Management control systems as interorganizational trust builders in evolving relationship

Description
Research on inter-organizational relationships argues that at mature stages, when trust has reached a high level, it
will be damaged by new management control systems (MCSs). This longitudinal case study provides evidence to the
contrary: in an open-ended and evolving relationship, even when trust is well established, MCSs can build it. High trust
provides a platform where success encourages the partners to cooperate further, demanding, in turn, more MCSs and
greater levels of trust to support cooperation.

Management control systems as inter-organizational
trust builders in evolving relationships: Evidence from
a longitudinal case study
q
Maria L. Ve´lez
a,
*
, Jose´ M. Sa´nchez
b,1
, Concha A
´
lvarez-Dardet
b,2
a
Universidad de Ca´ diz, Departamento Econom?´a de la Empresa, Gta. Carlos Cano S/N, 11002 Ca´ diz, Spain
b
Universidad Pablo de Olavide, Departamento Direccio´ n de Empresas, Ctra. Utrera, Km.1, 41013 Sevilla, Spain
Abstract
Research on inter-organizational relationships argues that at mature stages, when trust has reached a high level, it
will be damaged by new management control systems (MCSs). This longitudinal case study provides evidence to the
contrary: in an open-ended and evolving relationship, even when trust is well established, MCSs can build it. High trust
provides a platform where success encourages the partners to cooperate further, demanding, in turn, more MCSs and
greater levels of trust to support cooperation. By providing evidence with a greater appearance of objectivity than infor-
mal controls can yield, action and result controls improve partners’ perception of each other’s trustworthiness, and
build competence and goodwill-based trust. MCSs are used not only to supervise but also to coordinate, and this sec-
ond, more salient function avoids possible suspicions that could damage trust.
Ó 2008 Elsevier Ltd. All rights reserved.
One key for maintaining an inter-organizational
relationship (IOR), or preventing its failure, is that
partners have adequate con?dence in each other’s
cooperation (Das & Teng, 1998). In IORs, as col-
laborative arrangements to gain competitive
advantages (Coletti, Sedatole, & Towry, 2005),
?rms tend to have greater con?dence when they
perceive a suitable level of control over their part-
ners (Sohn, 1994) and when they trust each other
(Das & Teng, 2001; Ring & Van de Ven, 1992).
But the relationship between management control
systems (MCSs) and trust is complex and open to
debate (Coletti et al., 2005; Dekker, 2004; Tom-
kins, 2001; Van der Meer-Kooistra & Vosselman,
2006).
0361-3682/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.aos.2008.02.006
q
This research is partly ?nanced by the Research Projects
SEJ-366, SEJ-111 and SEJ2005-06652/ECON.
*
Corresponding author. Tel.: +34 956 01 54 35; fax: +34 956
01 53 86.
E-mail addresses: [email protected] (M.L. Ve´lez), josesan-
[email protected] (J.M. Sa´nchez), [email protected] (C. A
´
lvarez-Dar-
det).
1
Tel.: +34 954 34 98 54; fax: +34 954 34 83 53.
2
Tel.: +34 954 34 93 57; fax: +34 954 34 83 53.
Available online at www.sciencedirect.com
Accounting, Organizations and Society 33 (2008) 968–994
www.elsevier.com/locate/aos
As Knights, Noble, Vurdubakis, and Willmott
(2001) pointed out, there is a long tradition that
conceptualises MCSs and trust as alternatives. In
opposition to this idea, Das and Teng (1998) main-
tained that the introduction of an MCS does not
necessarily suppose a lowering of trust. Tomkins
(2001) argued that conceptualising MCSs and trust
as either substitutive or complementary re?ects a
static analysis that ignores the dynamic process
of building trust. He maintained that the in?uence
of MCSs on trust is not uniform, and is likely to be
characterised over the life cycle of a relationship
by an inverted U-shape. Thus, only in the IOR’s
early stages do MCSs have a positive association
with trust. In the later stages, when trust has
reached a higher level, the introduction of new
MCSs can cause harm.
When the IOR’s duration is limited to one spe-
ci?c project or venture, it is probably true to say
that as trust intensity becomes established at
higher levels the successful development of the
associated activity needs less control to sustain
that relationship, as Tomkins (2001, p. 170)
claimed. However, in open-ended IORs, the stabil-
ity of mature stages could provide a platform for
continuous evolution (Halinen, Salmi, & Havila,
1999). When the IOR’s feedback loops are positive
(Arin˜ o & de la Torre, 1998), trust will encourage
the partners to enlarge the collaboration (Inkpen
& Curral, 2004). Encouraged by favourable
results, the parties may expand the scope or com-
plexity of their activities (Doz, 1996), involving
more resources and time, and in turn increasing
their interdependence. This evolution could a?ect
the information needed for control of the IOR
and the perceived trust.
The aim of our paper is to increase knowledge
on the e?ect of MCSs development on already
established trust in mature stages, taking into con-
sideration the evolving nature of an open-ended
IOR. To grasp the complexity and dynamism of
IORs (Arin˜ o & de la Torre, 1998; Dekker, 2004),
we adopted a longitudinal case study approach
(Yin, 1984). We chose the long-standing and suc-
cessful relationship between a manufacturing ?rm
called CMD (a pseudonym) and its distribution
channel. Between 1997 and 2004, CMD, as a part
of its strategy to extend and control its degree of
externalisation, gradually introduced, to the entire
channel, various MCS tools that now make up a
management control system shared between
CMD and the channel. We study this process from
both perspectives, that of the manufacturer and
that of the distribution channel, and analyse a ser-
ies of events that occurred during the course of the
IOR, in order to explain the impact that MCS
development had on trust.
Our paper makes several contributions to the
existing literature in order to o?er a new point of
view on the complex association between MCSs
and trust. First, advancing on Tomkins thesis,
we observe the association between MCSs and
trust in a mature and open-ended IOR and address
the evolving nature of IORs. This case study pro-
vides evidence that, even when trust is well estab-
lished, MCSs foster conditions that favour and
build trust. The constant evolution of the IOR in
interdependence, scope, and complexity requires
greater con?dence in cooperation and, in turn,
demands new MCSs and greater levels of trust.
Second, given Sako’s (1992) proposal that compe-
tence-based and goodwill-based trust will be cre-
ated and maintained in di?erent ways, we
analyse whether action, result, and personnel-cul-
tural controls (Merchant & Van der Stede, 2003)
have di?erent e?ects on trust. Whereas Das and
Teng (2001) and Inkpen and Curral (2004) argued
that only informal controls can build trust, our
?ndings suggest that action and result controls
also build trust, by providing evidence with a
greater appearance of objectivity. Formal MCSs
do not necessarily cause in?exibility; instead, such
systems can improve agents’ independence and
increase their opportunities to demonstrate their
competences inside the established limits. Further-
more, because both parties work with the same
system, formal MCSs improve their perception of
each other as trustworthy. Third, in contrast to
most studies (Andaleeb, 1995; Inkpen & Curral,
2004; Ring & Van de Ven, 1994), which focus on
the monitoring function of MCS, even to the
extent of confusing it with the whole function
(e.g., see Coletti et al., 2005), we include the less-
studied coordination function of MCSs (Dekker,
2004; Tomkins, 2001). Our ?ndings show that
both parties use MCSs to coordinate the new com-
M.L. Ve´ lez et al. / Accounting, Organizations and Society 33 (2008) 968–994 969
plex tasks, and this use creates trust. The leader
?rm, ‘‘giving in order to receive”, keeps end mar-
ket control, improving joint results and satisfac-
tion. Although the manufacturer uses the same
MCS tools both to coordinate and to supervise,
because the tools are shared, the channel members
perceive much more coordination than monitor-
ing, and this avoids suspicions that could damage
trust.
In the next section, we consider di?erent de?ni-
tions of trust, its twofold dimensions, and the pre-
cursors of its increase. Second, we review the
literature that analyses the e?ects of MCSs on trust
in order to establish our research questions. We
present a longitudinal case study of how CMD
introduced MCS tools into its distribution channel
and the impact of these tools on trust. We o?er an
analysis in the discussion section, and end with
conclusions and directions for further research.
Inter-organizational trust: the concept and the
precursors of its evolution
Although trust has been de?ned in di?erent
ways, Moorman, Deshpande´, and Zaltman
(1992) recognized two general conceptions of trust
in the business literature: (1) trust as a belief or
expectation about a trustee and (2) trust as a
behavioural intention that re?ects a reliance on
the trustee and involves vulnerability and uncer-
tainty on the part of the trustor. Rousseau, Sitkin,
Burt, and Camerer (1998) concluded that ‘‘con?-
dent expectations” and a ‘‘willingness to be vulner-
able” are critical components of most de?nitions
of trust. Using Mayer, Davis, and Schoorman’s
(1995) de?nition, we understand trust as the will-
ingness of a party to be vulnerable to the actions
of another party on the expectation that the other
party will perform a particular action which is
important to the trustor.
Since trust is an inherent characteristic of
human nature (Mayer et al., 1995), it lends itself
to multiple levels of analysis. The origin of trust,
the trustor, is of course human, but may be either
an individual or a group; the object of trust, the
trustee, may be things as well as people and their
systems (Luhmann, 1979). Various authors (Cur-
ral & Inkpen, 2002; Das & Teng, 2001; Ganesan
& Hess, 1997) have treated trust across organiza-
tional boundaries as having multiple levels: (1)
interpersonal trust, in which individuals in organi-
zation A trust individuals in organization B; (2)
organizational trust, in which an individual in
organization A trusts organization B as a whole;
and (3) inter-organizational trust, in which mem-
bers of organization A have a collectively held
trust in organization B.
‘‘Trust . . . depends on time and context” (Noote-
boom, Berger, & Noorderhaven, 1997, p. 314). As
partners and partnership managers learn about
each other, the level of trust will change, which
means that trust should be viewed as dynamic
rather than static (Inkpen & Curral, 2004), and
its evolution should depend on perceptions about
distinct trustee characteristics (Mayer et al.,
1995). Accordingly, the literature is increasingly
conceptualising trust in multidimensional terms
(Das & Teng, 2001; Sako, 1992). Although precise
de?nitions and examples are beyond the scope of
this paper, keeping in mind our aim, we choose a
typology that describes the base of these expecta-
tions: (a) an expectancy that the other party’s
word can be relied on; and (b) a belief in the
motives of the other party. Sako (1992) catego-
rized three basic types of trust: contract-based
trust in written or oral agreements that are backed
by a mutual, universalistic agreement on ethical
codes; competence-based trust, the expectation of
a trading partner’s technical and managerial com-
petence; and goodwill-based trust, or somewhat
di?use, abstract expectations of open commitment
to each other. Therefore, as Sako (1992) argues,
trust can relate to distinct trustee characteristics,
among which, in IORs, a partner’s goodwill and
capabilities are of particular importance. Indeed,
Ganesan (1994) argued that competence-based
trust includes the contractual dimension. Various
scholars (Das & Teng, 1998, 2001; Dekker, 2004;
Ganesan, 1994) have connected competence-based
trust to relatively objective perceptions about part-
ners’ technical capabilities, skills, know-how, cred-
ibility, and reliability, and goodwill-based trust to
subjective expectations regarding the partner’s
moral responsibility, benevolence, and non-oppor-
tunistic behaviour in unforeseen situations.
970 M.L. Ve´ lez et al. / Accounting, Organizations and Society 33 (2008) 968–994
It is therefore necessary to understand how
trust can be enhanced. Sako (1992) argued that
each dimension of trust will be di?erently created
and maintained, though with some degree of inter-
action. Accordingly, the precursors of trust pro-
posed in previous research can be classi?ed
according to their e?ect on each dimension:
– Competence-based trust precursors may be
divided into technological, economic, and part-
nering competencies. According to Sako (1992),
?rms may generate competence-based trust
through investments that allow them to transfer
their knowledge and technology to their part-
ners. Reliability can be induced through educa-
tion (Sako, 1992). One competence-building
mechanism is proactive information collection
(Das & Teng, 2001). Other factors that build
trust are satisfaction with the relationship,
reaching acceptable levels of results (Kumar,
Scheer, & Steenkamp, 1995; Selnes, 1998), the
professionalism of the counterpart, the ability
to perform, and (large) size of the partner ?rm
(Doney & Cannon, 1997).
