Building thick trust in inter-organizational relationships: the role of relational signaling and formal controls

(1)

Building thick trust in inter-organizational relationships: the

role of relational signaling and formal controls

Brenden B.J. Vorenkamp

S2570602

MSc. BA – Organizational & Management Control University of Groningen

Faculty of Economics and Business Submitted: 25 June 2018

Supervisor: A. Rehman Abbasi

Co-assessor: prof. dr. J. van der Meer-Kooistra

(2)

2 ABSTRACT

(3)

3

4 1. INTRODUCTION

In essence, inter-organizational relationships (IOR) create value for companies through synergy gains and competitive advantages (Coletti, Sedatole, and Towry, 2005). However, several researchers notice that many of these collaborations eventually fail (Coletti et al., 2005; Faems, Janssens, Madhok, and Van Looy, 2008). Many factors attribute to these failures, of which some are predictable and/or possible to prevent and some are not. For instance, Coletti et al. (2005) and Das and Teng (2001), argue that even if partners act in accordance to an agreement, a collaboration can fail due to events happening in the environment such as changes in market competition and demand. These risks are not always manageable. Risks that can be managed in some extent come from within the relationship and are defined as relational risks, which includes for instance incentives to act in self-interest, also known as opportunistic behaviour. This behaviour results in lack of cooperation and eventually underperforming relationships. Relational risks are unique to a collaboration and in contrast to environmental risks easier to manage with control mechanisms.

Research in the field of IOR performance has shown that trust is an important mechanism for collaborations to become and maintain a success (Seppänen, Blomqvist, and Sundqvist, 2007). Trust is not seen as a mechanism to prevent relational risks, but is the result of one having positive expectations of the other party. In other words, one experiences the risks to be limited. The associated challenge for managers is to come to an optimal level of control to mitigate the risks while keeping the trust level high (Smets, Van Oorschot, and Langerak, 2013).

(5)

5

dynamics of trust in IORs by making causal inferences. Therefore, the purpose of this study is to get a better understanding of the factors that build high trust levels in an IOR, also known as thick trust (Vosselman and Van der Meer-Kooistra, 2009).

One of the factors examined in this research is control mechanisms. As mentioned before, managers should exercise control mechanisms in a way that builds or maintains thick trust. However, how control mechanisms and thick trust relate to each other and whether they can be seen as substitutes or complements, is still to debate. Early literature on this topic shows mixed evidence towards trust being a substitute for control (Das and Teng, 1998), while recent literature shows results towards a complementary relationship (Poppo and Zenger, 2002; Vosselman and Van der Meer-Kooistra, 2009; Woolthuis, Hillebrand and Nooteboom, 2005). The debate currently seems to be ended with the notion that the two are more than just substitutes and their association is more complex than initially thought (Vosselman and Van der Meer-Kooistra, 2009; Costa and Bijlsma-Frankema 2007). This study will contribute to the control-trust nexus by looking into the influence of control mechanisms on thick trust building. By considering the current insights in the literature on the control-trust nexus, an inverse u-shape relation is proposed and tested, adapted from Dekker (2004).

Formal control is not the only factor that may have important implications for the building of thick trust. Vosselman and Van der Meer-Kooistra (2009) proposed a framework including the importance of relational signals. Relational signaling comes from voluntary local decisions which stem from outside the contractual agreement and shows one is committed to the relationship. It is proposed to enhance thick trust building. To make clear, signals resulting from formal controls are not voluntary and therefore not considered as relational signals. The relationship between relational signaling and thick trust, proposed by Vosselman and Van der Meer-Kooistra (2009), is extended in this study with opportunistic behaviour. Opportunistic behaviour indicates lack of commitment and hence less relational signals are expected. Furthermore, it is expected that no matter how opportunistic the parties are, if the main goals in an IOR are aligned, thick trust can be build. In this way, the concept of enlightened self-interest will be tested. Besides that, both the effect of opportunistic behaviour and shared goals on thick trust building will be investigated.

It is relevant to understand how control mechanisms and behaviour of partners affect the trust building process, as it defines the success of an IOR. Examining the influence of the discussed factors on thick trust provides new and more generalizable insights in the process of thick trust building and will contribute to the control-trust nexus. The following research question examined in this study comprises the enhancement in understanding thick trust:

RQ: What are important factors in the building of thick trust, which leads to a successful

(6)

6

In order to give answer to this research question, two sub questions were developed explicitly considering formal control mechanisms and relational signaling:

Subquestion 1: Do formal controls have an inverse u-shape relationship to thick trust

Subquestion 2: Does relational signaling lead to thick trust and are shared goals decisive in the

extent of relational signals given?

By empirically testing the influence of formal controls on trust building, this study contributes to the current control-trust debate. An inverse u-shape relation has not been tested before and gives new insights into the building of thick trust. This study also contributes to the current literature by empirically testing the proposed influence of relational signaling on thick trust by Vosselman and Van der Meer-Kooistra (2009). Furthermore, it will contribute to a better understanding of developing and managing inter-organizational relationships.

The remainder of this study is structured as follows. In chapter 2, the concepts of trust, control and relational signaling will be discussed in depth and relevant literature will be reviewed for the development of hypotheses. Chapter 3 will discuss the methodology used in this study, including the data collection method, measurement of constructs, plan of analysis and associated validity and reliability tests. In chapter 4, the test results will be presented. Chapter 5 will answer the research question by relating the findings to the existing literature. Also the theoretical and managerial implications, limitations and suggestions for future research will be discussed.

2. LITERATURE REVIEW

Like Woolthuis et al. (2005) noticed, the disagreement in previous literature on the control-trust debate may be attributed to the confusion about the definition of control and trust. Therefore, I will first elaborate further on the concepts of trust and control used in this research. After that, the framing theory will be reviewed. This theory lies at the foundation of this study together with the conceptual framework of Vosselman and Van der Meer-Kooistra (2009). Next, recent developments in the literature and their relevance for the development of my hypotheses will be explained in depth. The first hypothesis is developed to give an answer to the first sub question. The remaining five hypotheses deal with the second sub question. At last, an overview of the developed hypotheses will be presented in the conceptual framework.

2.1 Concept of trust and control

(7)

7

controls. These formal control mechanisms are proven to be of great importance for the trust building process, but their actual influence is still up to debate and needs more empirical evidence. (Dekker, 2004; Tomkins, 2001; Vélez et al., 2008; Vosselman and Van der Meer-Kooistra, 2009) The views of various scholars, including recent developments in the field, with respect to the control-trust debate will be explored in paragraph 2.2.

