual pro?t is regressed on cost information, on power, on the moderation term
power
*
cost information, on information exchange, on the moderation term cost information
*
information exchange, on distributive
behavior, and on the moderation term power
*
distributive behavior. Regression coe?cients are reported.
A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266 259
(coe?cient = 700.11, p < 0.01) implies that this neg-
ative relation between information exchange and
individual pro?t is only true for buyers lacking
TCO information. Buyers who have access to
TCO information and who exchange this informa-
tion earn signi?cantly higher individual pro?ts. Dis-
tributive behavior has a positive e?ect on individual
pro?t (coe?cient = 1188.89, p < 0.05), but this is
not the case for buyers with power (coe?-
cient = À1181.94, p < 0.05). In addition, the mod-
el’s R
2
increases signi?cantly, from 0.41 to 0.52 (F
for change in R
2
= 5.76, p < 0.01). Together, these
results provide support for our expectation that
the moderation e?ect on the outcome variable is
mediated by the process variables. As shown by
the signi?cant moderation e?ect in equation 1, the
performance disadvantage of less powerful buyers
is less pronounced when the buyer has detailed
TCO information. We also expect this moderation
e?ect to be mediated by the process variables: med-
iated moderation implies that the overall modera-
tion e?ect is reduced once the mediating process is
taken into account (Muller et al., 2005). Equation
2 clearly indicates that the moderation e?ect of cost
information and power becomes insigni?cant when
the process variables are added.
Supplementary analyses of the supplier’s negotiation
outcome and behavior
Table 1 also reports the results of the analyses of
the supplier’s individual pro?t and of the joint
pro?t. An ANOVA on the supplier’s individual pro?t
reveals main e?ects for power and cost information.
Suppliers facing less powerful buyers earned higher
individual pro?ts (F(1, 100) = 17.46, p < 0.01).
However, supplier pro?ts were lower when they
negotiated with buyers with TCO information
(F(1, 100) = 15.80, p < 0.01). Joint pro?ts were sig-
ni?cantly higher when the buyer possessed TCO
information (F(1, 100) = 15.78, p < 0.01). Together,
these results indicate that buyers with TCO infor-
mation and no outside option obtained high indi-
vidual pro?ts not because they knew ‘‘how to
fool” the (more powerful) suppliers, but because
these dyads realized higher joint pro?ts than dyads
in which the buyer had an outside option (mean
joint pro?t = 2446.15 versus mean joint
pro?t = 2336.67). Suppliers facing buyers without
an outside option earned more than their less pow-
erful opponents (buyer’s mean individual
pro?t = 1007.69 versus supplier’s mean individual
pro?t = 1438.46), re?ecting the power imbalance.
However, less powerful buyers with TCO informa-
tion were able to earn similar individual pro?ts as
buyers with an outside option and TCO informa-
tion, a result that we explained through the detailed
analysis of the buyers’ negotiation behavior.
Finally, in order to assess the e?ects of buyers’
behavior on suppliers’ behavior, we present results
(not tabulated) on the suppliers’ bargaining behav-
ior. Recall that the experimental manipulation for
the suppliers was the same in each of the experimen-
tal conditions: suppliers always had an outside
option and full cost information. Accordingly, the
di?erences in suppliers’ bargaining behavior across
the di?erent experimental cells can be explained by
their interaction with buyers. An ANOVA on the
suppliers’ information exchange shows a main e?ect
for cost information (F(1, 100) = 6.92, p < 0.01):
suppliers facing buyers with TCO information dis-
closed more information than suppliers facing buy-
ers with traditional information. We ?nd no e?ect of
the manipulations with respect to the buyers’ condi-
tions on the suppliers’ integrative behavior, nor on
the number of explicit requests for cooperation
issued by suppliers. Furthermore, we observe that
suppliers facing a powerful buyer used distributive
behavior more frequently than suppliers in the other
experimental conditions (F(1, 100) = 0.58, p < 0.10).