– Goodwill-based trust precursors include the
establishment of mutual objectives and interests
(Anderson & Weitz, 1989; Das & Teng, 2001); a
common system of values and norms, together
with ties of friendship in a context of reciprocal
interchange (Sako, 1992); investment in
resources dedicated to the partner (Sako &
Helper, 1998); willingness to provide comple-
mentary services and undertake investment
(Anderson & Weitz, 1989; Doney & Cannon,
1997); technical assistance (Sako, 1992); and
IOR formalization and participation in deci-
sion-making (Dwyer & Oh, 1987). Bradach
and Eccles (1989) stated that the creation of
trust could be accelerated by regular contacts
(see also Das & Teng, 2001; Lang?eld-Smith
& Smith, 2003). Morgan and Hunt (1994) and
Sako and Helper (1998) have stressed a two-
way information ?ow, as one-way demands
for information evoke suspicion of opportun-
ism. The partner’s reputation also relates to
goodwill-based trust (Anderson & Weitz,
1989). Managerial style, openness, and the per-
ception of being treated with fairness, reciproc-
ity, and equity nurture trust in the partner (Das
& Teng, 1998; Kumar et al., 1995; Ring & Van
de Ven, 1992).
Management control systems’ e?ects
on inter-organizational trust
In an IOR, MCSs are used to create conditions
that motivate the partners to achieve desirable or
predetermined outcomes (Dekker, 2004). Trans-
lating Merchant and Van der Stede’s (2003, p.
4) de?nition to an IOR setting, we adopt a
broader vision of MCSs that comprises the vari-
ous policies and procedures (Coletti et al., 2005)
used to ensure that the partners’ behaviour and
decisions are consistent with the IOR objectives
and strategies.
MCSs and trust are both sources of con?dence
that a partner will cooperate (Das & Teng, 1998).
According to Das and Teng (1998), con?dence in
partner cooperation is de?ned as a ?rm’s perceived
level of certainty that its partner ?rms will pursue
mutually compatible interests in the IOR. Firms
tend to have greater con?dence in the IOR when
they perceive a suitable level of control over their
partners (Sohn, 1994) and when there is trust
between ?rms (Das & Teng, 2001; Ring & Van
de Ven, 1992). Nevertheless, a review of the litera-
ture highlights the complexity of the relationship
between the two constructs (Tomkins, 2001) and
indicates a need for further research (Coletti
et al., 2005; Dekker, 2004; Van der Meer-Kooistra
& Vosselman, 2006).
As Knights et al. (2001) noted, there is a long
tradition that conceptualises MCSs and trust as
opposite alternatives. It is argued that the intro-
duction of contracts and other types of MCSs
may be perceived as lack of trust (Neu, 1991;
Ring & Van de Ven, 1994). Das and Teng
(1998) maintained that this argument assumes
that con?dence in an IOR is given and static,
and is a zero-sum calculus: any increase in either
MCSs or trust supposes a decrease in the other.
But in practice, they defend, partners will demand
the con?dence level that they feel is needed, and
this level will di?er according to a number of fac-
tors including the partner ?rm’s risk propensity,
M.L. Ve´ lez et al. / Accounting, Organizations and Society 33 (2008) 968–994 971
perceived risk, the type of knowledge involved,
and the amount of resources committed. For that
reason, they argued (p. 495) that both MCSs and
trust contribute ‘‘jointly to the total level of con?-
dence one has in partner cooperation”. Therefore
the introduction of an MCS does not necessarily
suppose a lowering of trust. Similarly, Woolthuis,
Hillebrande, and Nooteboom (2005), focusing on
contracts as a form of formal control, ?nd trust
and contracts to be both complements and alter-
natives in IORs.
Coletti et al. (2005) criticized previous studies
that considered the cooperation level in collabora-
tive settings as constant. They found that the pres-
ence of MCSs enhances the level of trust and that
observed cooperation perfectly mediates this e?ect.
Their experimental results showed that MCSs
could directly increase cooperation and, in conse-
quence, indirectly increase trust. Taken together,
these results suggest an increasing marginal e?ect
of MCSs on the evolution of cooperation. These
results concur with arguments by other authors
(Lang?eld-Smith & Smith, 2003; Poppo & Zenger,
2002) that MCSs reinforce trust because they pro-
mote expectations of cooperation and generate
commitment in the IOR.
Arguing that the information generated by
MCSs intervenes in the dynamic process of build-
ing trust, Tomkins (2001) distinguished four stages
in the IOR life cycle—pre-relationship, explor-
atory, developing, and stable—and proposed that
the negative e?ect arises only at later stages, when
trust intensity becomes established at higher levels.
In closer IORs the information needed to success-
fully carry out the activity associated with the rela-
tionship is less, and therefore the introduction of
more MCS may harm trust.
Relating Tomkins (2001) proposals with Das
and Teng’s (1998) arguments, one might think
that, when the IOR reaches a maximum point in
the inverted U-shape function, the maximum con-
?dence in partner cooperation will be reached and
remain constant, and the introduction of an MCS
will damage the pre-existing trust. This argument
is probably true when the IOR’s duration is lim-
ited to one speci?c project or venture. However,
Halinen et al. (1999) argued that in IORs with
an expectation of continuity and an open-ended
nature, the stability of mature stages could provide
a platform for continuous evolution. The IOR will
be enhanced by the trust increment, which will
encourage the partners to accept bigger risks and
to enlarge the IOR scope (Inkpen & Curral,
2004). Doz (1996), referring to joint ventures,
argued that to continue being successful, an alli-
ance will need to go through a series of spiral tran-
sitions. When the initial conditions support
continued collaboration, the movements toward
deeper cooperation involve willingness by the part-
ner ?rms. Encouraged by the favourable results,
the activities associated with mature IORs can
increase or change in character; members can
invest more resources; and their interdependence
level can be a?ected. Arin˜ o and de la Torre
(1998) concluded that positive feedback loops are
critical in the evolution of alliances, and that
MCSs are critical from the start in fostering a cli-
mate for positive reinforcement and building trust
and con?dence in the IOR.
Dekker (2004)—in speaking of the level of trust
at a certain point in time and the need for control
that the IOR itself generates—pointed out that
trust will be damaged only when formal control
exceeds the level necessary to safeguard the trans-
action. This implies that until this threshold the
use of formal controls can be complementary
and conducive to trust. Keeping in mind his argu-
ment, we suggest that if the initial conditions
evolve, a higher level of con?dence may be needed,
requiring more information for control of the IOR
and more trust, in an additive association, even in
mature IORs. This argument is summarized in our
research question:
Does MCS introduction increase already
established trust in a mature but evolving
inter-organizational relationship?
To answer this question, we considered that ite-
mising the di?erent MCS types and MCS func-
tions could advance our knowledge because they
might have di?erent e?ects on trust.
On one hand, regarding the di?erent MCS
types, Inkpen and Curral (2004) argued that
action and result controls formalize limits on the
discretion of the parties and may thus discount
trustworthy action. Similarly, Das and Teng
972 M.L. Ve´ lez et al. / Accounting, Organizations and Society 33 (2008) 968–994
(2001) pointed out that strict rules and objectives
may undermine people’s autonomy and lead to
questioning their intentions, creating an atmo-
sphere of low trust. These control types hinder
the creation of competence-based trust, as they
prevent partners from designing their own behav-
iour and demonstrating their capacity to work in
the best interest of the IOR. Conversely, shared
objectives and norms allow partners to act freely
and boost trust based on competence and good-
will.
3
These studies therefore agree that only for-
mal controls impair trust. Clearly, we need to
analyse separately the association between each
MCS type and each trust dimension. One useful,
and all-inclusive, way to classify MCSs is to focus
on the object of control (Merchant & Van der
Stede, 2003). Speci?c action controls try to ensure
that certain desirable actions are carried out and
undesirable actions are prevented, using physical
and administrative constraints, action accountabil-
ity, pre-action reviews, and redundancy. Result
controls motivate the achievement of desirable
results by de?ning, measuring, and rewarding
them. Personnel-cultural controls mitigate control
problems by improving capacities, selection and
training programmes, communication, and com-
mitment to objectives.
On the other hand, these three MCS types can
develop di?erent functions in the management of
the IOR. For example, Woolthuis et al. (2005)
found that, in a trusting atmosphere, contracts
(as a form of formal control) can have a di?erent
function and meaning, concluding that contracts
and trust can be both alternatives and comple-
ments, depending on the intentions with which
contracts are drawn up and used. Intra- and
inter-organizational studies on MCSs (e.g., Aber-
nethy & Vagnoni, 2004; Dekker, 2004) have
de?ned two main functions of MCSs:
Monitoring. MCSs are used to measure and
reward performance by supervising the
accomplishment of objectives. MCSs are
intended to reduce opportunism and diver-
gence by providing warning signals and by
aligning incentives.
Coordination. The need for MCSs arises from
the division of work-loads and specialization
and from di?ering knowledge and capacities,
which create uncertainty and interdependence.
MCSs generate, update, and share the informa-
tion needed to reduce uncertainty, to direct
attention, and to assist planning—in general,
to make decisions and solve problems. MCSs
also communicate common objectives and
assist learning.
Monitoring is often believed to be detrimental
to trust because it implicitly assumes that opportu-
nistic behaviours by partners are possible, and this
assumption can create suspicion between them
(Sitkin, 1995). Most scholars (Andaleeb, 1995;
Inkpen & Curral, 2004; Neu, 1991; Ring & Van
de Ven, 1994) who regard MCSs as an alternative
to trust focus only on their monitoring function,
confusing it with the whole.
MCSs can also be developed to coordinate
tasks (Dekker, 2004; Gulati, Lawrence, & Pura-
nam, 2005). The relationship between coordina-
tion and trust has not been analysed extensively.
However, as the need for coordination is associ-
ated with complexity, strategic importance, or
interdependence (Dekker, 2004), we may assume
that the coordination function of MCSs lacks
the negative connotations associated with the
monitoring function. MCSs can be used to share
information (Abernethy & Vagnoni, 2004), and
communication is accepted in the literature as
an important determinant of trust (Morgan &
Hunt, 1994; Sako, 1992). Seal, Cullen, Dunlop,
Berry, and Ahmed (1999) and Kaju¨ ter and Kul-
mala (2005) claim that when MCSs increase the
3
The conceptual domains of social control mechanisms,
social control, and trust have signi?cant overlap. Social control
mechanisms are structural arrangements that foster socializa-
tion and interaction between partners. The outcome of these
arrangements is social control (i.e., the ability to in?uence
behaviours) based on mutual understanding and the develop-
ment of shared values and norms. The shared values and norms
generate trust (Sako, 1992), but trust involves a positive attitude
about others; also, it is not about in?uencing others’ behaviour,
but about believing that others will perform on the relation-
ship’s behalf. Therefore, we concur with Das and Teng’s (1998)
statement that trust should not be viewed simply as a control
mechanism.
M.L. Ve´ lez et al. / Accounting, Organizations and Society 33 (2008) 968–994 973
transparency and fairness of agreements, trust is
enhanced. Although Tomkins (2001) and Dekker
(2004) have advocated distinguishing between the
MCS functions, there is still little empirical
research relating MCSs to trust that makes this
distinction.
Case study research and description
Some authors have argued that longitudinal
qualitative research is needed to clarify the under-
lying core issues of IORs (Arin˜ o & de la Torre,
1998; Dekker, 2004). Accordingly, we adopted a
longitudinal case study approach (Yin, 1984),
selecting as our research site the relationship
between a manufacturer called CMD (a pseudo-
nym) and its distribution channel, formed by a
set of 176 small agents who carry out their
assigned activities in similar ways, in given geo-
graphical territories. This may be described as a
non-equity and open-ended inter-?rm relationship
in which a set of smaller downstream agents
assumes part of the value chain functions from a
bigger upstream manufacturer. As Stern, El-
Ansary, and Coughlan (1996) note, it is possible
to suppress or to replace agents, but their func-
tions cannot be eliminated, and many of the sys-
tem’s activities cannot be explained simply by
analysing each member. In fact, CMD has viewed
the channel as one entity, developing a set of
homogeneous MCS to manage it. So, we analyse
trust as both the manufacturer’s perceptions and
expectations regarding its distribution channel, as
a whole, and the distribution channel members’
perceptions and expectations regarding the
manufacturer.