Previous literature often links formal controls to transaction cost economics (TCE) theory, because it deals with two of its basic assumptions: bounded rationality and opportunism (Yang et al, 2010; Williamson 1996). Bounded rationality means that parties in an IOR are not able to predict every possible scenario ex ante, making them vulnerable to opportunistic behaviour. Formal controls allow parties to protect themselves from these undesirable situations by monitoring the processes in the relationship and consequently making results and partner’s behaviour more predictable (Vélez et al., 2008; Leifer and Mills, 1996; Williamson, 1985). Similar to TCE, formal controls also offer a solution for the classic agency problem. According to agency theory, people are inclined to pursuit self-interest and to act opportunistic if possible. Including formal controls in the relationship clarifies the degree to which a party complies to the contractual agreement (Houston and Johnson, 2000; Balakrishnan and Fox, 1993) and penalties can be made accordingly (Malhotra and Murnighan, 2002). Also contract theory considers a contract that includes formal controls as an incentive to limit opportunism (Woolthuis et al., 2005).

Formal controls can be subdivided into outcome and behaviour control mechanisms (Dekker, 2004). The function of outcome control is to keep track of achievements of targets and align partner’s interests by motivating them to achieve derisible outcomes through the creation of goal congruence and clarification of mutual expectations (Dekker, 2004; Das and Teng, 1998). Behaviour controls indicate how partners should act in the relationship and whether they have the appropriate behaviour specified in the contract. These controls consist of i.e. planning, rules, and operating procedures (Dekker, 2004). As can be noticed from these two mechanisms, formal controls do not solely function to protect the parties. They also help to coordinate, establish and communicate tasks in the IOR (Vlaar, Van den Bosch, and Volberda, 2007). For instance, when firms are combining their resources, it might be hard to manage the value creation process. Mellewigt, Madhok, and Weibel (2007) argue, from a resource-based view, that formal controls will coordinate the collaboration by clarifying the expectations and enable goal congruence. Therefore, previous literature often uses overarching terms that define the general functions of the two control mechanisms as coordinating and safeguarding (Costa and Bijlsma-Frankema, 2007; Mellewigt et al., 2007; Vélez et al., 2008).

(8)

8

lies at the foundation of a contractual arrangement. Parties in an IOR know that any negative experiences of the partners regarding the behaviour of another will be voiced to a network, leading to penalties of malicious and incompetent behaviour. Together with formal controls, which function as contractual safeguards, the involved parties are protected for the calculated potential opportunistic behaviour. Eventually, thin trust is built and the trustor accepts vulnerability to the potential opportunistic behaviour that could not be foreseen (Rousseau, Sitkin, Burt, and Camerer, 1998). This is consistent with transaction cost theory and contract theory. Both theories consider a contract as the basis for trust since it limits incentives for opportunism (Woolthuis et al., 2005).

The major difference between thin and thick trust is that the latter contributes to positive behavioural expectations by itself while thin trust is just based on reducing negative behavioural expectations (Vosselman and Van der Meer-Kooistra, 2009). The notion of thick trust is in line with Dekker (2004) who defines relational trust as trust that emerges over time by repeated interactions between partners in an IOR such as joint problem solving and decision making. The interactions result in partners having more commitment and interest in a well performing relationship. According to Poppo, Zhou, and Li (2016), relational trust is defined as a situation in which a shared identity between the partners is present and all partners can expect that the partner will act co-operative and according to the other’s preferences (Lewicki et al., 1996). This makes thick trust distinct from thin trust and therefore relational trust is essential for building thick trust. In this study the terms relational trust and thick trust are used interchangeably.

2.2 Framing theory

(9)

9

thing. In a gain frame, the actions are based on rational choice and increasing one’s efficiency and lower costs is the focal goal.

2.3 Formal control mechanisms and trust

The ongoing debate in the IOR literature on whether formal controls function as substitutes or complements for trust suggests the two are inversely related (Dekker 2004). In a substitutive view, relationships with high levels of trust in the partner need less control mechanisms in order to reduce possible opportunism, since both parties act in favour of the relationship (Ouchi, 1979 and Ebers, 2002). It also has economic relevance, because less costs are made regarding the monitoring tools (Cummings and Bromiley, 1996). Some scholars argue that formal controls can even signal misbelief, suspicion and distrust in the other party resulting in less relational trust (Poppo and Zenger, 2002; Das and Teng, 2008; Yang et al., 2010). In a complementary view, scholars suggest formal controls have a positive effect on the building of relational trust, because parties can track the other’s performance and so risk is mitigated. Sako (1992) argues that formal controls create structure which allows information and knowledge to transfer easier, which builds trust. Vélez et al. (2008) also find that an increase in formal control mechanisms enhance cooperation, which indirectly increases trust. Poppo and Zenger (2002) support this view by stating that controls promote expectations of cooperation and lead to commitment in the relationship. The contradicting substitutive and complementary views are proved to be too simplistic and their association is more complex than initially thought. (Vosselman and Van der Meer-Kooistra, 2009)

Vosselman and Van der Meer-Kooistra (2009) propose a framework which examines trust building. In their framework, they state that trust in IORs is built on a contractual arrangement that works as a safeguarding device and incentive system. This arrangement needs institutional trust that originates from outside the relationship, because it offers voice and exit threats. In line with the coordination function and the complementary view of control, formal controls produce trust in the early stage of a relationship, because the behaviours of the parties involved are directed through these formal controls. Institutional trust and formal controls together mitigate as much as possible against potential negative behaviours which results in thin trust.

(10)

10

relationship. Tomkins (2001) also investigated in detail the structure of an inter-organizational relationship by distinguishing four different stages. They suggest that in later stages of a relationship, the activity associated with the relationship needs less information in order to be successful. They conclude that relationships in more ‘stable’ stages have higher trust levels, and therefore an increase in control mechanisms may harm trust. In other words, the complementary effect of formal controls becomes substitutive when trust is higher. This is supported by Smets et al. (2013) who did research to the use of control mechanisms in the manufacturing sector. They concluded that overusing formal controls may decrease trust. They also claim that a “constant level of formal control appears more beneficial than alternating periods of high formal control” (p. 1156). This suggests that the introduction of any type of control mechanism may directly harm trust building after the contractual arrangement took place. Dekker (2004) investigated the effect on trust when formal controls exceed the level necessary to safeguard for opportunistic behaviour. He proposes formal controls to be complementary to trust until a certain threshold after which it will be damaged. Especially if the formal controls are intended to safeguard in more developed relationships, they appear to be substitutive. Therefore, the proposed relationship between control and thick trust is an inverse u-shaped one, meaning that formal controls will harm trust after a certain threshold:

H1: The effect of formal controls on the building of thick trust is inverse u-shaped

2.4 Opportunistic behaviour, relational signaling and trust

As discussed in paragraph 2.2 a normative frame implies one is willing to behave appropriate, trustworthy and is not seeking short-term individual goals. This suspends opportunistic behaviour and has important implications for the building of thick trust. In case the trustee acts opportunistic, trust building will be damaged since it conflicts with any positive expectations the trustor had. The trustee acts in self-interest rather than complying to the business agreement, which is typical for a gain frame. Also Vosselman and Van der Meer-Kooistra (2009) state that being in a gain frame has a negative effect on trust building. The trustor notices a lack of sings of the trustee having normative goals, which is a necessary condition for the building of trust as discussed before (Vosselman and Van der Meer-Kooistra, 2009; Lindenberg, 2000; Chaserant 2003).