More speci?cally, we ?nd that suppliers facing a
powerful buyer with TCO information used more
distributive threats and referred to their outside
option more often than suppliers in the other exper-
imental conditions. Suppliers’ negotiation behavior
is also assessed by the composite measure problem
solving approach (based on the post-bargaining
questionnaires of the buyers). Main e?ects are
found for cost information (F(1, 100) = 4.83,
p < 0.05) and power (F(1, 100) = 3.18, p < 0.10).
Thus, compared to buyers with traditional cost
information, buyers with TCO information
reported that suppliers used more problem solving
techniques and fewer distributive bargaining tech-
niques. Furthermore, less powerful buyers felt that
their opponent employed more problem solving
techniques than buyers with an outside option.
Overall, these results provide evidence that buyers’
behavior strongly a?ects suppliers’ behavior.
Discussion and conclusion
In this paper, we examine the moderation e?ect
of cost information (TCO information versus
260 A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266
traditional cost information) on a buyer’s individual
pro?t at both high and low levels of buyer power.
Our results indicate that the availability of detailed
TCO information may alleviate the disadvantage
that dependent buyers face vis-a`-vis a more power-
ful supplier. This ?nding has important managerial
implications. On the one hand, it implies that less
powerful buyers can compensate for their power
disadvantage by gathering more detailed cost infor-
mation. On the other hand, powerful buyers do not
seem to be able to use this more detailed cost infor-
mation to enhance their power advantage in order
to obtain even higher individual pro?ts.
We explore the driver behind this result by exam-
ining the negotiation process. Consistent with
exchange theory and recent literature on informa-
tion processing, we expect that buyers with detailed
cost information and less power than their oppo-
nent try to (re)gain control over their own outcomes
by sharing information in an e?ort to create integra-
tive solutions. Because of their higher epistemic
motivation, less powerful buyers have a greater
incentive to engage in information acquisition and
processing (e.g., Fiske, 1993; Van Kleef et al.,
2006). Accordingly, we expect that less powerful
buyers with TCO information seek a more integra-
tive bargaining situation than powerful buyers,
who are more prone to using distributive bargaining
techniques because of their low epistemic motiva-
tion (e.g., De Dreu & Van Kleef, 2004; Fiske,
1993; Weingart et al., 1996).
Support for these conjectures is found in our fol-
low-up analyses, in which we examine whether the
moderation e?ect of TCO information on the rela-
tion between power and individual pro?t can be
explained by the choice of negotiation strategy.
Using Baron and Kenny’s (1986) framework for
combining mediation and moderation, we ?nd that
the moderation e?ect of cost information on the
relation between power and individual pro?t is med-
iated by buyers’ bargaining behavior. Analysis of
the buyers’ communications suggests that buyers
with TCO information disclose more information
than buyers with traditional cost information. Con-
sistent with prior research (e.g., Perdue & Summer,
1991), we ?nd that more powerful buyers disclose
less information and use more distributive bargain-
ing techniques than less powerful buyers. The
results also reveal an unexpected positive e?ect of
TCO information on distributive behavior. How-
ever, this e?ect can be explained by the ?nding that
buyers with TCO information and power tend to
use distributive techniques. Furthermore, a limited
analysis of suppliers’ bargaining behavior revealed
its reciprocal nature. Suppliers facing buyers with
TCO information disclosed more information than
suppliers facing buyers with traditional information.
It also emerged that suppliers facing powerful buy-
ers with TCO information used more distributive
bargaining techniques than suppliers in the other
experimental conditions. As the experimental
manipulation for the suppliers is the same in each
of the experimental conditions, the di?erences in
suppliers’ bargaining behavior across the di?erent
experimental cells can only be explained by their
interaction with the buyers. These ?ndings provide
support for our conjecture that less powerful buyers
are able to create a cooperative relationship, into
which the supplier is willing to enter. This resulted
not only in higher individual pro?ts, but also in
higher joint pro?ts.