We chose this case because (1) the distribution
channel has longevity, success, and stability. The
manufacturing ?rm has used this sales channel
for more than 19 years, and as Table 1 shows,
97% of the channel members have been agents
for more than 5 years. (2) During the period
studied (1997–2004), the manufacturer gradually
introduced—uniformly to all agents—various
control tools that currently make up its MCSs.
(3) Our access to both manufacturer and distri-
bution channel members allowed us to gather
evidence from both perspectives.
Data collection
Extensive, detailed archival data allowed us to
identify the chronology of MCSs introduction, as
well as their objectives and content. The main
archival sources included the di?erent types of
contracts, the agent evaluations, minutes of all
the regular joint meetings, and projects and proce-
dures documents (see Appendix Table A.1). Inter-
views enabled us to gather the perceptions and
expectations of individual members, a key to eval-
uating trust, and served to triangulate our ?ndings.
Taking as a base the literature on MCSs and trust,
we developed guidelines for semi-structured inter-
views to orient data collection in two areas: (1)
the implementation of MCSs in the IOR and (2)
the e?ects of MCSs on trust. We held 12 interviews
with key CMD employees (see Appendix Table
A.2), interviewing boundary personnel who were
continually in contact with the channel, and other
top managers who were the designers of MCSs
and/or were involved in the channel management
at a higher level. We also interviewed eight agents
(see Appendix Table A.3), whom CMD catego-
rized into three groups: group A, with sales
exceeding 600 thousand euros per year; group B,
with sales of 360–600 thousand euros per year;
and group C, with sales of less than 360 thousand
euros per year. Our interviewees were chosen ran-
domly from these three groups (two agents from
group A, and three each from groups B and C).
Because we were focusing on a particular period
of time, we did not interview former agents. All
interviews were carried out between November
2002 and June 2003 and typically took 2–4 h.
Interviews were tape recorded and transcribed,
Table 1
Longevity of distribution channel
Longevity of agents
Less than 5 years 3.0% 3.0%
Between 6 and 10 years 12.9%
Between 11 and 15 years 27.7%
Between 16 and 20 years 21.8% 97.0%
Between 21 and 30 years 18.8%
Between 31 and 40 years 6.9%
More than 40 years 8.9%
Source: Extracted from Agents Database, 2004.
974 M.L. Ve´ lez et al. / Accounting, Organizations and Society 33 (2008) 968–994
with the complete write-up being sent to the inter-
viewees for comment. We also visited some work
centres and held informal discussions with both
IOR parties, taking advantage of the fact that we
were invited to four annual joint meetings (years
2001–2004).
Research site and antecedents
CMD was founded in the late 19th century
and is currently the leader in its branch of the
Spanish chemical industry. It is organized into
11 regional districts, and its main focus is com-
mercial, although it also produces and packages
commodities. This sector is characterised by a rel-
atively stable market, in which ?ve large compa-
nies share most of the world market and large
entry barriers discourage other potential manu-
facturers. CMD’s products are sold (a) directly
to major clients and (b) to a large number of
small and medium clients via its distribution
channel, which currently generates more than
75% of sales and caters to 90% of clients. CMD’s
ROI from its distribution channel is 3.4 times
higher than from direct sales.
The distribution channel is formed by micro-
?rms (physical people or ?rms with 1–4 workers).
The channel members receive the products on
consignment for distribution and sale. There is
no internal competition; each agent acts exclu-
sively in a given geographical area and receives
commissions on CMD’s invoicing, regardless of
who actually delivers the products to the client.
The commission system and the market stability
enable agents to obtain high and stable pro?ts.
In addition, serving as a CMD agent is compati-
ble with selling other industrial supplies, and
because of the synergy between businesses,
large-scale investments are not necessary to join
the CMD channel.
Although CMD had a distribution channel
practically from its foundation, this channel was
consolidated in the mid-80s. In 1985, CMD for-
malized a single contract type for the entire distri-
bution channel. This contract de?ned the
functions and responsibilities of distribution chan-
nel members, the exclusive geographical areas
assigned to them, and the commissions, which
functioned as both remuneration and a control
system.
At the same time, CMD established a commer-
cial department. The role of commercial sta? was
essential to the proper working of the distribution
channel. Besides conducting typical sales activities
with large direct clients, the commercial sta? also
supplied assistance in all the channel functions
and became the main liaisons in the IOR. They
also carried out (and still do) direct supervision
of channel members’ actions, supporting the
greater part of commercial activity: indicating
channel needs, supplying technical and commer-
cial support, and even selling on behalf of the
channel members. From this moment onwards
the IOR was structured as shown in Fig. 1.
From 1985 through 1996, CMD’s trust in its
distribution channel had been built up from expe-
rience and good results. ‘‘The fact that CMD has
had a distribution channel during so many years is
based on trust” (former CMD CEO). ‘‘CMD trusts
the channel from the beginning, because we gave
them our client portfolio, a ?xed revenue, before
they do anything” (Sales Manager). Before the per-
iod of study, the activities delegated to the distri-
bution channel were simple: warehousing and
delivery of products. The process of selecting
Agent 1.1-1
Agent 1.1-2
Agent 1.1-j
Commercial 1-1
Agent 1.2-1
Agent 1.2-i
Commercial 1-2
..........
Regional District 1
Agent n.1-1
Agent n.1-2
Agent n.1-3
Agent n.1-x
Commercial n-1
Agent n.2-z
Commercial n-2
.........
Regional District n
Agent 11.1-1
Agent 11.1-2
Agent 11.1-3
Agent 11.1-p
Commercial 11-1
Agent 11-w
Commercial 11-2
.......
Regional District 11
CMD
TMT and corporate departments
Fig. 1. Organization chart of CMD and its distribution channel
relationship.
M.L. Ve´ lez et al. / Accounting, Organizations and Society 33 (2008) 968–994 975
agents was based mainly on their reputation and
on previous contacts. The commercial sta?’s visits
favoured social relationships, which encouraged
the creation of trust. However, CMD related
separately with distribution channel members,
and usually ‘‘we tried to avoid agents’ meetings”
(CAC – Client Attention Centre-Manager).
Since the 1985 formalization, CMD has looked
to develop strategies and behaviours that favour a
good climate and trust. The former CMD CEO
referred to it in this way: ‘‘Agents know that we
do not look to abuse them, we look for the long-term
relationship. Although they always complain about
the commissions, as employees complain about sala-
ries, the relationship tone indicates that the channel
trusts us reasonably, and they know that in con?ict
situations that are not covered in the contracts, the
treatment that they will be given will be fair”.
For example, in 1992, CMD’s ?nancial state-
ments reported losses. The reasons were not clear,
but the losses were put down to macroeconomic
factors, excessive costs, and duplication of centres
and activities, among other causes. CMD hired a
consultant ?rm, which recommended that it aban-
don the channel model and use resellers as its com-
petitors did. However, on the basis of trust
developed over the years, CMD rejected the con-
sultant’s proposal. CMD considered that through
its channel ‘‘in spite of the distance, we had identi-
?ed and assisted small clients” (Sales Manager).
‘‘Our channel is better than those of the competitors.
CMD bet on it, even against some people who indi-
cated just the opposite” (District Director).
On their part, the agents trusted the reputation
of a bigger ?rm ‘‘that provides us an image, and rea-
sonable bene?ts” (Agent B-2). Their trust in CMD
is manifested in longstanding relationships and
low desertion. The widespread feeling of trust is
shown in comments such as ‘‘I have always felt a
friend of CMD, more than in a buyer–seller rela-
tionship” (Agent B-1). The channel members per-
ceive CMD’s commitment toward them, re?ected
in small sacri?ces that are very important for
them. ‘‘They have always helped us. I remember
that we had ?nancial problems and they helped us”
(Agent A-2). ‘‘They have always trusted us. For
example, they never disapproved me in a client
action” (Agent C-3).
Events in the evolution of the CMD – distribution
channel relationship
We have tracked seven events that show the
chronological evolution of MCSs. Table 2 summa-
rizes the time line and shows, for each event, the
MCS tools’ use over time and the IOR evolution.
– Event 1: 1997 Agents’ ?rst meeting. CMD called
a videoconference for all of its employees with the
aim of communicating the ?rm’s new strategy,
which changed focus from production to market-
ing, highlighting the need to ‘‘deliver a di?erence”
to end clients. ‘‘If we are able to sell better, at a bet-
ter price, it is because we are di?erent and we should
be able to give that di?erence to clients. . . Since
1997, CMD has realized that it cannot remain sta-
tic, that it owes users the added values of the channel
and needs to give them more” (District Director).
All channel members were also invited to attend
this conference, which was regarded by both par-
ties as a turning point in the IOR. Interviewees
saw this meeting as foreshadowing what the rela-
tionship would be in the future, in terms of moti-
vation and greater distribution channel alignment
with CMD. ‘‘We saw that the channel belonged to
the CMD value chain, and that it was necessary to
?nd communication, and to work as a team” (Com-
mercial Director). ‘‘CMD began to invest its e?ort
in creating a commercial network based on the dis-
tribution channel, and that is our advantage” (Sales
Manager). ‘‘From that moment onwards they valued
us and were closer to us” (Agent C-1).
– Event 2: 1998 Electronic integration. ‘‘We
needed systems that, by qualifying [Distribution
channel members] to execute the assigned func-
tions, would allow us to keep control over the mar-
ket” (District Director). In 1998, CMD supplied
its entire channel with computers and shared its
own software. Its goal was to homogenize tasks,
improve activities control, eliminate duplicated
functions (mainly administrative), and reduce its
inventory level. ‘‘It liberated us from an important
administrative work-load; we won in speed and reli-
ability of data, and decreased client complaints”
(CAC Manager). ‘‘We invest time, and CMD
saves it” (Agent C-1). These systems allowed
976 M.L. Ve´ lez et al. / Accounting, Organizations and Society 33 (2008) 968–994
CMD to externalise new functions (i.e., mechanis-
ing deliveries and invoices, creating and maintain-
ing client ?les, setting price discounts, managing
its own and clients’ warehouses and logistics).
‘‘It allows us to know the identity of clients assisted
by the channel. . . with this information, we feel
that this makes it harder for agents to abandon
us and contract with our competitors” (former
Agents Manager). ‘‘It allows us greater e?ciency,
even though the channel model is more costly” (for-
mer CMD CEO). The agents willingly accepted
CMD software as a means to automate their
tasks. ‘‘We have advanced a lot. They gave us a
very good tool, to see consumptions or stock levels.
Now we control the market” (Agent B-2).
– Event 3: 1999 Agents’ marketing plan. This plan
consisted of ?ve phases: (a) a new agent pro?le
de?nition; (b) an agents control scorecard; (c) the
creation of an agents database; (d) the beginning
of regular meetings; and (e) channel satisfaction
surveys.
(a) The de?nition of the agent pro?le and the
establishment of norms for agent selection
allowed CMD to formalize these processes,
and ‘‘adapted the existent channel” (Agents
Manager). The agent pro?le evolved to com-
Table 2
Events time line
1
Relationship formalization
Contract, commercial department,
and commission system establishment.
Agent numbers increase.
Warehouse and Delivery tasks Warehouse, Delivery, Administrative and Commercial tasks
Warehouse, Delivery and Administrative tasks
2
3
4
5
6
7
Legend
1. Change in the distribution channel policy. Turning point in the IOR. CMD reorganizes its activities from production to market focus.
2. Outsourcing of administrative tasks. Mechanization of activities, qualifying the distribution channel to control the end market.
3. Outsourcing of new commercial activities. More links, shared information, benchmarking and interface channel structures,
developing an agent profile, an agents control scorecard, an agents database, periodical agents’ meetings and channel satisfaction
surveys.
4. Development of new procedures in order to homogenize and constrain the activities. Channel training plan. Gradual reduction in
agent numbers.
5. Performance measurement system to appraise the channel activities, including 37 items. Ranking of agents to reward the best, and to
identify those who required more assistance. Gradual reduction in agent numbers.
6. New platform to include all of the former tools, as well as some new services. Gradual reduction in agent numbers.
7. Commitment and Excellence awards. Objectives and improvement projects to reach target congruence. Gradual reduction in agent
numbers.
1985 to 1996 1997 1998 1999 2000 2001 2002 2003 2004
Antecedents
First agents’ meeting
Electronic integration
Marketing plan
Management manual
Agents’ evaluation system
Web page integration
Joint targets development
1
Relationship formalization
Contract, commercial department,
and commission system establishment.
Agent numbers increase.