(11)

11

The trustor expects the trustee to handle conform the relationship and to hold common beliefs (Poppo et al., 2016). Therefore, I hypothesize that:

H2a: Opportunistic behaviour is negatively related to the building of thick trust

Relational signaling includes all voluntary actions executed by parties to show commitment to the relationship (Vosselman and Van der Meer-Kooistra 2009; Connelly, Certo, Ireland, and Reutzel, 2010; Wittek, 1999). This can be done through e.g. sharing information such as product quality, costs and performance (Vosselman and Van der Meer-Kooistra, 2009) and knowledge sharing (Das and Teng, 1998). If the trustee values the IOR, they will signal to the other parties they prefer a stable and well-performing relationship. According to framing theory, the trustee uses a normative frame to guide its actions, because this frame includes behaving appropriate, trustworthy and seeking long-term individual goals (Lindenberg, 2000). Therefore, parties acting in favour of the relationship will rationally show these relational signals.

In case a party acts in self-interest, it will have less commitment to the relationship and therefore rationally give less signals. The possible costs of not complying to the contractual agreement are offset by the rewards of acting in self-interest. Since relational signals stem from voluntary actions outside the contract, receiving these signals from an opportunistic party will be even less likely, because of the following. Opportunistic parties generally act from a gain frame, where the actions are based on rational choice. If a party is in trouble, the other party in a gain frame does not have the general obligation to help, but rather makes the best out of the situation with the lowest associated costs for themselves (Lindenberg, 2000). To conclude, opportunistic behaviour is expected to lower voluntary local actions resulting in less relational signaling:

H2b: Opportunistic behaviour is negatively related to relational signaling

2.5 Shared goals as moderator and the direct effect on trust building

(12)

12

H3a: Having shared goals in an IOR enhances the building of thick trust

As discussed before, relational signals show the trustee acts in a normative frame and is willing to behave appropriate, be trustworthy and not to seek short-term individual goals. Opportunistic behaviour leads to less relational signals as proposed in hypothesis 2b, because parties pursuit self-interests and will rationally not signal commitment to the relationship. A party generally acts from a gain frame and does not feel the urge to help/act voluntarily. In case the goals of the parties in the IOR are aligned, the goals in the gain frame will become rationally more aligned with the overall goals of the relationship. Hence, the actions resulting from a gain and normative frame get similar, because the actions resulting from both frames will be based on what enhances the IOR performance. An opportunistic party is then stimulated to involve in joined problem-solving and consequently share more knowledge and ideas voluntarily. This reasoning is defined in the literature as enlightened self-interest (Chaserant, 2003; Vosselman and Van der Meer-Kooistra, 2009). The individual gains of opportunistic parties are weakened by the benefits of a well-performing relationship and the pursuit of long-term self-interest. Having shared goals will result in one being in a normative frame or close to and enhance relational signaling even if the party is opportunistic in essence. Like Vosselman and Van der Meer-Kooistra (2009) argue, relational signaling does not remove the intentions to pursuit self-interest, but is consistent with it. So, shared goals are expected to positively moderate the relationship between opportunistic behaviour and relational signaling:

H3b: Having shared goals in an IOR positively moderates the relationship between opportunistic

behaviour and relational signaling

2.6 Relational signaling and trust

(13)

13

𝐻₄+ 𝐻1ᴖ 𝐻₂−_𝑏 𝐻3𝑏 + 𝐻₃+_𝑎 𝐻₂−_𝑎

The relationship between relational signaling and thick trust proposed by Vosselman and Van der Meer-Kooistra (2009) corresponds with relational signaling theory and the ‘framing’ theory developed by Lindenberg (2000). Vosselman and Van der Meer-Kooistra argue that signaling stems from voluntary local decisions and therefore directly builds trust by showing commitment to the relationship. Trust will be built when a party shows it acts based on a normative frame by giving relational signals that suggests positive expected behaviour. Where formal controls function as mechanisms to prevent negative expectations, relational signaling serves to increase positive expectations. Therefore, the hypothesized relationship between relational signaling and trust building is:

H4: Relational signaling has a positive effect on the building of thick trust

Fig. 1. Conceptual model of the hypothesized relationships.

Relational signaling Relational trust

(14)

14 3. METHODOLOGY

In this section, the data collection method, measurements and plan of analysis will be discussed. Also the validity and reliability analyses will be addressed. In this study, a theory testing approach is the most appropriate method to use, because more empirical evidence is needed in this literature field.

3.1 Data collection

The data in this study is collected through a survey instrument. The questionnaire is developed in English by my supervisor and is pre-tested in previous years. Translations of the survey from English to Dutch was done with care and checked by myself and other students to ensure conceptual equivalence. I conducted the survey at randomly selected Dutch (located) companies which I contacted by phone and e-mail. If possible, appointments were made to fill in the survey in order to enhance the quality of the data. Researchers were then able to provide extra information in case of unclear questions. In a period of two months, I contacted 105 companies of which 12 participated, which results in a participation rate of 11.43 percent. The remaining observations were collected by other students in the same year and theme, which brings the total sample size to 115 observations.

The participating firms cover various industries, but most are active in the service and manufacturing sector (20.9% and 20%, respectively). Besides that, all participating firms were privately owned commercial companies with an annual revenue exceeding 1 million euros. Other requirements were that the partners were privately owned commercial companies and a written contract underlying the partnership should be signed by all companies participating in the partnership. The respondents were active in various functions and departments and had full knowledge of the partnership and signed the associated contract. Most of my respondents were active as managers. No restrictions were made regarding the types of partnerships, so all kinds of partnerships such as buyer-supplier and outsourcing relations were investigated. Among them were mostly buyer-buyer-supplier relationships, which made up for 64.35% of the total observations. On average, companies were doing business together for 9,5 years.

3.2 Measurement

If possible, items of each construct were adopted from existing literature and new items were developed by my supervisor if no existing scales were available in the literature. All questionnaire items were measured on a seven-point Likert scale with anchors at 1 (“strongly disagree”) and 7 (“strongly agree”), unless stated otherwise. Table A1 provides the items of each construct.

3.2.1 formal control mechanisms. According to Yang et al. (2011) formal control is hard to

(15)

15

include precision of responsibilities, how each party is to perform and how behaviour is regulated in cooperation.

3.2.2 shared goals. This construct is measured with three items adopted from Li et al. (2010). The

items measure enthusiasm in pursuing collective goals, commitment to improve benefits for the relationship and level of joint ambition and vision.

