Control Mechanisms in Inter-organizational Relationships:
Complements or Substitutes?
An Empirical Analysis on the Interaction Between Control Mechanisms and the Influence of Market Uncertainty when Explaining Partner Performance in Inter-organizational Relationships
Master Thesis by Jasper Jimmink S3032639 20th June, 2017 University of Groningen MSc. Business Administration: Organizational and Management Control
Supervisor: A. R. Abbasi Co-assessor: dr. M.P. van der Steen
2 Abstract
The main determinants of inter-organizational relationships (IORs) performance are still an unsolved question. Although control mechanisms are seen as an important aspect, it is yet unclear how to optimally structure and govern the control mechanisms in an IOR. Conducted survey data from business professionals is used to test how control mechanisms explain partner performance, whether they act as complements or substitutes and how market uncertainty influences the control mechanisms used by partners in an IOR. The results show that outcome and social control
mechanisms positively influence partner performance. Furthermore, the results show a positive interaction between outcome and social control mechanisms, thus supporting the complements perspective. Additionally, when market uncertainty is examined, a negative influence of market uncertainty is shown. A new insight gained from this study is the negative influence of behavioral control mechanisms on explaining partner performance. The transaction cost theory, relational exchange theory, relational contract theory and social exchange theory are enriched by testing their predictability and assumptions on the structure of control mechanisms. The findings of this study contribute to finding the main determinants of explaining performance.
Acknowledgement: I thank my supervisor Abdul Rehman Abbasi for his support during the process of my master thesis and providing the survey which is used for data collection. Data was collected with 6 fellow students, which I want to thank, because without them it would be
3 Table of contents
1. Introduction ... 4
2. Literature review ... 6
2.1 Control ... 6
2.1.1 Formal and social control mechanisms ... 6
2.2 Theoretical background ... 7
2.2.1 Critiques on TCE ... 7
2.2.2 Relational theory and opportunistic behavior ... 9
2.2.3 Social Exchange Theory ... 11
2.2.4 Complements versus substitutes perspective ... 12
2.4 Hypotheses development ... 13
3. Research Method ... 16
3.1 Sample and data collection ... 16
3.2 Variables and measurement ... 17
3.3 Data cleaning, validity and reliability ... 17
4. Results ... 19
4.1 Assumptions before regression ... 19
4.2 Sample characteristics ... 20
4.3 Regression analysis ... 20
5. Discussion ... 23
6. Conclusion ... 25
6.1 Managerial implications ... 26
6.2 Limitations and future research ... 27
References ... 29
4 1. Introduction
The forming of alliances increased substantially across all economic sectors since the late 1980s (Mamavi, Meier & Zerbib, 2015). Since then, IORs have become a research topic of substantial importance in different social science disciplines (Dekker, 2004). However, according to Mamavi, Meier and Zerbib (2015) is the search for the main elements of performance still an important concern for both managers and researchers of strategic alliances and are the
determining factors of how to improve performance still an unsolved question. According to Dekker (2004), the governance and structure of control mechanisms used in an IOR is of significant importance and are often critical to its success. Therefore, a good bit of research has been done on control mechanism and its influence on performance. The literature agrees upon the distinction of control mechanisms created by Smith, Carrol and Ashford (1995) into formal control mechanisms, which can be subdivided into outcome and behavioral control mechanisms, and social control mechanisms (Dekker, 2004). However, literature is unclear about how control mechanisms should be structured in an IOR, and a distinction exists between followers of the complements perspective and substitutes perspective (Rhee, Kim & Lee, 2014). Followers of the complements perspective argue that an interaction between formal and social control mechanisms should be used by partners in an IOR, and followers of the substitute perspective argue that partners should use only one control method (Rhee et al., 2014). The first focus of this study is contributing to clarification of the indicated literature gap by examining the relation between control mechanisms and partner performance.
There are a few theoretical streams associated with predicting the control mechanisms used in an IOR. Transaction cost economics theory (TCE) provides economic rationale to predict the control mechanisms used, although commonly used, it has received critiques because it is unable to adequately predict the control mechanisms used and does not take relational and social contexts into account (Dekker, 2004; Carson, Madhok & Wu, 2006; Li, Xie, Teo & Peng, 2010). Relational exchange theory (RET), relational contract theory (RCT) and social exchange theory (SET) have been receiving more attention lately, and provide rationale to put more emphasis on relational norms and social control mechanisms. Using the different theoretical streams increases the predictability and knowledge of the governance structure chosen in an IOR, however
5 This research investigates the interaction effect of formal and social control mechanisms on partner performance, which is somewhat more emphasized in the literature (Rhee et al., 2014). In a similar manner is the influence of market uncertainty examined. Market uncertainty is about the unpredictable changes in external environments and increases the possibility of opportunistic behavior to occur, which can influence the control mechanisms used by partners in an IOR (Poppo, Zhou & Li, 2015; Carson et al., 2006). Therefore, the second focus of this study is investigating the effect of market uncertainty on control mechanisms and partner performance. The first and second focus will be investigated upon by answering the following research question:
“Does the interaction of formal and social control mechanisms lead to higher firm performance in an inter-organizational relationship, and what is the effect of market uncertainty?“
From a practical and managerial perspective, this study contributes to the understanding of control mechanisms used in an IOR, how they influence partner performance, and how control mechanisms and partner performance is influenced by market uncertainty. As indicated by Mamavi et al. (2015), the determining factors of how partners can improve partner performance are still an unsolved question. Therefore, practitioners could use knowledge gained by this study in their advantage by applying it to their own IOR. This research deepens the understanding of the complements versus substitutes perspective, and the rationale behind it based on several theories. TCE, RET, RCT and SET are enriched because these theories provide rationale to support the complements or substitutes perspective.