Warehouse and Delivery tasks Warehouse, Delivery, Administrative and Commercial tasks
Warehouse, Delivery and Administrative tasks
2
3
4
5
6
7
Legend
1. Change in the distribution channel policy. Turning point in the IOR. CMD reorganizes its activities from production to market focus.
2. Outsourcing of administrative tasks. Mechanization of activities, qualifying the distribution channel to control the end market.
3. Outsourcing of new commercial activities. More links, shared information, benchmarking and interface channel structures,
developing an agent profile, an agents control scorecard, an agents database, periodical agents’ meetings and channel satisfaction
surveys.
4. Development of new procedures in order to homogenize and constrain the activities. Channel training plan. Gradual reduction in
agent numbers.
5. Performance measurement system to appraise the channel activities, including 37 items. Ranking of agents to reward the best, and to
identify those who required more assistance. Gradual reduction in agent numbers.
6. New platform to include all of the former tools, as well as some new services. Gradual reduction in agent numbers.
7. Commitment and Excellence awards. Objectives and improvement projects to reach target congruence. Gradual reduction in agent
numbers.
1985 to 1996 1997 1998 1999 2000 2001 2002 2003 2004
Antecedents
First agents’ meeting
Electronic integration
Marketing plan
Management manual
Agents’ evaluation system
Web page integration
Joint targets development
1985 to 1996 1997 1998 1999 2000 2001 2002 2003 2004
Antecedents
First agents’ meeting
1985 to 1996 1997 1998 1999 2000 2001 2002 2003 2004
Antecedents
First agents’ meeting
Electronic integration
Marketing plan
Electronic integration
Marketing plan
Management manual
Agents’ evaluation system
Management manual
Agents’ evaluation system
Web page integration
Joint targets development
Web page integration
Joint targets development
Chronology of events (start at top).
M.L. Ve´ lez et al. / Accounting, Organizations and Society 33 (2008) 968–994 977
prise the following basic characteristics: A
CMD agent must (1) be an industrial supply
?rm; (2) know its territory; (3) have a good
reputation; and (4) conduct other related
business, which should compose approxi-
mately 75% of its total (Agent selection
norm). ‘‘We have consolidated our channel.
We de?ned what the channel type must be,
describing how [agents] should work. This
takes us to the following step, to evaluate the
channel against this de?nition” (former
Agents Manager). This new pro?le refocused
the channel towards commercial activities,
which highlighted that CMD needed to begin
to supply commercial information to the
channel. ‘‘The new pro?le de?ned a profes-
sional looking for new opportunities and not
just maintaining our portfolio” (District
Director). The channel members saw this
process as normal: ‘‘CMD was looking for
our synergies” (Agent B-3).
(b) With the introduction of the agents control
scorecard, CMD began to share information
with the channel (information that some
agents were already demanding). ‘‘It gives
them information to improve their manage-
ment” (District Director). ‘‘It is a support tool
where each agent’s situation is analysed, and it
provides a great quantity of information about
the agent’s sales, decreases in individual cli-
ents’ orders, stocks, sales analysis regarding
the previous year, itemized sales analysis. . .”
(Agents control scorecard report). ‘‘Elabo-
rating tools to help them, we told them that
the clients should be managed, requesting that
they take responsibility for doing so” (Agents
Manager). The feedback received from the
channel is ‘‘completely satisfactory. Our
channel members have agile information that
allows them to develop their work in a more
coordinated, e?ective, quick and satisfactory
way, using a common language” (Agents con-
trol scorecard report). As expressed by the
designers, the philosophy was ‘‘give in order
to receive”. ‘‘If they were partners, they
needed to receive professional-to-professional
treatment” (District Director). ‘‘We began
to transform the channel into a partnership,
not only into operational executors” (former
CMD CEO). ‘‘There were people who feared
the sharing of information. However, it is very
positive that the channel members see that we
have not hidden anything, because we believe
that they are trustworthy and we should not
fear that they will leave for competitor ?rms”
(former Agents Manager). ‘‘When we give
them information, we give them the decision
power, to de?ne how their future will be and
how they will negotiate the client portfolio that
CMD has given them. Now they depend less
on us to work” (District Director).
Channel members discovered that this tool
helped them to acquire new commercial abil-
ities, ‘‘to detect problems; it helps us in our
management, allowing us to take quicker deci-
sions and to establish plans. It is information
that allows us to have our independence and
not to depend on CMD” (Agent B-3). ‘‘It is
very useful, to analyse, to compare, to negoti-
ate” (Agent A-1). On the other hand, the
commercial department used the information
to carry out its functions of monitoring and
assisting the agents. ‘‘Again, the commercial
member helps us to understand this informa-
tion. He is the teacher, motivating us to use
this new information” (Agent C-2).
(c) CMD developed a database containing the
available information about the channel.
This database, with access restricted to
CMD, was used to compare agents and to
analyse the evolution of channel results. ‘‘In
this database we register each agent’s charac-
teristics, with quantitative and qualitative
data. It is a benchmarking database” (Agents
Manager).
(d) As a continuation of the ?rst meeting, CMD
began holding regular joint meetings
between channel members and its own sta?.
‘‘Before, it was not considered proper. That
changed; it is necessary to communicate so
that everyone receives the same message
because not all the commercial sta? can or
want to transmit things in the same way”
(Commercial Director). ‘‘It created uncer-
tainty, about whether the agents might use
the meetings to mount group protests. In fact,
978 M.L. Ve´ lez et al. / Accounting, Organizations and Society 33 (2008) 968–994
there were some managers who thought this
way, but now they do not admit it” (CAC
Manager). These meetings were used to com-
municate results and budgets, commercial
and price policies, marketing campaigns,
and technical, legal, and logistic issues, as
well as administrative matters. Furthermore,
the meetings were used to train agents in new
activities, and gradually incorporated per-
sonnel from other CMD departments (logis-
tics, CAC, and commercial). ‘‘We needed to
create a team feeling. There should be total
transparency between them and us” (Agents
Manager).
Both parties highlighted the meetings’ bene-
?ts, as shown by the following comments:
‘‘The channel is part of the family and loyal;
the communication should be open” (former
Agents Manager). ‘‘The culture of the ?rms
is changing. We transmit feedback to them
and they contribute with their suggestions
and complaints, and they know each other”
(former CMD CEO). ‘‘The objective is to
make the channel identify with CMD, to unify
concepts so that we all give the same image”
(Agent C-3). ‘‘CMD looks for contact,
because we conduct its activities. There are
always topics to speak about, problems to
solve” (Agent A-2). ‘‘It allows us to be
informed of the results, of the future. And it
is good to talk about experiences with other
agents” (Agent B-3). ‘‘In these meetings we
approach the thoughts of each party, and more
unity is achieved with the CMD sta? and
among the agents” (Agent B-2). Through a
questionnaire, CMD evaluates satisfaction
with the meetings. Both parties consider that
‘‘the global result is very satisfactory, demon-
strated by the channel attendance (more than
90%), participation, and valuation” (E-mail
from the Agents Manager).
(e) In that same year, CMD began to carry out
channel satisfaction surveys with two objec-
tives: (1) to evaluate the global satisfaction
level and (2) to identify those areas in the
IOR that caused agent concern.
For the 139 questionnaires returned
(response rate of 65%) (see Table 3), CMD
highlights that most of the channel members
(69%) were satis?ed or very satis?ed, and
only 5% were not satis?ed. In the free-com-
ment section of the questionnaire, agents
noted that their main concerns were training,
stock levels and supplies, and more aggres-
sive marketing. Additionally, CMD began
to publish periodicals, such as ‘‘Product
News” for the channel and ‘‘CMD Maga-
zine” for sta?, channel, and clients.
In a marketing plan summary, a former
Agents Manager outlined the reasons for this
initiative: ‘‘It was a very important channel,
and it was not ?tting to have so little control”.
Table 3
Channel satisfaction survey results
Items Average
a
St.d. % 1–2
b
No. of agents
Commercial department support 3.99 0.95 5%
Clarity of the commercial strategy On new clients 3.26 1.10 22%
On prices 3.10 1.07 26%
On commissions 2.86 1.18 38%
CMD personal support 3.77 1.05 10%
Warehouse information satisfaction 4.01 1.06 9%
Computer systems satisfaction 3.65 1.09 14%
. . ./. . .
General degree of satisfaction with CMD 3.92 0.86 5%
Would you want to receive information on. . .? No. of Yes Sales methods 101
Service quality 94
Client attention 98
Source: Channel satisfaction survey, 1999, rate of response 65%, 139 questionaries returned.
a
Likert scale from 1 (not satis?ed) to 5 (very satis?ed).
b
Percentage of agents that answer 1 or 2, showing low satisfaction.
M.L. Ve´ lez et al. / Accounting, Organizations and Society 33 (2008) 968–994 979
– Event 4: 2000 Agents’ management manual. To
complete the electronic integration, CMD sent an
operations manual to all agents, outlining the steps
to follow in storage, distribution, administration,
and commercialisation. Additionally, CMD estab-
lished ‘‘a training plan in order to reach the high lev-
els of knowledge that will enable them to provide the
level of service that we want to o?er to our clients”
(Channel Training Plan report).
– Event 5: 2001 Agents evaluation system. CMD
set up an agents evaluation system (AES). ‘‘Before
we had sales data, but we could not foresee the
appropriate management, the area potentiality. We
had data but dispersed. Now we have them together
and known by both parties” (CAC Manager). ‘‘It
has provided us information to know objectively
how our channel is doing” (Sales Manager). Com-
pleting the former commission systems, the AES
achieved two objectives: to reward those agents
who obtained the best results, and to identify those
who required more assistance. ‘‘The AES allows
foreseeing possible con?icts that sooner or later we
will have to confront” (Commercial Director).
The AES included 37 performance measures: inter-
nal and external, and ?nancial and non-?nancial
items that evaluated all the channel activities
(warehousing, delivery, administration, and com-
mercialisation). Eleven items are obtained through
surveys of client satisfaction with agent services.
The results obtained from 2002 to 2004 show a
notable improvement in clients’ satisfaction. When
the AES was ?rst introduced, more than half its
indicators referred to commercial and sales activi-
ties and the ?nal market service appraisal; but
CMD explained that these objectives and weights
would change over time to guide performance pri-
orities. ‘‘It is a system to enhance the strong points
and to solve the weak ones. We want to boost those
improvements” (Agents Manager). ‘‘It is a profes-
sional view, easier to evaluate, where both parties
know what items will be valued, and agree yearly
on where to arrive on every item” (District
Director).
CMD perceives a high level of acceptance by
the channel members. ‘‘After the ?rst evaluations
all their comments have been very satisfactory”
(CMD CEO). ‘‘The information is used for control
by CMD and it also helps us to improve our work. It
is normal for the ?rm that pays us every month to
make certain demands” (Agent B-3). ‘‘CMD is
appreciating the channel. Choosing this system,
CMD has increased the trust placed in us” (Agent
B-1). ‘‘CMD now demands more, they want us to
visit and assist clients” (Agent B-2). ‘‘We can see
our defects and try to correct them” (Agent C-3).
‘‘Now we know in what position we are inside
CMD. We can compare ourself with other agents.
It is a motivation tool” (Agent A-2).
– Event 6: 2002 Agents’ web page. Continuing with
the electronic integration, CMD designed a web
page that included all of the former tools, as well
as some new services (optional low-cost responsi-
bility insurance, a safety advisor, ISO-9000 con-
sulting, and email) and the foundations of future
services (a complaint resolution programme, pro-
cessing of client debts, and ISO-14000 consulting).
‘‘The change is: take information, and tell me what
you ?nd. CMD is not afraid to give open informa-
tion; the channel has to act as CMD, as an autono-
mous centre. We have to create a community where
all feel comfortable, proud of belonging to the chan-
nel. To be able to demand more from them you have
to give them more” (District Manager). ‘‘To be able
to use the web page and the new applications more
e?ciently, a speci?c training program has been elab-
orated . . . it is an open environment that will incor-
porate all those contents that are considered of
interest to the channel. Please, send me any sugges-
tions” (Agents Manager letter to the channel).
Channel members regarded this as another step
forward in the IOR, appreciated the improved
communications and assistance supplied by
CMD, and considered that the IOR had become
more ?uent. ‘‘I believe that CMD is valuing us prop-
erly” (Agent B-2). ‘‘Issues are now better managed,
there is a better atmosphere, and everything is per-
fectly under control thanks to the new tools. It is
better to solve a ‘little problem’ now than a ‘major
problem’ later. I think that trust has grown” (Agent
A-2).