3.2.3 opportunistic behaviour. This construct contains six items developed by my supervisor. The

items describe the behaviour of the partner firm and touch upon cooperation, tendency to alter facts, taking advantages of holes in the contract, getting a larger share of the gains in the relationship, promising to do things without actually doing them later and using unexpected events to extract concessions.

3.2.4 relational signaling. This measurement is borrowed from Kingshott (2006) which view

relational signaling as behaviour that enhances cooperation and joint outcomes. This perspective is consistent with the way this study views relational signaling. The three items included describe the extent to which the partner firm would try to cover up if they had a setback, tends to be secretive about the politics in their company and to sidestep any talk about the weaknesses in their company. These items were reverse coded in the data set due to opposite way of questioning.

3.2.5 relational trust. The measure of relational trust is adopted from Poppo et al. (2016), which

view the highest level of relational trust as one in which shared identity is at place. This view is consistent with how this study treats relational trust, since shared identity would measure the level of thick trust in the relationship. Three items measure this construct and capture the extent to which both parties can effectively act for the other because both share the same understanding of what matters. It shows their confidence that their interests will be fully protected, because both share a common identity.

3.2.6 control variables. This study includes several control variables in order to account for

sources of heterogeneity: propensity to trust, alliance duration, industry, and relationship type. First of all, propensity to trust was considered, because the respondent’s propensity to trust may have significant implications for how one experiences the relational trust at the firm level. The respondent easier adopts a so-called trust-as-choice strategy assuming a better outcome is obtained by assuming the partner to be reliable (Lu and Yan, 2015; McKnight, Cummings, and Chervany, 1998). Hence, the trust-building process is enhanced. Lu and Yan (2015) find that propensity to trust has a significant positive effect on both cognitive and affect based trust. This is in line with work done by Mayer, Davis, and Schoorman (1995), which state that propensity to trust is an important predictor of trust building. Five items developed by Huff and Kelly (2003) were used to measure this construct.

(16)

16

to predict (Selnes 1998). Kotabe, Martin, and Domoto (2003), also suggest that longer relationships positively relates to knowledge sharing, which consequently leads to trust building. This construct was measured by asking the respondent how long the companies were working together. In the initial data this construct had issues regarding skewness and kurtosis, so a log transformation was performed to overcome this problem.

The third control variable industry is included as a dummy variable and controls for various industries: agriculture, forestry, and fishing; mining; construction; transportation and public utilities; wholesale trade; retail trade; banking, finance, insurance, and real estate; services; public administration; and other, of which service and manufacturing have the largest amount in the data set.

The last variable included controls for the relationship type, because of the significant implications it has on the building of trust (Das and Teng, 1998). Respondents in the survey had to choose between buyer-supplier, joint venture, outsourcing and research and development relationships, or other. This variable is also included as a dummy variable.

3.3 Reliability and validity

Validity and reliability analyses were performed to assess the validity of the data (Van Aken, Berends, and Van der Bij, 2012). First, the Kaiser-Meyer-Olkin (KMO) and Bartlett’s sphericity test were performed to check if the data is appropriate for analysis. The KMO score of 0.77 exceeds the threshold of 0.5 indicating the sample is adequate for a factor analysis (Frohlich and Westbrook, 2001). Also Bartlett’s sphericity test was significant at a 0.00 significance level, so the null hypothesis that the dataset has no correlations between the variables is rejected. Therefore, the data is sufficient to perform a factor analysis (Field 2013). The associated determinant of the correlation matrix was below the threshold of 0.00001, indicating possible multicollinearity (Field 2013). An exploratory factor analysis was conducted including all multiple-item constructs to make sure all items measure what they are supposed to measure. The rotated component matrix resulted in seven factor solutions of which some items had high loadings on wrong or multiple factors.

In order to solve the cross loading problems, three items were deleted: SIGNBHVR4, SHRDGOALS4 and RETLTRUST1. Another factor analysis was performed using a principal component extraction with varimax rotation. Absolute values below 0.45 were suppressed. Six factors were left over and all items were loading on the right construct without cross loadings. The factor loadings after deleting the items can be found in table A1. The determinant of the correlation matrix became 0.00002689, so the data did not suffer from possible multicollinearity anymore. Besides that, the item-total correlations lies above 0.20, indicating that the items correlate with the overall score from the scale (Bagby, Parker, and Taylor, 1994). The KMO and Bartlett’s test stay fine and the components explain 67.74 percent of the total variance.

(17)

17

higher than 0.70 ensures that the construct reliability of the items is satisfactory (Nunnally and Bernstein, 1994). In the initial dataset, every subscale exceeded the generally accepted threshold of 0.7, except for relational trust, which was 0.55 including all three items. After deleting the necessary items from the data set, the Cronbach’s alpha’s of both relational signaling and shared goals slightly increased. For relational trust, the value dropped to 0.48. Although the value is far from the threshold of 0.70, I decided to continue working with this construct because of the following reasons. The construct functions as a dependent variable in this study and the items are adopted from Poppo et al. (2016), where the indicators were assumed to be equally reliable. Furthermore, Hair, Ringle, and Sarstedt (2011) state that Cronbach’s alpha does assume all indicators to be equally reliable and have equal outer loadings on the construct. Therefore, a composite reliability test (CR) would be more appropriate to use because it takes into account different outer loadings of the indicator variables. Values below 0.60 point to a lack of reliability (Hair et al., 2011). The associated values are around 0.85 except for relational trust, which has a value of 0.60, so continue using this construct would be valid (see table A1).

3.3.1 common method bias. A common issue that occurs when the independent and dependent

variables stem from the same respondent is that the data suffers from common method bias. A Harman’s single factor test is used to test whether the data suffers from this inconsistency (Podsakoff and Organ, 1986). In order to test this, a factor analysis is performed with a fixed number of factors of 1 without any rotation. In case a single factor accounts for 50 percent or more of the total variance in the variables, the data is subject to common method bias. The results show that the variance explained by one single factor is 25.53 percent, so no common method bias is present.

3.3.2 normality, homoscedasticity and outliers. At last, the data set was checked for normality,

homoscedasticity and outliers by inspecting the P-P plot, scatterplot, histogram and boxplot. When heteroscedasticity or outliers are present, findings can be distorted and the possibility of a Type I error increases (Osborne and Waters, 2002). Inferences made from the regression output can become useless when analyses will be performed with a Type I error. The examination of the P-P plot shows that the independent and dependent variables are linear related which is an indication of normal distribution. Besides that, there were no signs of heteroscedasticity, because the residuals were scattered around the line evenly (Osborne and Waters, 2002). The histogram and boxplot do suggest there are outliers in the data set, however by following the outlier-labeling rule (Hoaglin and Iglewicz, 1987) no outliers were found.

(18)

18 4. RESULTS

After the constructs were proved to be satisfactory regarding their validity and reliability, a multiple regression analysis could be performed that tests the hypotheses presented in chapter two.