From these results, we conclude that the manip-
ulation of power and cost information resulted in
buyers using di?erent negotiation techniques. Less
powerful buyers who have access to TCO data
are more likely to resort to problem solving tech-
niques, whereas powerful buyers tend to rely on
distributive bargaining techniques. Particularly
interesting is that the problem solving strategy
adopted by less powerful buyers with TCO infor-
mation seems to be e?ective, whereas the distribu-
tive bargaining strategy adopted by more
powerful buyers with TCO results in less informa-
tion sharing and in less e?ective negotiation out-
comes. Less powerful buyers seem to be
motivated to create a cooperative and coordinated
relationship, in which the supplier is willing to con-
sider the buyer’s objectives. When a buyer shares
information about needs and preferences and/or
makes concessions that facilitate the development
of a solution, the supplier is likely to reciprocate
(e.g., Campbell et al., 1988; Dekker, 2004). This
may explain the higher individual pro?ts for less
powerful buyers with TCO information. Less pow-
erful buyers lacking TCO information are less able
or willing to communicate the right information
and therefore to create integrative situations. More
powerful buyers, on the other hand, believe to have
power and choose a distributive bargaining strategy
that, in a tit-for-tat move, the supplier responds to
by also resorting to a distributive bargaining strat-
egy. Thus, a powerful buyer’s distributive bargain-
ing strategy may not be e?ective when he is facing
an equally powerful supplier.
A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266 261
Overall, our results suggest that powerful buyers
with re?ned TCO information may not maximize
possible bene?ts from buyer–supplier interactions
due to their bargaining strategy. Their (false) feeling
of power causes an increase in distributive bargain-
ing techniques and a reluctance to share the infor-
mation necessary for inter?rm cost minimization.
This implies that powerful buyers may only bene?t
from more re?ned accounting information systems
if these ?rms undertake e?orts that encourage infor-
mation sharing and discourage distributive bargain-
ing strategies. Future research should examine
whether, and under what conditions, inter?rm con-
trol systems and incentive systems may motivate
buyers and suppliers to share cost information in
such a way that supply chain performance is
improved.
We conclude with remarks on several limitations
of this study and further directions for future
research. First, while the experimental context
induced by a simple negotiation game allows us to
maintain control over exogenous variables, the
scope for generalizing the results is somewhat lim-
ited. Other factors, such as the incentive system,
past negotiation history, and future negotiation
probabilities, have been shown to impact negotiated
outcomes but are not included in the present analy-
sis. Future research is needed to determine the sen-
sitivity of our results to several parameters excluded
in the current study.
Second, this study manipulates the experimental
conditions of the buyers, but not the suppliers, that
is, the suppliers’ experimental conditions were held
constant. In particular, suppliers were always fully
informed and powerful. As a consequence, our con-
clusions do not generalize beyond negotiation set-
tings in which the supplier is always at least as
powerful as the buyer. Further research can modify
these experimental conditions and examine the role
of TCO information from both buyer and supplier
perspectives.
Third, although optimal joint outcomes are
introduced, our study focuses primarily on buyer
outcomes. This focus underplays the cost/bene?t
trade-o? of obtaining the additional information
needed for TCO. Since TCO information is not
costless, it would be interesting to consider whether
buyers would be willing to incur the cost of obtain-
ing TCO information if it is e?cient to do so.
