Our first aim in this paper is to map the dynamics in vertical
and horizontal control packages in municipal JVs and
to investigate the relational factors that affect them. Based
on our three case studies of municipal JVs in the waste
disposal business, all of which were founded to exploit
economies of scale, we conclude that four relational factors
affect vertical control packages. A higher degree of vertical
goodwill and competence trust reduces the need for
broad and tight action controls, and (inversely) a reduction
in vertical goodwill trust may result in broader and tighter
action control and results control. Broader results control
and action control (although not necessarily more tightly
used) may be caused by parent differences in size (in combination
with control competence) and in management
style. Parent diversification, that is, a limited relationship
between the JV’s activity and the parents’ other activities,
leads to narrow and loose control since it increases
information asymmetry, reduces the need for coordination,
and dilutes parents’ attention (especially if the activity is
politically uninteresting). Also, a well-designed horizontal
control package (e.g., rules for parent interaction and
the distribution of work) leads to proactive controls and
a readiness to impose new formal controls in the vertical
dimension. Regarding the horizontal control packages, we
conclude that the existence of goodwill trust, system trust
and calculative trust limit control to fairly narrow and loose
action control. We also conclude that parent differences in
size make it easier to develop structures for parent interaction
and social control, and that equality of treatment
among the parents may drive horizontal action control.
Our second aim in this paper is to contribute to the
discussion on the relationship between trust and control
in inter-organizational relationships. In our case studies,
we observed that over time goodwill trust and competence
trust made action control narrower and looser
and that diminished trust was followed by broader and
tighter action control and results control. We found many
instances where new, formal control did not destroy trust.
This conclusion supports the theoretical view that new,
formal controls may be implemented without damage to
established trust, that is, without signalling distrust. Furthermore,
we observed an inverted “crowding out” effect
of trust on control (i.e., the high ambition to maintain
trust may lead to underdeveloped formal controls). Trust
appears to lead to loose and narrow control that perhaps is
suitable for the present situation but insufficient for future
contingencies. This may be because trust inhibits consideration
of the need for control and because control is thought
to endanger goodwill trust. Well-designed horizontal controls
make it easier to deal with emerging control gaps in
vertical relationships.
In addition, we found that the ambition to preserve trust
may inhibit the realisation of economies of scale. JV Companies
that value autonomy and manage their activities to
preserve vertical and horizontal trust may not take advantage
of economically justified opportunities. This strategy
contradicts the very reason for the creation of municipal
JVs.
An important generalization of this study is that the
relationship between control and trust is multifaceted.
However, this generalization (as well as the interrelations
presented above) should be regarded as tentative due to
the nature of case study research. While we offer some
insights about how control processes create or impair trust,
we call for additional studies in this area. We conclude that
formal control is not a homogenous factor that influences
trust. We need to pay attention to the different processes
and techniques of control. In general, research needs to
acknowledge the potential of control as a positive force
in coordination and as an indication of trust. Furthermore,
our choice of case studies narrows our findings to the situation
of fairly homogeneous parents. Future research may
determine if vertical and horizontal control packages for
heterogeneous parents are broader and tighter.