In table 1, the descriptive statistics of the constructs are presented. As can be seen, the independent variable shared goals shows high significant correlations with the opportunistic behaviour and the two endogenous variables relational signaling and relational trust (r = 0.45, p < 0.01; r = 0.22, p < 0.01; r = -0.35, p < 0.01, relatively). Opportunistic behaviour was negatively related to the two endogenous variables (r = -0.27, p < 0.01; r = -0.55, p < 0.01). When independent variables are correlated too high, multicollinearity may occur, which makes interpretations from the regression analysis impossible due to inaccurate coefficient estimates and standard errors (Grewal, Cote, and Baumgartner, 2004). Therefore, the data was tested for multicollinearity after the regression analysis was performed. In case the tolerance factor is less than 0.2 and the variance inflation factor (VIF) is below 10, the data suffers from multicollinearity (Menard, 1995). In this case there were no signs of multicollinearity except for alliance duration. A log transformation was performed to overcome this problem. Consequently, the highest VIF score was 1.54 and the lowest tolerance level was 0.65. Also no signs of skewness or kurtosis were present, so the data set is normally distributed. Only alliance duration had skewness and kurtosis problems in the initial dataset, but the log transformation solved the problem.

A well-known problem in previous literature is the issue of endogeneity, where the sign of regression coefficients can be biased if not corrected for (Bascle, 2008). One way to correct for endogeneity is to use a two-stage least squares procedure for the regression analysis (Poppo et al., 2016). This procedure is adopted from Poppo et al., 2016. In the first stage of the two-stage least

Table 1. Basic descriptive statistics of constructs (Pearson correlation)

Construct 1 2 3 4 5 6 7 1 Relational trust 1.00 2 Relational signaling 0.13 1.00 3 Shared goals 0.45** 0.22** 1.00 4 Opportunistic behaviour -0.27** -0.55** -0.35** 1.00 5 Formal controls 0.07 -0.18* 0.14 0.15 1.00 6 Propensity to trust -0.03 0.01 0.09 -0.06 -0.13 1.00 7 Alliance duration*** -0.03 -0.18 -0.15 0.14 -0.06 -0.07 1.00 Mean 4.30 4.53 5.07 2.79 4.31 4.84 0.75 SD 1.37 1.25 1.06 1.22 1.44 0.82 0.49

N = 115 (missing observations: 1 for shared goals and formal controls, 2 for opportunistic behaviour and 4 for alliance duration).

Dummies for industry and relationship type are not included in the table. * p < 0.05 (two-tailed)

** p < 0.01 (two-tailed)

(19)

19

squares analysis, relational signaling (RS) was regressed against the instrumental variable opportunistic behaviour (OB), interaction term shared goals (SG) and control variables:

RS = 𝑏₀+ 𝑏₁(OB) + 𝑏2(OB × SG) + 𝑏controls(Controls) + e, to obtain RS_residual= RS − RS_predicted

The standardized residuals were saved and used in the second stage. In the second stage, relational trust (dependent variable) was regressed against formal controls (FC), opportunistic behaviour, relational signaling (where the standardized residuals obtained in the first stage were used as the indicator of relational signaling) and control variables:

Relational trust = 𝑏0+ 𝑏1(RSresidual) + 𝑏2(FC²) + 𝑏3(OB) + 𝑏4(SG) + 𝑏controls(Controls) + e, Table 2 gives an overview of the results, including a first stage and second stage model. The first stage model consists of three models. In model 1 the endogenous variable relational signaling is regressed against the control variables. Model 2 adds the independent variable opportunistic behaviour and model 3 adds the interaction term shared goals. In the second stage, the dependent variable is relational trust and includes two models. Model 4 only includes the control variables and model 5 adds the direct effects formal controls, opportunistic behaviour, shared goals and relational signaling. Both model 1 and 4, which include control variables only, have a low adjusted R-squared and insignificant F-value, indicating lack of model fit. As can be seen in table 2, model 2, 3 and 4 have highly significant F-values (F = 4.26, p < 0.01; F = 4.16, p < 0.01; F = 1.98, p < 0.01) and model 5 is also significant (F = 1.72, p < 0.05)

Hypothesis 3a proposes that opportunistic behaviour is negatively related to relational signaling. In model 2 can be seen that the associated coefficient is highly significant and has a negative sign (b = -0.55, p < 0.01), which supports hypothesis 3a.

Hypothesis 4a predicts that the variable shared goals has a moderation effect on the relationship between opportunistic behaviour and relational signaling. The variables opportunistic behaviour and shared goals were standardized before the interaction term was made, because a product term can result in collinearity (Aiken and West, 1991). Model 3 shows that the interaction effect OB x SG is insignificant and therefore does not support hypothesis 4a.

(20)

20

Table 2. Regression analysis

Relational signaling Relational trust

Model 1 Model 2 Model 3 Model 4 Model 5

Independent variables B S.E. B S.E. B S.E. B S.E. B S.E.

Control variables

Propensity to trust 0.10 0.16 0.11 0.17

Alliance duration -0.27 0.26 -0.15 0.22 -0.17 0.22 0.23 0.28 0.22 0.28

Direct effects

Opportunistic behaviour Supported H2b & H2a: -0.55*** 0.09 -0.53*** 0.09 -0.21** 0.12 -0.24** 0.12

Shared goals Supported H3a: 0.07 0.11 0.57*** 0.14 0.53*** 0.14

Formal controls 0.11 0.11

Formal controls² Not supported H1: 0.04 0.13

Relational signaling Not supported H4: 0.02 0.14

Interaction effect OB x SG Not supported H3b: 0.09 0.08 -0.09 0.10 Model fit R-squared 0.16 0.44 0.44 0.30 0.30 Adjusted R-squared 0.03 0.33 0.33 0.15 0.12 R-squared change 0.27 0.01 0.22 Highest VIF 1.46 1.46 1.46 1.48 1.54 F-value 1.20 4.12*** 3.97*** 1.98*** 1.72** Change in F-value 21.94 1.28 5.29

N = 115 (missing observations: 1 for shared goals and formal controls, 2 for opportunistic behaviour and 4 for alliance duration) Dummies for industry and relationship type are not included in the table.

(21)

21

Hypothesis 4b examines the positive relationship between shared goals and relational trust. The results show shared goals are strongly positively related to relational trust (b = 0.53, p < 0.01), which supports hypothesis 4b.

Hypothesis 1 predicts that the relationship between formal controls and relational trust is inverse u-shaped. First, it will enhance thick trust building, but it will eventually damage it when too much formal controls are implemented. In order to test this, the variable formal controls was standardized and squared (Barnett and Salomon, 2006). The results show insignificant results, so no support is found for hypothesis 1. In hypothesis 2 the relationship between relational signaling and thick trust was examined. Also here, no significant results were found, so hypothesis 2 is not supported.