Fourth, our manipulation of TCO information is
obviously a simpli?cation of reality, providing many
avenues for further research. In this paper, we take a
deterministic approach to TCO information, which
allows us to study the impact of cost information
on negotiation behavior and outcomes. Our aim
was not to describe a value chain and ABC analysis,
nor to discuss data collection mechanisms and chal-
lenges. Neither did we seek to identify and categorize
important TCO cost drivers. However, more
research is needed on these issues to e?ectively iden-
tify critical cost drivers for estimating TCO. Future
research may consider aspects such as the manner
in which buyers collect TCOdata (e.g., fromthe sup-
plier, benchmarking, past experience), the type of
TCOsystems implemented (e.g., formal vs. informal,
standardized vs. unique models), and how these
design aspects impact buyer–supplier relations and
negotiations. In addition, examining the e?ects of
imperfect TCO information would also be interest-
ing. In this study, perfect TCO information was pro-
vided to decision-makers. In reality, TCO
information is characterized by mistakes and simpli-
?cations, which may have important implications on
the negotiation process and its outcomes. Similarly,
it would be interesting to consider how costs that are
di?cult to measure (e.g., opportunity costs of
reduced sales, productivity losses due to downtime)
can be included in TCO systems. Because of the dif-
?culty of measuring these costs reliably, they are typ-
ically not recorded in a buyer’s accounting system.
However, they may represent a substantial cost in
total costs. An investigation of how the negotiation
process and outcomes are a?ected when less objec-
tive or less reliable cost estimates are explicitly
included in the TCO system would add useful
insights to the literature. Finally, further research
could consider whether audited TCO information
(either by a third party or by the partner) would be
considered more reliable than unaudited TCO infor-
mation, and if so, what the implications for the
buyer–supplier relation would be.
Acknowledgements
The comments of Shannon Anderson, Eddy Car-
dinaels, Henri Dekker, Luc Sels, the participants at
the BAA conference in Edinburgh, the EAA confer-
ence in Gothenburg, the AOS-Bocconi conference
in Milan, the MAS conference in Clearwater Beach,
and two anonymous reviewers are gratefully
acknowledged, as is the research funding provided
by the Fund for Scienti?c Research – Flanders (Pro-
ject G.0261.00) and by the National Bank of Bel-
gium (Project NB/05/01).
262 A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266
Appendix A. Experimental cells and corresponding cost tables
Power
Low power (no outside option) Equal power (outside option)
Cost information
TCO information Cell 1 (n = 30) Cell 2 (n = 26)
Traditional cost information Cell 3 (n = 26) Cell 4 (n = 22)
Cost table for the supplier (in cell 1, 2, 3 and 4)
a
Price (=income) Maintenance Spare parts
Contract 1 600 Contract A
*
1350 Contract R 2250
Contract 2 1200 Contract B 1200 Contract S 2000
Contract 3 1800 Contract C 1050 Contract T 1750
Contract 4 2400 Contract D 900 Contract U 1500
Contract 5
*
3000 Contract E 750 Contract V 1250
Contract 6 3600 Contract F 600 Contract W 1000
Contract 7 4200 Contract G 450 Contract X 750
Contract 8 4800 Contract H 300 Contract Y 500
Contract 9 5400 Contract I 150 Contract Z
*
250
a
The tables were constructed such that the minimum (À3000 Euro) and maximum pro?ts (5000 Euro) that buyers and suppliers could
earn were the same. Pareto-optimal joint outcomes correspond to the buyer and the supplier agreeing to contract 5AZ. This situation is
denoted with an asterisk (
*
).
Cost table for the buyer with TCO information (in cell 1 and 2):
Income = 6000
Price (=cost) Maintenance Spare parts
Contract 1 600 Contract A
*
250 Contract R 150
Contract 2 1200 Contract B 500 Contract S 300
Contract 3 1800 Contract C 750 Contract T 450
Contract 4 2400 Contract D 1000 Contract U 600
Contract 5
*
3000 Contract E 1250 Contract V 750
Contract 6 3600 Contract F 1500 Contract W 900
Contract 7 4200 Contract G 1750 Contract X 1050
Contract 8 4800 Contract H 2000 Contract Y 1200
Contract 9 5400 Contract I 2250 Contract Z
*
1350
Cost table for the buyer with traditional cost information (in cell 3 and 4):
Income = 6000
Price (=cost) Maintenance
a
Spare parts
b
Contract 1 600 Contract A
*
1 Contract R 3
Contract 2 1200 Contract B 2 Contract S 6
Contract 3 1800 Contract C 3 Contract T 9
Contract 4 2400 Contract D 4 Contract U 12
Contract 5
*
3000 Contract E 5 Contract V 15
Contract 6 3600 Contract F 6 Contract W 18
Contract 7 4200 Contract G 7 Contract X 21
Contract 8 4800 Contract H 8 Contract Y 24
Contract 9 5400 Contract I 9 Contract Z
*
27
a
Number of maintenance sessions performed by the buyer each month.