– Event 7: 2003 Joint targets and improvement
projects. CMD changed its incentives system to
an award system based on the AES. It established
980 M.L. Ve´ lez et al. / Accounting, Organizations and Society 33 (2008) 968–994
the ‘‘commitment award, for those agents who
exceeded the average evaluation” and the ‘‘excel-
lence award, for those who surpassed the level of
commitment”. In 2003, 9% of agents obtained the
excellence award and 12% the commitment award.
The channel members appreciated this award sys-
tem, albeit of small monetary value, as recognition
of their work. The introduction of the AES and
control scorecard also led to agreement on annual
improvement plans. These projects focused mainly
on operations, not on sales targets, allowing agents
to eliminate weaknesses that the AES highlighted.
Although it is made clear in the annual meetings
that CMD seeks only two projects per year from
each agent, as an incremental improvement, in
2003 the agents voluntarily established on average
5.3 new projects (see Table 4). The total of 934
improvement projects established by the channel
was received by CMD as a signal of the agents’
interest and alacrity. As Table 4 summarises, only
311 of these projects were achieved; however,
CMD accepted that at the beginning of this pro-
cess, agents were likely to bite o? more than they
could chew. The completed projects were distrib-
uted evenly across agents, for an average of 1.8
projects per agent, very nearly reaching in the ?rst
year the target of two projects per year from each
agent; 57% of agents reached or exceeded the stan-
dard. These ?rst results were valued positively by
both sides. ‘‘In the channel it is important to see
the tendency; these tools allow us to set down objec-
tives” (former CMD CEO). ‘‘Now, the channel
gives you solutions, and if it is good for them, it is
good for CMD. This is true for most of the agents;
there are still some with the older pro?le” (Agents
Manager). ‘‘A professionalization is expected, to
introduce us into the CMD world” (Agent B-1).
‘‘Each agent knows what it has to do and where it
has to act” (Agent A-2).
As a direct consequence of MCS development,
after 1999 there was a gradual fall in the number
of agents (see Table 5), mainly in types A and C
(see Table 6). MCS development highlighted new
challenges and commitments for agents. Although
circumstances varied across ?rms, interviewees
argued that most of the discontinued type A
agents had shown little interest in the new activi-
ties, or had not been pro-active about their
CMD business. Regarding type C agents, intervie-
wees argued that those who were discontinued
lacked the sales capacities or resources to succeed
in the new activities. Through MCS, e.g., with
the AES measures or the agents database, CMD
and the agents themselves acquired the informa-
tion they needed to evaluate the situation of each
agent ?rm, analyzing its performance over time
and its strong and weak points. As a result, some
agents voluntarily left the channel because they
lacked interest or resources to engage in the new
activities; in other cases CMD initiated agreements
(normally friendly) to induce agents to leave the
relationship.
Table 4
2003 annual improvement projects of agents
Agents’ categorization by sales volume A B C Total
1. Number of agents 46 54 76 176
2. Annual projects jointly established 253 286 395 934
3. Projects established per agent (2/1) 5.5 5.3 5.2 5.3
4. Annual projects achieved 92 97 122 311
5. Projects achieved per agent (4/1) 2 1.8 1.6 1.8
Number of projects achieved per agent
2 or more 1 0
Annual improvement projects achieved
Number of agents 100 51 25
Percentage of agents 57% 29% 14%
Source: Results extracted from AES, 2003.
M.L. Ve´ lez et al. / Accounting, Organizations and Society 33 (2008) 968–994 981
According to the new agent pro?le, CMD was
seeking agent ?rms with proper capabilities and
interest, for which CMD activities would consti-
tute 25% of their business, and which would main-
tain their proximity to end clients. These qualities
suggested a better ?t with B type agents. ‘‘If we
want a certain quality of performance, agents must
have a minimum level of size and market” (CMD
CEO). ‘‘Large agents are usually less committed
and integrated” (Logistic director). Surviving
agents regarded this concentration positively, as
they had increased in size, territory, number of cli-
ents, and pro?ts.
In summary, the distribution channel evolved to
a more complex IOR, with new administrative and
commercial tasks accompanied by sales increases
(see Table 5). The new commercial tasks, as well
as improving client service, allowed the introduc-
tion of new items of billing and the maintenance
of premium prices, which provided bigger bene?ts
for both parties. With the introduction of these
MCSs, CMD personnel felt that they had obtained
su?cient control over the distribution channel.
Likewise, both parties highlight that the devel-
opment of MCSs increased trust and the sense of
partnership. Table 7 presents selected quotations
that show various aspects of how the MCSs built
trust. CMD managers perceive that MCSs were
needed to open up the relationship, let them better
understand agents’ needs, and clarify their own
expectations about agents’ abilities. CMD manag-
ers recognize that MCS development allows them
to see the distribution channel as their best collab-
orator, enclosing the channel in CMD’s chain of
value and improving performance, which in turn
creates more trust. Equally, the interviews show
how the channel members freely accept the CMD
philosophy and perceive the MCS development
as a CMD bet on them, feeling that CMD increas-
ingly appreciates them; they feel that CMD makes
sacri?ces for them beyond what is called for in the
contract. Agents highlight that MCSs improved
relationship ?uency and atmosphere, helping to
solve problems and demonstrating CMD’s capac-
ity to establish guidelines. Furthermore, from our
visits to agents’ o?ces and annual joint meetings,
we noted the high degree of integration with
CMD in terms of notices, work clothes, and tele-
phone conversations. Additionally, both parties
valued highly the role of the commercial depart-
ment in the IOR, and its importance in creating
and maintaining trust.
Analysis and discussion
Before our study period the distribution channel
tasks were simple, so the MCSs used by the manu-
facturer were basic and commonplace in distribu-
tion channel management (Stern et al., 1996): the
Table 5
Number of agents per year
Year No. of agents Reduction % Total sales €MM Increase €MM %
1999 214 57
2001 192 À22 À10 75 18 32
2002 185 À7 À4 78 3 4
2003 176 À9 À5 84 6 8
2004 173 À3 À2 90 6 7
Total 5 years À41 À20 33 58
Source: Extracted from agents database.
Table 6
Reduction in number of agents by type, 2001–2003
Year Type A % Var % Sales Type B % Var % Sales Type C % Var % Sales No. of agents
2001 48 59% 32 20% 112 21% 192
2003 44 À8% 56% 56 75% 32% 76 À32% 12% 176
Source: Extracted from AES, 2001 and 2003.
982 M.L. Ve´ lez et al. / Accounting, Organizations and Society 33 (2008) 968–994
Table 7
MCSs e?ects on trust: empirical evidence
MCSs e?ects on trust: illustrative quotations
CMD ‘‘We have our control systems, the same that we have for our personnel. The work is based on trust, and
also on those MCSs necessary to conduct the activity. . . We want to include them in our chain of value,
[for them] to participate in other business of CMD. Our challenge is to see how to involve them in this
project. We have communication and information tools; we have tools to reach objectives. . . The channel
is the best collaborator, and it is necessary to treat [members] as if they were employees. . . Trust has been
increased and improved totally. The relationship is open, except for some scattered cases. The future strat-
egy is very clear, to maintain this communication level” (CMD CEO)
‘‘It is one of our features, to create a network to give service to the clients, as an integrated community.
Trust depends on better information for both parties. It is unavoidable that we have to evaluate somehow
the performance and results that they obtain. There are some basic norms for all. But obviously some
agents are putting out a lot of e?ort, are getting better results, and others need more help” (former
CMD CEO)
‘‘Through the channel satisfaction surveys, we engaged in self-criticism to establish a trusting relationship.
We actively listen to them, and keep their opinion in mind. . . The system is good, and we are moving in
the direction of integrating the channel into CMD, and CMD into the channel. . . Trust plays an impor-
tant role. But controls are established to ensure that trust exists” (District Director)
‘‘Before, information was power. Now, you must give them the information that they need to be able to
manage their business. . . The agents trust CMD. All these changes have varied the trust toward more
trust. Now the channel has the same programs, the same databases, and the same access to clients’ data,
as we do. . . The channel is our face and eyes; thus we want a win/win relationship. With a total trust in the
channel, we should think about what we have to do to make things good for both parties” (Commercial
Director)
‘‘In 1997, we realized that we needed collaborators, and by improving the systems, we began the philos-
ophy of giving to receive, to modernize the relationship. The revolution of the channel began with the
cultural change from agents to partners. . . The channel members have seen our intention to help them.
And it has improved the trust in the IOR” (CAC Manager)
‘‘We see that they are loyal and that they help us in our business. . . Many contact tools have been devel-
oped. The channel is more open and we have better ?ows of information that allow us to make better
decisions. . . Our channel is di?erent in treatment and identity” (Sales Manager)
‘‘We have been working together for many years and want to remain loyal to them” (Agents Manager)
‘‘The great success of CMD was to continue with the channel. And the success is that this channel is so
much closer and that it carries out all the tasks, thanks to MCSs” (former Agents Manager)
‘‘We protect the channel members to avoid losing markets. We have formed, advised and informed them.
And tools allow us to speak in the same language” (Logistics Manager)
Distribution channel ‘‘I have other business with other multinationals with di?erent philosophies. . . and I keep with the CMD
philosophy. . . Now the mechanisms are better, but the control is the same. I think that they help me more
and they control me. Control is not a problem for us. It is logical because CMD trusts us. CMD is our
adviser” (Agent A-1)
‘‘More friendship ties. Although CMD has to impose things, we accept them, because we are its collab-
orators. CMD has given us its client portfolio and a not excessively oppressive norm” (Agent A-2).
‘‘CMD values us better. There is a better atmosphere. . . CMD has increased its trust in the channel in that
it has chosen this network, this system. In my relationship with CMD I value trust, because a relationship
without trust does not mean anything” (Agent B-1)
‘‘CMD trusts the channel. . . and we trust CMD. Now the tasks are made better, not because before they
were bad, but because it was more complicated to communicate and to solve problems. . . The new systems
allow us to have a better relationship with CMD. These tools allow us to be more integrated” (Agent B-2)
‘‘Trust has grown. . . Trust in the system and in CMD people. . . I believe that CMD has to mark the
guidelines clearly. I feel more a friend of CMD. With CMD people, my trust is clear. With the organiza-
tion, the problem is that CMD is very big, and it is very di?cult to say. I have always been supported, and
the daily contact makes you to be integrated” (Agent B-3)
(continued on next page)
M.L. Ve´ lez et al. / Accounting, Organizations and Society 33 (2008) 968–994 983
contract, the commission system, and direct super-
vision. At the beginning of our period of study,
with di?erent degrees for each agent, CMD had
developed a high trust in the channel as a whole,
based on its good work in assigned tasks and on
CMD’s expectations of channel members’ contin-
ued commitment (with some reservations about
whether agents should have meetings). On their
part, although with ?ne distinctions, the channel
members shared positive expectations regarding
CMD’s market and business knowledge and its
capacity to make small but signi?cant sacri?ces
for them. It was in this setting that the manufac-
turer undertook to introduce MCSs in a homoge-
neous way to the entire channel. Analysing the
common aspects in the relationships between
CMD and its agents, in this section, we ?rst discuss
how each MCS type in?uenced the twofold dimen-
sions of trust in this mature stage. Second, we ana-
lyse how each MCS function a?ected trust. Finally,
we discuss the general association between MCSs
development and already established trust in evolv-
ing inter-organizational relationships.
Analysis of the association between each MCS type
and each trust dimension
Table 8 classi?es MCS types (action, results,
and personnel-cultural controls) and analyses,
from the perspectives of both manufacturer and
channel, how each MCS type enhanced the tru-
stor’s expectation of the trustee’s competence and
goodwill.