The dummy variables transport and public utilities and retail trade were significant in model 2 and 3, which is not presented in table 2 (b = -0.62, p < 0.10; b = -0.49, p < 0.10). In model 4 and 5, the dummy variable manufacturing was also significant (b = -0.66, p < 0.05).

5. DISCUSSION AND CONCLUSION

In this section, the test results will be related to existing literature. As two of the main hypotheses comprising the two sub questions were found to be insignificant, no conclusive answer can be given to the main research question. Plausible reasons will be presented as to why some of the results were insignificant. The results, however, confirm the importance of opportunistic behaviour and shared goals on thick trust. Also the negative influence of opportunistic behaviour on relational signaling is statistically proven. Furthermore, the theoretical and managerial implications of this study will be discussed. At the end of this section, the limitations of this study will be presented which steers the suggestions for future research.

5.1 Discussion

This study investigates the relevance of formal control mechanisms and behaviour of partners in achieving high levels of trust in an IOR. Two sub questions were constructed which help to give an answer to the main research question. The first sub question comprises the effect of formal control on thick trust building and the second question investigates the effect of relational signaling on thick trust and includes the role of opportunistic behaviour and the decisive function of shared goals. Six hypotheses were developed in order to give an answer to the research question and de data used comes from a questionnaire.

(22)

22

assumed that it would harm thick trust. No support was found for this hypothesis, so no conclusive inferences can be made on the test results and the first sub question stays unanswered. Hence, the inconclusive results are not in line with the current literature on the control-trust debate. Both the complementary and substitutive view are not supported. A plausible reason for the insignificant results could be attributed to the two different functions formal control mechanisms: safeguarding and coordination. The safeguarding function mitigates the risks of opportunistic behaviour. In more developed relationships, this function appears to be substitutive, because implementing mechanisms to safeguard the relationship may signal distrust. The coordination function helps establishing and communicating tasks, which is necessary throughout the entire relationship. Therefore, increasing formal controls to coordinate, may not signal distrust to the other partner. The question remains, whether partners can distinguish the two functions of control when they are implemented. Another explanation for the insignificant results could therefore be that the effect of formal controls on thick trust may be negligible. From a theoretical perspective, this study does not confirm that formal controls are substitutive or complementary to thick trust. This suggests that their association is even more complex. More theory building is necessary on the relation between formal controls and trust before empirical testing can be done. From a managerial view, this study was purposed to help managers with the challenge to come to an optimal level of control to mitigate the risks while keeping the trust level high. However, the actual effect of formal controls on thick trust remains unanswered.

Second, the effect of opportunistic behaviour on both relational signaling and thick trust is investigated. The results show that opportunistic behaviour is negatively related to thick trust, which is consistent with previous studies. If a partner acts opportunistic, thick trust building will be damaged since it conflicts with any positive expectations the trustor had. The effect of opportunistic behaviour on relational signaling is also negatively related. This finding is consistent with the framing theory. Parties that act from a normative frame to guide their actions prefer a stable and well-performing relationship. Parties with opportunistic behaviour do not act from a normative frame and rationally give off less voluntary signals, which shows commitment.

Third, the moderator effect of shared goals on the relationship between opportunistic behaviour and relational signaling is tested. In accordance with existing literature on enlightened self-interest, an opportunistic party is expected to show more commitment to the relationship when goals in an IOR are aligned. The rational choice to act in favour of a well performing relationship includes voluntary local decisions, resulting in relational signals. However, the results do not confirm this reasoning, because the moderator effect is insignificant. This outcome is surprising and suggests more theory building is necessary. Furthermore, the direct effect of shared goals on thick trust is proven to be significant and therefore an important antecedent of thick trust building.

(23)

23

is not significant and therefore does not confirm the proposed relationship by Vosselman and Van der Meer-Kooistra (2009). A possible reason for this result is that the two are interdependent and in fact reciprocal, meaning that it is built by each other.

Altogether, the results provide no conclusive answer to the second sub question. Likewise, the answer to the main research question is limited. Opportunistic behaviour is found to be an important antecedent to relational signals given, but aligning goals does not appear to be a decisive factor in the extent of relational signals given. Also more theory building is needed on the effect relational signals may have on thick trust building. No theoretical implications can be made based on the insignificant relation between relational signaling and trust. Only the importance of opportunistic behaviour and shared goals on thick trust is confirmed (Seppänen et al., 2007). From a managerial view, this suggests that it is in the interest of a manager to align the goals in an IOR if a long-term relationship is preferred and to keep track of incompetent behaviour.

5.2 Limitations and future research

Next, a couple of limitations will be discussed which might be the reason that insignificant results were found in this study. Based on these limitations and the findings in this study, suggestions for future research will be provided.

First of all, this study uses cross-sectional data. The observations are collected from different companies at the same point in time. Besides that, only one respondent per company was surveyed and only one party per IOR filled in the survey. The disadvantage of cross-sectional data is that it may suffer from endogeneity. Although the two-stage least square method used in this study can correct for possible endogeneity, it cannot take into account learning processes and timing patterns that influence control mechanisms (Stouthuysen, Slabbinck, and Roodhooft, 2017). Future research should make use of longitudinal data in order to account for these learning processes and timing patterns. Also, surveying multiple employees that take part in the IOR or including more than one partner might improve the validity of the data (Stouthuysen et al., 2017). The importance of a longitudinal study is confirmed with the views of Smets et al. (2013). They find that alternating the amount of formal controls used in a relationship has a negative influence on trust. Investigating how formal controls affect thick trust over time gives better insights into the control-trust nexus.

Second, the sample size (115 observations) in this study was small, which could affect significant results. Redoing this research with a bigger sample size might result in more reliable results. Besides that, the data set used in this study only includes Dutch located companies in various industries. Future research might consider gathering data globally in order to test for cultural differences (Fryxell et al., 2002). This might improve the generalizability of results.

(24)

24

formal controls to coordinate tasks and need less to safeguard. Similarly, outsourcing relationships might need different control structures including more formal controls to safeguard. Increasing formal controls can have different implications in both relationships (Mellewigt et al., 2007, Das and Teng, 1998). Therefore, more research is needed to understand how formal controls affect thick trust in different types of relationships.

Fourth, this study does not distinct the safeguarding from the coordination function of control. Based on the available data it was impossible to distinguish between these two functions of control. As existing literature notices, both functions might have different effects on the building of thick trust (Vélez et al., 2008). Hence, more research should be done at the effect of these functions on the building of thick trust.

(25)

25 REFERENCES

Aiken, L. S., West, S. G., & Reno, R. R. (1991). Multiple regression: Testing and interpreting interactions

Bagby, R. M., Parker, D. A., & Taylor, G. J. (1994). The twenty-item toronto alexithymia scale—I. item selection and cross-validation of the factor structure. Journal of Psychosomatic

Research, 38(1), 23-32.