b
Spare parts procured by the buyer from a third party each month.
A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266 263
Appendix B. Measuring negotiation behavior
B.1. Behavioral coding categories
Category Examples
Information exchange – Maintenance is more expensive for my company than spare parts
– Maintenance contracts start at €250 (=contract A) and increase by €250 until €2250 (=contract I); con-
tracts for spare parts start at €150 (= contract R) and increase by €150 until €1350 (=contract Z)
Rewards – I am pleased with the concessions made thus far
Positive normative appeals – Your o?ers have been fair and equitable
Request for cooperation – Let us cooperate
General threats – Make a concession or you will be in trouble
Exit threats – Respond with a concession or I will call another supplier
Punishment – This negotiation is going nowhere
Warnings – My company has a policy against uncooperative suppliers
Positional commitment – I refuse to concede any further
– I refuse to drop my price below the present level
B.2. Buyers’ problem solving approach
Observed ratings from suppliers’ questionnaires (items were reverse coded)
Do you feel that the person with whom you were paired was more interested in solving your mutual
problem, or more self-interested?
1 2 3 4 5
Solving a mutual problem Self-interested
Rate your partner’s bargaining strategies on the following scales:
1 2 3 4 5
Accommodating Exploitative
1 2 3 4 5
Honest Deceptive
1 2 3 4 5
Information sharing Withholding
information
Appendix C. Bivariate correlation matrix (n = 104)
1 2 3 4 5 6 7 8 9 10
1 Buyer’s individual pro?t
a
1
2 Supplier’s individual pro?t
b
À0.71 (
**
) 1
3 Joint pro?t
c
0.54 (
**
) 0.21 (
*
) 1
4 Information exchange
d
0.14 0.19 0.41 (
**
) 1
5 Integrative behavior
e
0.06 À0.02 0.05 0.02 1
6 Requests for cooperation
f
0.15 0.05 0.27 (
**
) 0.30 (
**
) 0.36 (
**
) 1
7 Distributive behavior
g
0.22 (
*
) À0.20 (
*
) 0.07 0.09 À0.11 À0.11 1
8 Problem solving approach
h
0.41 (
**
) À0.10 0.44 (
**
) 0.35 (
**
) 0.16 0.38 (
**
) 0.09 1
9 Cost information
i
0.55 (
**
) À0.34 (
**
) 0.36 (
**
) 0.23 (
*
) 0.13 0.23 (
**
) 0.23 (
*
) 0.82 (
**
) 1
10 Power
j
0.27 (
**
) À0.35 (
**
) À0.04 À0.20 (
*
) À0.21 (
*
) À0.25 (
**
) 0.35 (
**
) À0.23 (
*
) 0.00 1
(
*
) Correlation is signi?cant at the 0.05 level (2-tailed).
(
**
) Correlation is signi?cant at the 0.01 level (2-tailed).
a
Buyer’s individual pro?t at the end of the game (based on the cost table provided in Appendix A).
b
Supplier’s individual pro?t at the end of the game (based on the cost table provided in Appendix A).
c
Joint pro?t is the sum of buyer’s individual pro?t and supplier’s individual pro?t at the end of the game.
d
Information exchange is coded ‘‘0” for participants not revealing any information about their cost structure, ‘‘1” for participants
revealing the relative importance (but not the numerical values) of each of the three issues under negotiation, and ‘‘2” for participants
revealing the relative importance and the numerical values of each of the three issues under negotiation.