MCS types and competence-based trust
Personnel-cultural controls (e.g., channel satis-
faction surveys, meetings, and training plans)
transferred capabilities and enhanced each party’s
knowledge about the other. The new action con-
trols (e.g., electronic integration and norms) also
improved the amount and quality of information
available. The channel members received knowl-
edge in order to carry out their tasks e?ectively
and e?ciently, and, in turn, CMD increased its
expectations that the channel would perform its
assigned tasks properly. Additionally, the avail-
able information gave the channel members a bet-
ter understanding of CMD, its market knowledge,
its manner of doing business, and its management
prowess, which added to their perception of ‘‘pro-
fessionalism” (Doney & Cannon, 1997; Kumar
et al., 1995). Action and result controls formalized
the IOR by narrowing the domain and severity of
risk (Poppo & Zenger, 2002), increasing the pre-
dictability of results and behaviours. Results con-
trols also enhanced members’ mutual knowledge
through the participation they generated, as Lang-
?eld-Smith and Smith (2003) have argued. Addi-
tionally, the satisfactory results obtained—
evaluated by the control scorecard and agents
evaluation system—led both parties to perceive
that they had chosen a reliable partner who was
capable of delivering the new goals established
for the IOR (Kumar et al., 1995; Selnes, 1998).
As Sako (1992) claims, MCSs create a frame
where information transfers knowledge, and in
turn builds competence-based trust. All MCS
types—not only informal controls—contribute to
building this more objective dimension of trust.
In this case formal MCSs provided a common lan-
guage and evidence with a greater appearance of
objectivity, evidence that was needed to reinforce
expectations about trustee abilities and compe-
tences. The channel members, through MCSs,
Table 7 (continued)
MCSs e?ects on trust: illustrative quotations
‘‘Trust has grown. I know agents who had been very critical of CMD. And I see that the level of criticism
has decreased, because the relationship has improved” (Agent C-1)
‘‘With MCSs, CMD bets on us, protects us, and sacri?ces for us; they outline things with transparency
and human value. . . The systems and people have made all concepts very clear, and that gives us con?-
dence” (Agent C-2)
‘‘CMD has marketing, R + D, administration departments, and a department for its relationship with the
channel. CMD establishes plans on its own and the channel’s behalf. . . CMD helps us and sacri?ces for
us. Unquestionably, CMD’s employees belong to an organization. However, I think that they have
trusted us exceedingly” (Agent C-3)
984 M.L. Ve´ lez et al. / Accounting, Organizations and Society 33 (2008) 968–994
Table 8
How all MCS types act as trust builder mechanisms
Type of
control
Events Control mechanisms Competence-based trust Goodwill-based trust
Channel towards
manufacturer
? Manufacturer
towards channel
Channel towards
manufacturer
? Manufacturer
towards channel
Actions
control
1998: Electronic
integration
Mechanisation and
homogenization of
processes. Norms and
procedures to limit
actions. Action
accountability. Pre-
action review
It increases knowledge
about the
manufacturer’s
competences (Sako,
1992). Through it, the
channel improves
market control. The
IOR is formalized
(Dwyer & Oh, 1987;
Poppo & Zenger, 2002)
It quali?es the
channel, transferring
knowledge and
capacities (Sako,
1992). It improves
channel and market
control, gaining in
speed and reliability
of data. CMD sees
the market through
the channel’ eyes.
The IOR is
formalized (Dwyer
& Oh, 1987; Poppo
& Zenger, 2002)
It builds the
perception of
integration in the
IOR (Dwyer & Oh,
1987). Unilateral
investment creates
routines towards
long-term
integration.
Assistance and
technical support
improve
communication
?uency with more
departments. It acts
as signal of
commitment (Sako
& Helper, 1998)
It increases
communication
between di?erent
departments and the
channel, allowing
better prediction of
behaviours. Giving
to receive allows
reducing
uncertainty.
Qualifying the
channel, CMD feels
that is more di?cult
for agents to leave
for competitor ?rms
1999: Agents’
marketing plan
Agents’ control
scorecard. Feedback
control
It increases knowledge
about the manufacturer.
It improves their
management. It gives
feedback information.
Communication takes
place in a common
language (Das & Teng,
2001; Morgan & Hunt,
1994)
It quali?es the
channel, transferring
knowledge and
capacities (Sako,
1992). It improves
coordination and
results
Shared information
formalizes the IOR
and enhances
participation in
target-setting.
Assistance and
technical support
allow a professional
treatment
It allows establishing
and aligning
objectives with each
agent (Anderson &
Weitz, 1989). Giving
to receive makes
behaviours more
predictable
2000: Agents’
management
manual
Homogenisation of
tasks. Norms and
procedures to limit
actions
It increases knowledge
about the manufacturer.
It improves their
management.
Communication takes
place in a common
language (Das & Teng,
2001; Morgan & Hunt,
1994)
It quali?es the
channel, transferring
knowledge and
capacities (Sako,
1992). It improves
coordination and
results
Assistance and
technical support act
as signals of
commitment (Sako
& Helper, 1998)
Giving to receive
makes behaviours
more predictable
(continued on next page)
M
.
L
.
V
e´
l
e
z
e
t
a
l
.
/
A
c
c
o
u
n
t
i
n
g
,
O
r
g
a
n
i
z
a
t
i
o
n
s
a
n
d
S
o
c
i
e
t
y
3
3
(
2
0
0
8
)
9
6
8
–
9
9
4
9
8
5
Table 8 (continued)
Type of
control
Events Control
mechanisms
Competence-based trust Goodwill-based trust
Channel towards
manufacturer
? Manufacturer
towards channel
Channel towards
manufacturer
? Manufacturer towards
channel
2002:
Agents’ web
page
Norms and
procedures to
limit actions.
Action
accountability.
Pre-action
review
It increases knowledge
about the
manufacturer.
Through it, the
channel improves
market control.
Communication takes
place in a common
language (Das &
Teng, 2001; Morgan
& Hunt, 1994)
It quali?es the channel,
transferring knowledge
and capacities (Sako,
1992). It allows CMD to
externalise new activities,
and to see the market
through the channel’s
eyes
It builds the perception of
integration in the IOR
(Dwyer & Oh, 1987).
Unilateral investment
creates routines towards
long-term integration. It
o?ers assistance and
technical support (Anderson
& Weitz, 1989; Doney &
Cannon, 1997)
It increases
communication, allowing
better prediction of
behaviours. Giving to
receive reduces
uncertainty, and quali?es
the channel
Results
control
1999:
Agents’
marketing
plan
Agents control
scorecard.
Feedback
control
It increases knowledge
about the
manufacturer.
Through it, the
channel improves
market control. The
results obtained are
satisfactory (Kumar
et al., 1995; Selnes,
1998)
It quali?es the channel,
transferring knowledge
and capacities (Sako,
1992). It improves
market and channel
control. The results
obtained are satisfactory.
It increases knowledge
about the channel’s
ability to perform
Assistance and technical
support allow increases in
social contacts, opening a
forum for debate
(Lang?eld-Smith & Smith,
2003). It creates equitable
and fair treatment (Kumar
et al., 1995; Ring & Van de
Ven, 1992)
Improving communication,
the manufacturer sees the
channel members as
collaborators. The forum
for debate establishes and
aligns objectives with each
agent, reducing the
perception of uncertainty.
It allows externalising new
activities
Agents
database.
Feedback
control
It increases knowledge
about the channel,
improving market and
channel control. The
results obtained are
satisfactory
>
2001: Agent
evaluation
system
Evaluation
system.
Feedback
control
It increases knowledge
about the
manufacturer. The
results obtained are
satisfactory
It quali?es the channel,
transferring knowledge
and capacities (Sako,
1992). The results
obtained are
satisfactory. It increases
knowledge about the
channel’s ability to
perform, improving
market and channel
control
Assistance and technical
support allow increases in
social contacts (Bradach &
Eccles, 1989). It creates a
perception of equitable and
fair treatment, reducing the
perception of uncertainty. It
creates a common system of
values and norms (Sako,
1992). The channel
evaluates the
manufacturer’s ability to
appreciate complementarity
It improves
communication (Das &
Teng, 2001); increasing
social contacts (Bradach &
Eccles, 1989). Creating a
common system of values
and norms (Sako, 1992),
the manufacturer
appreciates that the
channel accepts the
interdependencies
9
8
6
M
.
L
.
V
e´
l
e
z
e
t
a
l
.
/
A
c
c
o
u
n
t
i
n
g
,
O
r
g
a
n
i
z
a
t
i
o
n
s
a
n
d
S
o
c
i
e
t
y
3
3
(
2
0
0
8
)
9
6
8
–
9
9
4
2003: Joint
targets and
improvement
projects
Common
objectives.
Feedback
control
They increase knowledge
about the manufacturer.
Through them, the channel
improves market control.
The results obtained are
satisfactory
They qualify the channel,
transferring knowledge and
capacities (Sako, 1992). The
results obtained are
satisfactory. They increase
knowledge about the
channel’s ability to perform,
improving market and
channel control
They create a forum for
debate that shows
interest and sensitivity
towards the channel,
facilitating the
development of mutual
interests. They stimulate
participation in decision-
making, allowing a
higher congruence
between goals (Anderson
& Weitz, 1989). They
create a common system
of values and norms
(Sako, 1992)
Improving two-way
communication, they
create common interests,
increasing social contacts
(Bradach & Eccles, 1989).
The forum for debate
allows better prediction of
behaviours. The channel is
willing to provide
complementary services
and investments (Doney &
Cannon, 1997)
Personnel-
cultural
control
1997: Agents’
?rst meeting and
marketing plan
1999
Agents’
meetings
They increase knowledge
about the manufacturer.
They allow the formalization
of the IOR and allow an
increase in regular contacts
(Bradach & Eccles, 1989;
Dwyer & Oh, 1987; Sako,
1992)
They qualify the channel,
transferring knowledge and
capacities (Sako, 1992)
This unilateral
investment improves
communication with
di?erent departments,
increasing social
contacts. Creating a
common system of
values and norms, and
encouraging
participation in decision
taking, it reduces
uncertainty
Socialization proceeds
through the development
of social contacts and the
active participation of the
channel. It creates a
common system of values
and norms
Agent
pro?le
Selection is based on
knowledge and abilities
(Sako, 1992). The new pro?le
allows the manufacturer to
de?ne its expectations about
the channel
Channel members
perceive the
manufacturer’s trust in
them, and it increases
their trust in the
manufacturer
Selection is based on
reputation (Anderson &
Weitz, 1989)
Satisfaction
survey
It allows the formalization of
the IOR
Results are good. The
manufacturer receives the
channel’s opinion and
concerns
Through this unilateral
investment, the channel
members perceive the
manufacturer’s concern
for their satisfaction
Increasing information
shows good results. The
manufacturer receives the
channel’s opinion and
concerns
Training
plans
They increase knowledge
about the manufacturer
They qualify the channel,
transferring knowledge and
capacities (Sako, 1992)
They improve
communication, enhance
social contacts, and
create a common system
of values and norms
Socialization proceeds
through communication,
social contacts, and a
common system of values
and norms
M
.
L
.
V
e´
l
e
z
e
t
a
l
.
/
A
c
c
o
u
n
t
i
n
g
,
O
r
g
a
n
i
z
a
t
i
o
n
s
a
n
d
S
o
c
i
e
t
y
3
3
(
2
0
0
8
)
9
6
8
–
9
9
4
9
8
7
could assess CMD competences, increasing their
trust in the expertise of the leader ?rm. For
CMD, MCS supplied a track record and an objec-
tive evaluation process for the new tasks, which is
better for generating trust than a subjective evalu-
ation process (Das & Teng, 1998; Dekker, 2004).
This objective information allowed CMD to
remove incompetent agents and to concentrate
on medium-sized agents that it considered better
prepared to ful?ll their responsibilities (type B in
CMD’s classi?cation, see Tables 5 and 6). More
directly, MCSs improved CMD’s trust by training
the survivor channel members. Through the agent
pro?le de?nition, CMD standardized its processes
of agent recruitment and contracting, and there-
fore guaranteed the required competence in those
agents who accepted the new tools and pro?le
changes, and remained in the IOR.
Contravening the opinion expressed by Das and
Teng (2001) and Inkpen and Curral (2004) action
and result controls, as well as personnel-cultural
controls, had favourable e?ects on competence-
based trust. This ?nding may be due to the good
results obtained by both parties, and to the di?er-
ence in size and business knowledge between CMD
and the agents. Channel members recognized
CMD’s superior judgment and accepted some loss
of autonomy as the price of increased abilities to
attend to clients and to make better-informed deci-
sions inside the established limits. Furthermore,
through the client satisfaction surveys and contin-
ual contacts, the channel members still found it
possible to demonstrate their levels of service and
ability, and at the same time to communicate their
needs and areas for improvement.