Barnett, M. L., & Salomon, R. M. (2006). Beyond dichotomy: The curvilinear relationship between social responsibility and financial performance. Strategic Management Journal, 27(11),

1101-1122. 10.1002/smj.557 Retrieved from https://doi.org/10.1002/smj.557

Bascle, G. (2008). Controlling for endogeneity with instrumental variables in strategic management research. Strategic Organization, 6(3), 285-327.

Chaserant, C. (2003). Cooperation, contracts and social networks: From a bounded to a procedural rationality approach. Journal of Management and Governance, 7(2), 163-186.

Connelly, B. L., Certo, S. T., Ireland, R. D., & Reutzel, C. R. (2010). Signaling theory: A review and assessment. Journal of Management, 37(1), 39-67.

Costa, A. C., & Bijlsma-Frankema, K. (2007). Trust and control interrelations: New perspectives on the Trust—Control nexus. Group & Organization Management, 32(4), 392-406.

Cummings, L. L., & Bromiley, P. (1996). The organizational trust inventory (OTI): Development and validation. In R. M. Kramer, & T. R. Tyler (Eds.), Trust in organizations: Frontiers of theory and research (pp. 302-330) Thousands Oaks, CA: Sage Publications.

Das, T. K., & Teng, B. (2001). Trust, control, and risk in strategic alliances: An integrated framework. Organization Studies, 22(2), 251-283.

Das, T. K., & Teng, B. (1998). Between trust and control: Developing confidence in partner cooperation in alliances. Academy of Management Review, 23(3)

Dekker, H. C. (2004). Control of inter-organizational relationships: Evidence on appropriation concerns and coordination requirements. Accounting, Organizations and Society, 29(1), 27-49. Ebers, M. (2002). How different types of trust and attitudes towards risk affect the efficient design of

inter-organizational governance structures.

Faems, D., Janssens, M., Madhok, A., & Looy, B. V. (2008). Toward an integrative perspective on alliance governance: Connecting contract design, trust dynamics, and contract

application. Amj, 51(6), 1053-1078. 10.5465/amj.2008.35732527 Retrieved

from https://doi.org/10.5465/amj.2008.35732527

Field, A. (2013). Discovering statistics using IBM SPSS statistics (4th ed.) Sage Publications Ltd. Frohlich, M. T., & Westbrook, R. (2001). Arcs of integration: An international study of supply chain

strategies//doi.org/10.1016/S0272-6963(00)00055-3 Retrieved

(26)

26

Hair, J. F., Ringle, C. M., & Sarstedt, M. (2011). PLS-SEM: Indeed a silver bullet, journal of marketing theory and practice. Journal of Marketing Theory and Practice, 19(2), 139-152. Herzog, V. L. (2001). Trust building on corporate collaborative project

teams. Project Management Journal, 23(1), 28-37.

Hoaglin, D. C., & Iglewicz, B. (1987). Fine-tuning some resistant rules for outlier labeling. Journal of the American Statistical Association, 82(400), 1147-1149. 10.1080/01621459.1987.10478551

Retrieved from https://www.tandfonline.com/doi/abs/10.1080/01621459.1987.10478551

Houston, M. B. (2000). Buyer–Supplier contracts versus joint ventures: Determinants and consequences of transaction structure.37(1), 1-15.

Huff, L., & Kelly, L. (2003). Levels of organizational trust in individualist versus collectivist societies: A seven-nation study. Organization Science, 14, 81-90.

Kingshott, R. P. J. (2006). The impact of psychological contracts upon trust and commitment within supplier–buyer relationships: A social exchange view//doi.org/10.1016/j.indmarman.2005.06.006

Retrieved from http://www.sciencedirect.com/science/article/pii/S0019850105000921

Kotabe, M., Martin, X., & Domoto, H. (2003). Gaining from vertical partnerships: Knowledge transfer, relationship duration and supplier performance improvement in the U.S. and japanese automotive industries. Strategic Management Journal, 24(4), 293-316.

Poppo, L., Zhou, K. Z., & Li, J. J. (2016). When can you trust “trust”? calculative trust, relational trust, and supplier performance.Strategic Management Journal, 37(4), 724-741.

10.1002/smj.2374 Retrieved from https://doi.org/10.1002/smj.2374

Leifer, R., & Mills, P. K. (1996). An information processing approach for deciding upon control strategies and reducing control loss in emerging organizations. Journal of Management, 22(1), 113-137.

Lewicki, R. J., & Bunker, B. B. (Eds.). (1996). Developing and maintaining trust in work relationships. CA: Sage: Thousand Oaks.

Li, J. J., Poppo, L., & Zhou, K. Z. (2010). Relational mechanisms, formal contracts, and local knowledge acquisition by international subsidiaries. Strategic Management Journal, 31(4),

349-370. 10.1002/smj.813 Retrieved from https://doi.org/10.1002/smj.813

Lindenberg, S. (2000). It takes both trust and lack of mistrust: The workings of cooperation and relational signaling in contractual relationships. Journal of Management and Governance, 4(1-2), 11-33.

Lu, S. K., & Yan, H. (2016). Contractual control, the propensity to trust, active trust development: Construction industry. Jnl of Bus & Indus Marketing, 31(4), 459-471.

10.1108/JBIM-10-2014-0200 Retrieved from https://doi.org/10.1108/JBIM-10-2014-0200

Lui, S. S., & Ngo, H. (2004). The role of trust and contractual safeguards on cooperation in non-equity alliances//doi.org/10.1016/j.jm.2004.02.002 Retrieved

(27)

27

Madhok, A., & Weibel, A. (2007). Trust and formal contracts in interorganizational relationships — substitutes and complements.Managerial and Decision Economics, 28(8), 833-847.

Malhotra, D., & Murnighan, J. K. (2002). The effects of contracts on interpersonal trust. Administrative Science Quarterly, 47(3), 534-559.

Masaaki, K., Xavier, M., & Hiroshi, D. (2003). Gaining from vertical partnerships: Knowledge transfer, relationship duration, and supplier performance improvement in the U.S. and japanese automotive industries. Strategic Management Journal, 24(4), 293-316. 10.1002/smj.297

Retrieved from https://doi.org/10.1002/smj.297

Mayer, R. C., Davis, J. H., & Schoorman, F. D. (1995). An integrative model of organizational trust. Amr, 20(3), 709-734. 10.5465/amr.1995.9508080335 Retrieved

from https://doi.org/10.5465/amr.1995.9508080335

McKnight, D. H., Cummings, L. L., & Chervany, N. L. (1998). Initial trust formation in new organizational relationships. Amr, 23(3), 473-490. 10.5465/amr.1998.926622 Retrieved

from https://doi.org/10.5465/amr.1998.926622

Menard, S. (1995). Applied logistic regression analysis Thousand Oaks, CA: Sage.

Nooteboom, B. (2002). Trust: Forms, foundations, functions, failures and figures Edward Elgar Publishing.

Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric theory (3rd ed.) New York: McGraw-Hill. Osborne, J. W., & Waters, E. (2002). Multiple regression assumptions. ERIC digest.

Ouchi, W. G. (1979). A conceptual framework for the design of organizational control mechanisms. Management Science, 25(9), 833-848.

Podsakoff, P. M., & Organ, D. W. (1986). Self-reports in organizational research: Problems and prospects. Journal of Management, 12(4), 531-544.

Poppo, L., & Zenger, T. (2002). Do formal contracts and relational governance function as substitutes or complements? Strategic Management Journal, 23(8), 707-725.

Rousseau, D. M., Sitkin, S. B., Burt, R. S., & Camerer, C. (1998). Not so different after all: A cross-discipline view of trust. Academy of Management Review, 23(3), 393-404.

Sako, M. (1992). Price, quality and trust: Inter-firm relations in britain and japan Cambridge University Press.

Selnes, F. (1998). Antecedents and consequences of trust and satisfaction in buyer‐seller

relationships. European Journal of Marketing, 32(3), 305-322. 10.1108/03090569810204580

Retrieved from https://doi.org/10.1108/03090569810204580

Seppänen, R., Blomqvist, K., & Sundqvist, S. (2007). Measuring inter-organizational trust—a critical review of the empirical research in 1990–2003//doi.org/10.1016/j.indmarman.2005.09.003

(28)

28

Smets, L. P. M., van Oorschot, K. E., & Langerak, F. (2013). Don't trust trust: A dynamic approach to controlling supplier involvement in new product development. Journal of Product Innovation Management, 30(6), 1145-1158.

Smith, K. G., Carroll, S. J., & Ashford, S. J. (207). Intra- and interorganizational cooperation: Toward a research agenda. Academy of Management Journal, 38(1)

Srinivasan, B., & Isaac, F. (1993). Asset specificity, firm heterogeneity and capital structure. Strategic Management Journal, 14(1), 3-16. 10.1002/smj.4250140103 Retrieved

from https://doi.org/10.1002/smj.4250140103

Stouthuysen, K., Slabbinck, H., & Roodhooft, F. (2017). Formal controls and alliance performance: The effects of alliance motivation and informal controls//doi.org/10.1016/j.mar.2017.03.002

Retrieved from http://www.sciencedirect.com/science/article/pii/S1044500517300112

Thomas, M., Anoop, M., & Antoinette, W. (2007). Trust and formal contracts in interorganizational relationships — substitutes and complements. Managerial and Decision Economics, 28(8),

833-847. 10.1002/mde.1321 Retrieved from https://doi.org/10.1002/mde.1321

Tomkins, C. (2001). Interdependencies, trust and information in relationships, alliances and networks. Accounting, Organizations and Society, 26(2), 161-191.

Tsai, W., & Ghoshal, S. (2017). Social capital and value creation: The role of intrafirm networks. Academy of Management Journal, 41(4), 464-476.

Van Aken, J., Berends, J. H., & Van der Bij, H. (2012). Student projects. in problem solving in organizations: A methodological handbook for business and management students. Cambridge University Press.

Van Veen-Dirks, P., Minnaar, R., Vosselman, E., & Hassan, M K Z U. (2010). Control, relational signalling and trust building in an inter-firm

relationship: Field study reflections

Vélez, M. L., Sánchez, J. M., & Álvarez-Dardet, C. (2008). Management control systems as inter-organizational trust builders in evolving relationships: Evidence from a longitudinal case study. Accounting, Organizations and Society, 33(7-8), 968-994.

Vlaar, P. W. L., van den Bosch, F A J, & Volberda, H. W. (2007). On the evolution of trust, distrust, and formal coordination and control in interorganizational relationships: Toward an integrative framework. Group & Organization Management, 32(4), 407-428.

Vosselman, E., & van der Meer-Kooistra, J. (2009). Accounting for control and trust building in interfirm transactional relationships. Accounting, Organizations and Society, 34(2), 267-283. Williamson, O. E. (1985). The economic institutions of capitalism. New York, NY: The Free Press. Williamson, O. E. (1996). The mechanisms of governance Oxford University Press.

Wittek, R. P. M. (1999). Interdependence and informal control in organizations.

(29)

29

Yang, Z., Zhou, C., & Jiang, L. (2011). When do formal control and trust matter? A context-based analysis of the effects on marketing channel relationships in china. Industrial Marketing Management, 40(1), 86-96.

(30)

30 APPENDIX

Table A1. Factor analysis

Construct and item wording Factor loading

Relational trust (Cronbach’s alpha: 0.48, AVE: 0.43, CR: 0.60)

2. Both parties can effectively act for the other because both share the same understanding of what matters.

0.73

3. Both parties are confident that their interests will be fully protected, because both share a common identity.

0.57

Relational signaling (Cronbach’s alpha: 0.78, AVE: 0.55, CR: 0.78)

1. In our relationship, the partner firm would try to cover up if they had a setback.

0.66

2. In our relationship, the partner firm tends to be secretive about the politics in their company

0.78

3. In our relationship, the partner firm tends to sidestep any talk about the weaknesses in their company

0.78

Shared goals (Cronbach’s alpha: 0.76, AVE: 0.59, CR: 0.81)

1. Both parties in this relationship are enthusiastic about pursuing the collective goals.

0.82

2. Both parties (our firm & the partner firm) are committed to improvements that may benefit the relationship as a whole, and not only the individual parties.

0.75

3. The parties share the same ambition and vision. 0.73

Formal control mechanisms (Cronbach’s alpha: 0.76, AVE: 0.64, CR: 0.84) 1. The contract precisely defines the role/responsibilities of the partner and our

firm

0.86

2. The contract precisely states how each party is to perform 0.75 3. Generally, the contract is a primary mechanism to regulate the behaviour of

the partner

0.79

Opportunistic behaviour (Cronbach’s alpha: 0.89, AVE: 0.60, CR: 0.90)

1. The partner firm sometimes altered facts to get what it wanted. 0.84 2. The partner firm was less and less cooperative as the project/relationship

progressed.

0.74

3. The partner firm would do anything within its means to get a larger share of the gains from our relationship.

0.81

4. This partner firm sometimes promises to do things without actually doing them later.

0.81

5. This partner firm will try to take advantage of “holes” in our contract to further their own interests.

0.62

6. This partner firm sometimes uses unexpected events to extract concessions from our firm.

0.82

Propensity to trust (Cronbach’s alpha: 0.72, AVE: 0.48, CR: 0.82)

1. When I order something I’ve never seen through the mail or telephone, I am confident that the product will arrive as promised.

0.64

2. I believe that people usually keep their promises. 0.64 3. Most companies genuinely care about their customers. 0.49

4. Most salespeople are honest. 0.82

5. Most people can be trusted. 0.83

Building thick trust in inter-organizational relationships: the role of relational signaling and formal controls