264 A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266
e
Coded from the messages sent by buyers (based on the behavioral coding scheme in Appendix B1). Summated scale determined by
adding up two integrative behavior types (rewards and positive normative appeals) and solving for the mean value. Cronbach’s
alpha = 0.70.
f
Coded from the messages sent by buyers (based on the behavioral coding scheme in Appendix B1). Single scale item (requests for
cooperation).
g
Coded from the messages sent by buyers (based on the behavioral coding scheme in Appendix B1). Summated scale determined by
adding ?ve distributive behavior types (general threats, exit threats, positional commitment, punishments, warnings) together and solving
for the mean value. Cronbach’s alpha = 0.81.
h
Buyer’s problem solving approach based on observed ratings from supplier’s questionnaire. Summated scale determined by adding the
four post-bargaining questionnaire items included in Appendix B2 together and solving for the mean value. Cronbach’s alpha = 0.88.
i
Dummy variable, experimental manipulation: 0 for traditional cost information, 1 for TCO information.
j
Dummy variable, experimental manipulation: 0 for buyers with no outside option, 1 for buyers with an outside option.
References
Anderson, S. W., & Dekker, H. C. (2005). Management
control for market transactions: The relation between
transaction characteristics, incomplete contract design and
subsequent performance. Management Science, 51,
1734–1752.
Baiman, S., & Rajan, M. V. (2002). Incentive issues in inter-?rm
relationships. Accounting, Organizations and Society, 27,
213–238.
Bame-Aldred, C. W., & Kida, T. (2007). A comparison of auditor
and client initial negotiation positions and tactics. Accounting,
Organizations and Society, 32, 497–511.
Baron, R., & Kenny, D. (1986). The moderator–mediator
variable distinction in social psychological research: Concep-
tual, strategic, and statistical considerations. Journal of
Personality and Social Psychology, 51, 1173–1182.
Bazerman, M. H., Curhan, J. R., Moore, D. A., & Valley, K. L.
(2000). Negotiation. Annual Review of Psychology, 51,
279–314.
Bottom, W. P., Holloway, J., Miller, G. J., Mislin, A., &
Whitford, A. (2006). Building a pathway to cooperation:
Negotiation and social exchange between principal and agent.
Administrative Science Quarterly, 51, 29–58.
Campbell, N. C. G., Graham, J. L., Jolibert, A., & Meissner, H.
G. (1988). Marketing negotiations in France, Germany, the
United Kingdom, and the United States. Journal of Market-
ing, 52, 49–62.
Carr, L. P., & Ittner, C. D. (1992). Measuring the cost of
ownership. Journal of Cost Management, 6, 7–13.
Craft, J. A. (1981). Information disclosure and the role of the
accountant in collective bargaining. Accounting, Organiza-
tions and Society, 6, 97–107.
De Dreu, C. K. W., Beersma, B., Stroebe, K., & Euwema, M. C.
(2006). Motivated information processing, strategic choice,
and the quality of negotiated agreement. Journal of Person-
ality and Social Psychology, 90, 927–943.
De Dreu, C. K. W., & Carnevale, P. (2003). Motivational bases
of information processing and strategy in con?ict and
negotiation. Advances in Experimental Social Psychology,
35, 235–291.
De Dreu, C. K. W., & Van Kleef, G. A. (2004). The in?uence of
power on the information search, impression formation, and
demands in negotiation. Journal of Experimental Social
Psychology, 40, 303–319.
Degraeve, Z., & Roodhooft, F. (1999). E?ectively selecting
suppliers using total cost of ownership. Journal of Supply
Chain Management, 35, 5–10.
Dekker, H. C. (2003). Value chain analysis in inter?rm relation-
ships: A ?eld study. Management Accounting Research, 14,
1–23.