MCS types and goodwill-based trust
As Table 8 shows, using personnel-cultural con-
trols, the manufacturer was able to transmit its
values and transfer its image to the channel, and
it increased its con?dence that channel members
would act in the best interests of CMD because
they felt integrated into the company. Addition-
ally, because both parties worked with the same
systems and information (action and result con-
trols), both parties improved their integration
and their perception of each other as trustworthy
business partners. In the process of discussing
and establishing joint objectives, CMD showed
its sensitivity towards the channel by creating a
forum for debate. This action strengthened the
channel members’ perception that CMD was act-
ing in their best interest, increasing their trust
(see Lang?eld-Smith & Smith, 2003). At the same
time, the result controls gave CMD objective evi-
dence about how much each agent wished to pro-
gress in the IOR. Furthermore, formal controls
reduced channel members’ uncertainties about
the criteria on which they were evaluated. MCSs
allow broader evaluation criteria, which are per-
ceived as more objective, equitable, and fair.
Several authors (Anderson & Weitz, 1989;
Doney & Cannon, 1997; Sako & Helper, 1998)
have stated that investment in resources dedicated
to the partner makes mutual commitment credible
and so reinforces trust. The unilateral introduction
of computers, software, the evaluation system,
rewards, and training constituted a high degree
of speci?c investment in the channel, which
encouraged agents to consider themselves as key
players and to trust CMD’s intentions. Agents
considered the sharing of MCS information to be
risky behaviour for CMD, and therefore a signal
of trustworthiness (cf. Das & Teng, 1998).
In addition, the reduction in the number of agents
could be seen as a goodwill-based trust builder. The
reduction of bigger agents (type A in CMD’s classi-
?cation, see Tables 5 and 6), who had greater nego-
tiating power, increased CMD’s trust in the
surviving agents, whom it perceived to be more inte-
grated and committed to CMD and to have less
power of negotiation. In return, medium-sized
agents regarded CMD as a major manufacturer
‘‘who looks after” them. The surviving agents were
satis?ed with the higher level of business delegated
and commented that their trust in the manufacturer
was increased when they were reselected.
Again, we found that both formal (action and
result) and informal (personnel-cultural) controls
had favourable e?ects on this more di?use trust
dimension. According to Das and Teng (2001)
and Inkpen and Curral (2004), formal controls
establish rules and procedures to limit the potential
autonomy of the trustee. However, by accepting
information, tools, and more complex tasks,
CMD’s agents increased their independence in their
988 M.L. Ve´ lez et al. / Accounting, Organizations and Society 33 (2008) 968–994
geographical areas, their control over their markets,
and their involvement in the decision-making pro-
cess. Barney and Hansen (1994) argue that MCSs,
if accepted, generate trust. In our case, the absence
of resistance, the willingness to accept the gradual
implementation of MCSs, and even the channel
members’ active participation in work groups,
where they o?ered ideas and knowledge, contrib-
uted to CMD’s favourable perception of their inten-
tions and dispelled any doubts that may have been
harboured: ‘‘they are trustworthy and we should
not fear that they will leave for competitor ?rms”.
Analysis of the association between each MCS
function and trust
Abernethy and Vagnoni (2004) assigned the
coordination function to subunits that manage
day-to-day activities, and the monitoring function
to superiors who use information to control the sub-
units’ behaviour. CMD developed and used MCSs
to coordinate and monitor the channel’s behaviour,
whereas agents used MCSs to coordinate tasks with
CMD. That is, both parties used MCSs to coordi-
nate new and more interdependent tasks, a process
that encouraged commitment (Anderson & Weitz,
1992) and created an atmosphere of assistance and
respect. MCS development creates a frame where
information ?ows to facilitate problem solving,
resolve disputes promptly, improve IOR ?uency,
and manage interdependencies. CMD, ‘‘giving in
order to receive” through newMCSs, improved task
coordination, better supported distribution channel
management, and, at the same time, kept control of
the endmarket, improving joint results and satisfac-
tion. Our results con?rm that MCSs generate trust
when they are used to coordinate activities by
improving the provision of relevant information
(cf. Morgan & Hunt, 1994; Sako, 1992).
Likewise, faced with greater dependency, CMD
needed to supervise the channel’s activities in its
end market in order to con?ne the scope of the
IOR. As Tomkins (2001, p. 171) claimed, one
can never dispense completely with the need for
some information on the other party’s actions
unless trust is absolute, blind and unthinking.
Hence, CMD used its MCSs to evaluate its chan-
nel members’ behaviours in their new functions
and to see the end market through their eyes. As
the agents’ evaluations were satisfactory, the
supervision reinforced CMD’s trust in the surviv-
ing channel members.
Although CMD used every MCS tool both to
coordinate and to supervise, because all the tools
except the agent database were shared, the channel
perceivedmuchmore coordinationthansupervision,
and MCS development was considered necessary,
thus avoiding possible suspicions (Sitkin, 1995).
Importantly, the interviews clearly revealed that
commercial employees sold the new MCSs to the
channel. Some authors have argued that trust
between individuals is critical for IOR cooperation,
and is key to developing goodwill-based trust
between ?rms (Das & Teng, 2001; Ring & Van de
Ven, 1994). The pre-established trust that the chan-
nel hadinthe commercial employees—insome cases,
it is even possible to speak of friendship—contrib-
uted to a positive attitude towards new MCSs and
to their acceptance as systems of assistance rather
than of supervision. When monitoring was per-
ceived, surviving agents saw it as a form of protec-
tion against unfair reward sharing and as a means
of gaining recognition for their work (Ring & Van
de Ven, 1992), and for that reason the MCS-trust
relationship maintained an additive association.
Analysis of the association between MCSs
development and already established trust in
evolving inter-organizational relationships
Our case study reveals that, in a mature and
close IOR, the introduction of new MCSs does
not necessarily reduce trust. To explain this from
an evolutionary perspective, we explore how these
two critical factors favour cooperation and the
IOR evolution processes so that they are self-rein-
forcing in a series of positive loops. We found that,
although the relationship between the manufac-
turer and its distribution channel was longstand-
ing, its objectives, activities, dependencies, and
assumed risks evolved over the period of study.
These evolutions caused the level of con?dence
required by the parties to be dynamic and have
no zero-sum constraints (Das & Teng, 1998), and
this might have been the reason why the MCSs
increased trust.
M.L. Ve´ lez et al. / Accounting, Organizations and Society 33 (2008) 968–994 989
The initial conditions of an IOR are shaped by
an assessment of the likelihood that the partner
will behave in a trustworthy manner (Inkpen &
Curral, 2004), which is in turn based on judgments
about the partner’s goodwill and/or competence.
However, as Arin˜ o and de la Torre (1998) argued,
the conditions dictating the form of the IOR will
evolve over time, and such developments may also
prompt adaptation in the IOR’s governance struc-
ture. In our case, the pre-existing trust encouraged
CMD to externalise tasks in order to create a lea-
ner structure in the context of its losses in 1992,
against the advice of the consultant. When, mainly
after 1997, CMD intentionally externalised admin-
istrative tasks, these tasks altered the required
channel capacities. The greater outsourcing also
increased CMD’s dependence on its channel and
the perceived risk of opportunistic behaviours
(Van der Meer-Kooistra & Vosselman, 2000). Fur-
thermore, this increased interdependence led the
agents to demand more information to allow them
to better manage their activities. The level of con-
?dence required in the IOR a?ected both parties,
and they saw the introduction of new MCSs as
necessary to facilitate the coordination and moni-
toring of the IOR’s desired evolution.
The MCSs encouraged cooperation in the new
conditions and made it visible to both parties. In
a series of positive loops, MCSs built trust directly
and reinforced it indirectly, in the terms described
by Coletti et al. (2005). With the growth of trust,
there is an increasing willingness to accept risk
and to augment the IOR’s scope (Inkpen & Curral,
2004). Thus, increased trust made CMD more will-
ing to expand its externalisation, and encouraged
the channel members to accept the new activities.
The new and more complex tasks again shifted
the required channel capacities and the risks per-
ceived by CMD. The reduction in the number of
agents during the study period increased the prox-
imity of the IOR, and thus channel involvement
and dependency, in turn requiring new MCSs for
coordination and monitoring. Therefore, by facili-
tating cooperation, the introduction of MCSs
increased trust, directly and indirectly.
Our empirical ?ndings concur with the theoret-
ical statements of Inkpen and Curral (2004) that,
paradoxically, in an established IOR, increments
in trust may lead to a greater need for control. Ink-
pen and Curral argue that greater control would
be carried out through cultural mechanisms. How-
ever, the CMD case provides evidence that all
three types of MCSs (not only personnel-cultural
ones) can develop and maintain the trust that the
new IOR conditions demand. According to Dek-
ker’s (2004) proposal, in every time moment, trust
will be damaged only when the use of formal con-
trol exceeds the need for control that the IOR itself
generates. Our results seem to verify that the exist-
ing trust did not cause the introduction of new
MCSs to be interpreted as a lack of trust, but pos-
itively in?uenced the perceptions of both parties.
Fig. 2 modi?es Tomkins’s ?gure to show that
the association between aggregate MCSs and trust
increases over time. As Tomkins (2001) argues, at
an early stage of an IOR, the need for either MCSs
or trust is lower. However, when the IOR matures
and trust reaches higher levels, the success of IOR
exchanges encourages the partners to engage in
riskier operations and stronger cooperation. In
the CMD case, this encouraged constant develop-
ments in the IOR itself. These developments, as
pinch points (Fig. 2), needed greater con?dence
in cooperation and, in turn, demanded more
MCSs and greater levels of trust to support and
maintain cooperation, and so on. Our case study
shows that a manufacturer can use MCSs not only
to enable the conditions that favour trust to be
built, but as a tool to build trust. When the IOR
character evolves, the initial conditions also
develop, a?ecting the information needed for con-
trol of the IOR and trust, and this explains why
the association between MCSs and trust continues
to be positive, building trust.
Conclusions
This case study provides evidence that, even
when trust is well established, MCSs enable condi-
tions that favour trust, and directly build trust. In
mature, open-ended IORs, high trust and stability
provide a platform where the initial conditions
evolve. Success encourages the partners to cooper-
ate further, requiring greater con?dence and, in
turn, demanding more MCSs and greater levels
of trust to support cooperation. In this setting,
990 M.L. Ve´ lez et al. / Accounting, Organizations and Society 33 (2008) 968–994
where the level of con?dence is dynamic, the leader
?rm can develop MCSs to reach the new intended
stage without damaging the existing trust.
From both sides of the IOR, the ?ndings
illustrate that it is not only informal controls that
act as trust builders, as claimed by Das and
Teng (2001). Action and result controls do not nec-
essarily cause in?exibility, but can improve agents’
independence and increase their chances to demon-
strate their competences inside the established lim-
its. Action and result controls provide evidence
with a greater appearance of objectivity. Further-
more, because both parties work with the same
systems (action and result controls), formal
MCSs improve their perception of each other’s
trustworthiness.
We found that both parties used MCSs to coor-
dinate the new complex tasks, which encouraged
commitment and created trust. MCSs create a
frame to advance the ability to solve problems,
and to manage interdependencies. The leader ?rm,
‘‘giving in order to receive”, at the same time kept
control of the end market, improving the joint
results and satisfaction. Although the manufacturer
used the same MCS tools to coordinate and to
supervise, owing to their shared nature, the channel
perceived much more coordination than monitor-
ing, and this avoided possible suspicions that could
damage trust. Furthermore, the manufacturer’s use
of non-coercive tactics helped the channel members
to see monitoring as a mechanism for overall
improvement, as protection against unfair reward
sharing, and as recognition of their work.
The results and the constraints of our study sug-
gest avenues for further research. First, our case
study highlights that agents’ trust in the commer-
cial employees contributed to their ready accep-
tance of new MCSs and in?uenced their
perception of these as systems of assistance; it
would be interesting to analyse the role of MCSs
in facilitating the transmission between interper-
sonal and inter-organizational trust by developing
routines and cultural processes. Second, this paper
focuses on how MCSs can create positive expecta-
tions about the trustee’s characteristics. Future
research could use other di?erent constructs of
trust, such as system trust (Luhmann, 1979) or
institutional based trust, and could observe how
MCSs can generate institutional trust and give
the trustor con?dence in the trustee, over the evo-
lution of the IOR. Finally, the case study results
show that MCSs have acted as trust builders even
when trust was well-established. The conditions of
the case have allowed us to study this association
centring on common aspects of the relationships
between the manufacturing ?rm and its 176 distri-
bution channel members. An extension of this
work would carry out another research strategy
to consider every dyadic relationships in order to
analyse di?erent rates of growth in trust provoked
by the MCSs development, as well as moderating
variables that can explain such di?erences.