Dekker, H. C. (2004). Control of inter-organizational relation-
ships: Evidence on appropriation concerns and coordination
requirements. Accounting, Organizations and Society, 29,
27–49.
Drake, A., & Haka, S. (2008). Does ABC exacerbate the hold-up
problem between buyers and suppliers?. The Accounting
Review, 83, 29–60.
Ellram, L. M. (1995). Activity based costing and total cost of
ownership: A critical linkage. Journal of Cost Management, 9,
22–30.
Emerson, R. M. (1976). Social exchange theory. Annual Review of
Sociology, 2, 335–362.
Fisher, J., Frederickson, J. R., & Pe?er, S. A. (2000). Budgeting:
An experimental investigation of the e?ects of negotiation.
The Accounting Review, 75, 93–114.
Fisher, J., Frederickson, J. R., & Pe?er, S. A. (2006). Budget
negotiations in multi-period settings. Accounting, Organiza-
tions and Society, 31, 511–528.
Fiske, S. T. (1993). Controlling other people: The impact of
power on stereotyping. American Psychologist, 48, 621–628.
Gelfand, M., & Christakopoulou, S. (1999). Culture and nego-
tiation cognition: Judgment accuracy and negotiation pro-
cesses in individualistic and collectivistic cultures.
Organizational Behavior and Human Decision Processes, 79,
248–269.
Gibbins, M., McCracken, S. A., & Salterio, S. E. (2005).
Negotiations over accounting issues: The congruency of audit
partner and chief ?nancial o?cer recalls. Auditing: A Journal
of Practice and Theory, 24, 171–193.
Gibbins, M., Salterio, S. E., & Webb, A. (2001). Evidence about
auditor–client management negotiation concerning client’s
?nancial reporting. Journal of Accounting Research, 39,
553–563.
Giebels, E., De Dreu, C. K. W., & Van de Vliert, E. (2000).
Interdependence in negotiation: E?ects of exit options and
social motive on distributive and integrative negotiation.
European Journal of Social Psychology, 30, 255–272.
Goodwin, S. A., Gubin, A., Fiske, S. T., & Yzerbyt, V. Y. (2000).
Power can bias impression processes: Stereotyping subordi-
A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266 265
nates by default and by design. Group Processes and
Intergroup Relations, 3, 227–256.
Ittner, C. D., Larcker, D. F., & Meyer, M. W. (2003). Subjectivity
and the weighting of performance measures: Evidence from
the balanced scorecard. The Accounting Review, 78, 725–758.
Kachelmeier, S. J., & Towry, K. L. (2002). Negotiated transfer
pricing: Is fairness easier said than done?. The Accounting
Review, 77, 571–593.
Kadous, K., Koonce, L., & Towry, K. L. (2005). Quanti?cation
and persuasion in managerial judgment. Contemporary
Accounting Research, 22, 643–686.
Kelley, H. H. (1966). A classroom study of the dilemmas in
interpersonal negotiation. In K. Archibald (Ed.), Strategic
interaction and con?ict. Berkerley, California: Institute of
International Studies.
Kelley, H. H., & Thibaut, J. (1978). Interpersonal relations: A
theory of interdependence. New York: Wiley.
Keltner, D., Gruenfeld, D. H., & Anderson, C. (2003). Power,
approach, and inhibition. Psychological Review, 110, 265–284.
Kersten, G. (2001). Modeling distributive and integrative nego-
tiations: Reviewed and revised characterization. Group Deci-
sion and Negotiation, 10, 493–514.
Luft, J., & Libby, R. (1997). Pro?t comparisons, market prices
and managers’ judgments about negotiated transfer prices.
The Accounting Review, 72, 217–229.
Luft, J., & Shields, M. D. (2003). Mapping management
accounting: Graphics and guidelines for theory-consistent
empirical research. Accounting, Organizations and Society, 28,
169–249.
Muller, D., Judd, C. M., & Yzerbyt, V. Y. (2005). When
moderation is mediated and mediation is moderated. Journal
of Personality and Social Psychology, 89, 852–863.
Murnigham, J., & Bazerman, M. (1990). A perspective on
negotiation research in accounting and auditing. The Account-
ing Review, 65, 642–657.
Ng, T. B-P., & Tan, H-T. (2003). E?ects of authoritative guidance
availability and audit committee e?ectiveness on auditors’
judgment in an auditor–client negotiation context. The
Accounting Review, 78, 801–818.
Northcraft, G. B., & Neale, M. A. (1986). Opportunity costs and
the framing of resource allocation decisions. Organizational
Behavior and Human Decision Processes, 37, 348–356.
Perdue, B. C., & Summer, J. O. (1991). Purchasing agents’ use of
negotiation strategies. Journal of Marketing Research, 28,
175–189.
Pinkley, R. L., Neale, M. A., & Bennett, R. J. (1994). The impact
of alternatives to settlement in dyadic negotiation. Organiza-
tional Behavior and Human Decision Processes, 57, 97–116.
Porter, M. E. (1985). Competitive advantage: Creating and
sustaining superior performance. New York: The Free Press.
Putman, L. L., & Fairhurst, G. T. (2001). Discourse analysis in
organization. In F. Jablin & L. Putman (Eds.), The New
Handbook of Organizational Communication. Thousand Oaks,
CA: Sage.
Roodhooft, F., Hiel, G., Van den Abbeele, A., & van Doveren, B.
(2003). Optimised sourcing strategies using total cost of
ownership. Cost Management, 17, 28–35.
Roodhooft, F., Van den Abbeele, A., & Peeters, F. (2005).
Calculating the total cost of ownership of utilities: A case of
inter-?rm cost management. Cost Management, 19, 13–23.
Shank, J. K. (1989). Strategic cost management: New wine, or
just new bottles?. Management Accounting Research 1, 47–65.
Shank, J. K., & Govindarajan, V. (1992). Strategic cost manage-
ment: The value chain perspective. Journal of Management
Accounting Research, 4, 179–197.
Thibaut, J. W., & Kelley, H. H. (1959). The social psychology of
groups. NY: John Wiley & Sons.
Tomkins, C. (2001). Interdependencies, trust and information in
relationships, alliances and networks. Accounting, Organiza-
tions and Society, 26, 161–191.
Van Kleef, G. A., De Dreu, C. W., Pietroni, D., & Manstead, A.
S. R. (2006). Power and emotion in negotiation: Power
moderates the interpersonal e?ects of anger and happiness
and concession making. European Journal of Social Psychol-
ogy, 36, 557–581.
Walton, R. E., & McKersie, R. B. (1966). Behavioral dilemmas in
mixed-motive decision making. Behavioral Science, 11,
370–384.
Waterhouse, J., Gibbins, M., & Richardson, A. J. (1993).
Strategic ?nancial disclosure: Evidence from labor negotia-
tions. Contemporary Accounting Research, 9, 526–550.
Weingart, L. R., Hyder, E. B., & Prietula, M. J. (1996).
Knowledge matters: The e?ect of tactical descriptions on
negotiation behavior and outcome. Journal of Personality and
Social Psychology, 70, 1204–1217.
Windsor, C. A., & Ashkanasy, N. M. (1995). The e?ect of client
management bargaining power, moral reasoning develop-
ment, and belief in a just world on auditor independence.
Accounting, Organizations and Society, 20, 701–720.
Wolfe, R., & McGinn, K. L. (2005). Perceived, relative power
and its in?uence on negotiations. Group Decisions and
Negotiation, 14, 3–20.
Wouters, M., Anderson, J. C., & Wynstra, F. (2005). The
adoption of total cost of ownership for sourcing decisions – A
structural equations analysis. Accounting, Organizations and
Society, 30, 167–191.
266 A. Van den Abbeele et al. / Accounting, Organizations and Society 34 (2009) 245–266
doc_110177264.pdf