Acknowledgements
We gratefully acknowledge the contributions of
the two anonymous reviewers for their insightful
suggestions. It is di?cult to overstate our gratitude
for the comments received on earlier versions of
this paper from Henry C. Dekker, and the partici-
pants at the AOS-Bocconi Conference on Sustain-
ing Organizational Combinations, and at
management accounting seminars of the Work-
shop ‘‘Memorial Raymond Konopka”. We also
would like to thank the participants of the case
study companies for their time and data gener-
ously provided.
Fig. 2. Proposed relationship between MCSs and level of trust
increasing over time.
M.L. Ve´ lez et al. / Accounting, Organizations and Society 33 (2008) 968–994 991
Appendix
See Tables A.1–A.3.
Table A.1
Documents analysed during the research period
Documents analysed
Agent selection norm
Agents control scorecard. Improvement project report
Agents control scorecard report
Channel Training Plan report
Journals ‘‘Product News” for agents, and ‘‘CMD Magazine” for sta?, agents and clients
Agents control scorecard (?scal years 1999–2004)
Software ‘‘Agents control scorecard”
Software ‘‘Gescom”, commercial information for commercial department
Channel satisfaction survey results (1999 and 2001)
Report on Agents’ Marketing Plan
Report on the presentation of Agents Evaluation System
Agents database
Slides of videoconference from presentation of Agents Evaluation System
Agents Evaluation System results. Years 2002 and 2003
Slides from Agents’ meetings, years 2001, 2002 and 2003
CMD’s Annual Reports from 1989 to 2002
Every type of contract from the period studied
CMD’s earning reports (?scal years 1997–2002)
Information extracted from CMD’s web page
CMD’s sta? emails and letters
Table A.2
CMD interviews
Number Position Description Time
1 District Manager Head of zone, with 28 agents. Former commercial employee, with 21 years at CMD 4 h
2 CAC Manager Commercial department. Manager of Client Attention Centre, with 55 agents and
29 years at CMD
3 h
3 CMD CEO Director. CMD general manager. Former controller, commercial employee, district
manager, and commercial director, with 14 years at CMD
2 h
4 Former CMD CEO Director. Former CMD general manager, in the period 1994–2001, with 16 years at
CMD
3 h
5 CAC Director Commercial department. Head of Client Attention Centre and the superior of
agents manager. Former commercial employee, with 23 years at CMD
3 h
6 Commercial Director Commercial department. Head of commercial department. Former commercial
employee, with 22 years at CMD
2 h
7 Former Agents Manager and
former Marketing Director
Commercial department. Former manager and designer of agents’ marketing plan.
Former controller, with eight years at CMD
3 h
8 Sales Manager Commercial department. Former commercial employee, with 16 agents. 21 years at
CMD
3 h
9 Agents Manager Commercial department. Former commercial employee, with eight years at CMD 3 h
10 Logistics Director Head of logistics department, with 19 years at CMD 2 h
11 Former Agents Manager 2 Former head of agents management. Former commercial, with 22 years at CMD 2 h
12 Former CMD CEO 2 Former CMD CEO in the period 1986 to 1993, with 39 years at CMD 3 h
992 M.L. Ve´ lez et al. / Accounting, Organizations and Society 33 (2008) 968–994
References
Abernethy, M. A., & Vagnoni, E. (2004). Power, organization
design and managerial behaviour. Accounting, Organiza-
tions and Society, 29, 207–225.
Andaleeb, S. S. (1995). Dependence relations and the moder-
ating role of trust: Implications for behavioral intentions in
marketing channels. International Journal of Research in
Marketing, 12, 157–172.
Anderson, E., & Weitz, B. (1989). Determinants of continuity in
conventional industrial channel dyads. Marketing Science,
8(4), 310–323.
Anderson, E., & Weitz, B. (1992). The use of pledges to build
and sustain commitment in distribution channels. Journal of
Marketing Research, 29(1), 18–34.
Arin˜ o, A., & de la Torre, J. (1998). Learning from failure:
Toward an evolutionary model of collaborative ventures.
Organization Science, 9(3), 306–325.
Barney, J., & Hansen, M. (1994). Trustworthiness as a source of
competitive advantage. Strategic Management Journal, 15,
175–190.
Bradach, J. L., & Eccles, R. G. (1989). Market versus
hierarchies: From ideal types to plural forms. In W. R.
Scott (Ed.), Annual review of sociology (Vol. 15, pp. 97–
118).
Coletti, A. L., Sedatole, K. L., & Towry, K. L. (2005). The
e?ect of control systems on trust and cooperation in
collaborative environments. The Accounting Review, 80(2),
477–500.
Curral, S. C., & Inkpen, A. C. (2002). A multilevel approach to
trust in joint ventures. Journal of International Business
Studies, 33, 479–495.
Das, T. K., & Teng, B. S. (1998). Between trust and control:
Developing con?dence in partner cooperation in alliances.
Academy of Management Review, 23(3), 491–512.
Das, T. K., & Teng, B. S. (2001). Trust, control and risk in
strategic alliances: An integrated framework. Organization
Studies, 22(2), 251–283.
Dekker, H. (2004). Control of inter-organizational relation-
ships: Evidence on appropriation concerns and coordina-
tion requirements. Accounting, Organizations and Society, 1,
27–49.
Doney, P., & Cannon, J. P. (1997). An examination of the
nature of trust in buyer–seller relationships. Journal of
Marketing, 61(April), 35–51.
Doz, Y. L. (1996). The evolution of cooperation in strategic
alliances: Initial conditions or learning processes. Strategic
Management Journal, 17, 55–83.
Dwyer, F. R., & Oh, S. (1987). Output sector muni?cence
e?ects on the internal political economy of marketing
channels. Journal of Marketing Research, 24(4), 347–358.
Ganesan, S. (1994). Determinants of long-term orientation
in buyer–seller relationships. Journal of Marketing, 58,
1–19.
Ganesan, S., & Hess, R. (1997). Dimensions and levels of trust:
Implications for commitment to a relationship. Marketing
Letters, 8(4), 439–448.
Gulati, R., Lawrence, P. R., & Puranam, P. (2005). Adaptation
in vertical relationships: Beyond incentive con?ict. Strategic
Management Journal, 26, 415–440.
Halinen, A., Salmi, A., & Havila, V. (1999). From dyadic
change to changing business networks: An analytical
framework. Journal of Management Studies, 36(6),
779–794.
Inkpen, A. C., & Curral, S. C. (2004). The coevolution of trust,
control, and learning in joint ventures. Organization Sci-
ence, 15(5), 586–599.
Kaju¨ ter, P., & Kulmala, H. I. (2005). Open-book accounting in
networks: Potential achievements and reasons for failures.
Management Accounting Research, 16(2), 179–204.
Table A.3
List of agents interviewed
Number Type of agent Description Time
1 Agent A-1 Length of relationship 14 years. Principal activity: CMD agent. Share of business
with CMD: 50%
4 h
2 Agent A-2 Length of relationship 10 years. Principal activity: Industrial supplies. Share of
business with CMD: 40%
3 h
3 Agent B-1 Length of relationship 14 years. Principal activity: CMD agent. Share of business
with CMD: 92%
3 h
4 Agent B-2 Length of relationship 10 years. Principal activity: Industrial supplies. Share of
business with CMD: 10%
4 h
5 Agent B-3 Length of relationship 13 years. Principal activity: CMD agent. Share of business
with CMD: 80%
3 h
6 Agent C-1 Length of relationship 14 years. Principal activity: CMD agent. Share of business
with CMD: 40%
4 h
7 Agent C-2 Length of relationship 40 years. Principal activity: Motor repairs. Share of business
with CMD: 15%
3 h
8 Agent C-3 Length of relationship 12 years. Principal activity: Industrial supplies. Share of
business with CMD: 30%
3 h
M.L. Ve´ lez et al. / Accounting, Organizations and Society 33 (2008) 968–994 993
Knights, D., Noble, F., Vurdubakis, T., & Willmott, H. (2001).
Chasing shadows: Control, virtuality and the production of
trust. Organization Studies, 22(2), 311–336.
Kumar, N., Scheer, L., & Steenkamp, J. B. (1995). The e?ects of
supplier fairness on vulnerable resellers. Journal of Market-
ing Research, 32(1), 54.
Lang?eld-Smith, K., & Smith, D. (2003). Management control
systems and trust in outsourcing relationships. Management
Accounting Research, 14, 281–307.
Luhmann, N. (1979). Trust and power. New York: John Wiley.
Mayer, R. C., Davis, J. H., & Schoorman, F. D. (1995). An
integrative model of organizational trust. Academy of
Management Review, 20(3), 709–734.
Merchant, K. A., & Van der Stede, W. A. (2003). Management
control systems, performance measurement, evaluation and
incentives. Harlow, England; New York: FT Prentice Hall.
Moorman, C., Deshpande´, R., & Zaltman, G. (1992). Factors
a?ecting trust in market research relationships. Journal of
Marketing, 57, 81–101.
Morgan, R. M., & Hunt, S. D. (1994). The commitment-trust
theory of relationship marketing. Journal of Marketing, 58,
20–38.
Neu, D. (1991). Trust, contracting and the prospectus process.
Accounting, Organizations and Society, 16(3), 241–256.
Nooteboom, B., Berger, H., & Noorderhaven, H. G. (1997).
E?ects of trust and governance on relational risk. Academy
of Management Journal, 40(2), 308–338.
Poppo, L., & Zenger, T. (2002). Do formal contracts and
relational governance function as substitutes or comple-
ments?. Strategic Management Journal 23, 707–725.
Ring, P. S., & Van de Ven, A. H. (1992). Structuring
cooperative relationships between organizations. Strategic
Management Journal, 13, 483–498.
Ring, P. S., & Van de Ven, A. H. (1994). Developmental
processes of cooperative interorganizational relationships.
Academy of Management Review, 19(1), 90–118.
Rousseau, D., Sitkin, S., Burt, R., & Camerer, C. (1998). Not so
di?erent after all: A crossdiscipline view of trust. The
Academy of Management Review, 23(3), 393–404.
Sako, M. (1992). Prices, quality and trust: Inter-?rm relations in
Britain and Japan. Cambridge: Cambridge University Press.
Sako, M., & Helper, S. (1998). Determinants of trust in supplier
relations: Evidence from the automotive industry in Japan
and the United States. Journal of Economic Behavior and
Organization, 34, 387–417.
Seal, W., Cullen, J., Dunlop, A., Berry, T., & Ahmed, M.
(1999). Enacting a European supply chain: A case study on
the role of management accounting. Management Account-
ing Research, 10(3), 303–322.
Selnes, F. (1998). Antecedents and consequences of trust and
satisfaction in buyer–seller relationships. European Journal
of Marketing, 32(3/4), 305–322.
Sitkin, S. B. (1995). On the positive e?ect of legalization on
trust. Research on Negotiation in Organization, 5,
185–217.
Sohn, J. H. D. (1994). Social knowledge as a control
system: A proposition and evidence from the Japanese
FDI behavior. Journal of International Business Studies,
2Q, 295–324.
Stern, L., El-Ansary, A., & Coughlan, A. (1996). Marketing
channels (5th ed.). Upper Saddle River, NJ: Prentice Hall.
Tomkins, C. (2001). Interdependences, trust and information in
relationships, alliances and networks. Accounting, Organi-
zations and Society, 26, 161–191.
Van der Meer-Kooistra, J., & Vosselman, G. J. (2000).
Management control of inter?rm transactional relation-
ships: The case of industrial renovation and maintenance.
Accounting, Organizations and Society, 25, 51–77.
Van der Meer-Kooistra, J., & Vosselman, G. J. (2006).
Research on management control of inter?rm transactional
relationships: Whence and whither. Management Account-
ing Research, 17(3), 227–237.
Woolthuis, R. K., Hillebrande, B., & Nooteboom, B. (2005).
Trust, contract and relationship development. Organization
Studies, 26(6), 813–840.
Yin, R. K. (1984). Case study research: Design and methods.
Beverly Hills, CA: Sage Publications.
994 M.L. Ve´ lez et al. / Accounting, Organizations and Society 33 (2008) 968–994

doc_817874953.pdf