6 2. Literature review
The previous section introduced the complements versus supplements perspective which is elaborated upon in this section. First the purpose of control in an IOR is outlined, setting a basis for explaining formal and social control mechanisms. Thereafter are the theories discussed associated with control mechanisms and are hypotheses built upon the discussed theories and literature.
2.1 Control
The main focus of this study is to contribute to closing the gap of the complements versus substitutes perspectives of control mechanisms used in an IOR. To do this, first the purpose of control is defined, as various explanations exist in the literature. The explanation of control used in this study is the one of Dekker (2004). According to Dekker is the purpose of control “to create the conditions that motivate partners in an IOR to achieve desirable or predetermined outcomes” (p. 29, 30). Control mechanisms may play an important role in how achievements and outcomes are realized and how partners are motivated. In an IOR different control mechanisms can exist. A distinction is made between two types of control mechanisms: formal control mechanisms and social control mechanisms (Smith, Carrol & Ashford, 1995). This distinction will be further looked into in the next paragraph.
2.1.1 Formal and social control mechanisms
There is a clear distinction in the literature between formal and social control mechanisms (Smith et al., 1995). Formal control mechanisms focus on contractual obligations and formal organizational mechanisms to enhance collaboration between partners in the IOR, and can be subdivided into outcome and behavior control mechanisms (Dekker, 2004). Examples of formal control mechanisms are goal setting, incentive systems and performance monitoring. The contractual aspect of formal control mechanisms is an important part, because formal control mechanisms primarily rely on contracts (Li et al., 2010). However, formal organizational mechanisms are also an important part of formal controls, for example monitoring of the IOR (Anderson & Oliver, 1987).
One subdivision of formal control mechanisms is outcome control mechanisms (Dekker, 2004). Outcome control mechanisms focus on precise and consistent valuation of the
performance of partners in an IOR (Das & Teng, 2001). Furthermore, outcome control
7 (Dekker, 2004). Outcome control mechanisms enriches the IOR with goal setting, making
expectations more clear and increase goal alignment (Das & Teng, 1998). The other subdivision of formal control mechanisms is behavior control mechanisms (Dekker, 2004). Behavior control mechanisms focusses on making sure that the underlying process is correct (Das & Teng, 2001), and how the partners of an IOR act according to earlier made agreements (Dekker, 2004).
Furthermore, behavior control mechanisms put emphasis on the procedures and activities that are necessary to reach a goal (Lew & Sinkovics, 2013). According to Das and Teng (1998), behavior control mechanisms can be important in an IOR because they ensure that the partners perform desirable behavior when goals are not aligned and performance ambiguity occurs.
The other distinction indicated by Smith et al. (1995) are the social control mechanisms. Social control mechanisms are about creating and building trust between partners, “to enhance desirable behavior, and often takes form of joint problem solving, participatory decision making, thorough information exchange, and fulfillment of promises” (Li et al., 2010, p.334). Comparing social control mechanisms with formal control mechanisms, the social control mechanisms do not rely on contractual obligations (Li et al., 2010) and monitoring of actions and objectives
(Anderson & Oliver, 1987), but more on social control mechanisms like problem solving and participatory decision making based on trust (Li et al., 2010).
2.2 Theoretical background
This paragraph discusses the theoretical background, which forms the foundation for developing hypotheses. First, TCE is discussed and are critiques on this theory explained. Secondly, RET and the RCT are discussed and how they influence control mechanisms used in an IOR. Furthermore, the relation between RET and RCT is discussed with opportunistic behavior. Thirdly, SET is discussed and how SET predicts the governance structure of an IOR. 2.2.1 Critiques on TCE
TCE is widely used in many studies to explain the governance and mechanisms of an IOR (Cao & Lumineau, 2015). In explaining the mechanisms used in an IOR with TCE, there is a lot of focus on the contractual aspects of formal control mechanisms to restrain opportunistic behavior. According to TCE, “a well-established contractual governance can be an effective mechanism to control exchange hazards by specifying each party’s roles in both stable and
8 everything can be specified in a contract. This aligns with the definition of formal control
mechanisms by Dekker (2004) to also put emphasis on other formal control mechanisms, rather than contracts alone. TCE has received critiques because it does not take social and relational aspects into account and is therefore unable to fully predict the governance and mechanisms of an IOR (Dekker, 2004; Li et al., 2010). The critiques are explained next.
The first critiques have to do with the way TCE explains the governance structure by transaction cost characteristics (Chiles & McMackin, 1996). Firstly, TCE is not able to explain the differences that can exist in an IOR’s forms and goals (Dekker, 2004). Secondly, TCE’s static nature has led to carelessness of the organizational mechanisms used in IOR governance (Dekker, 2004). Thirdly, TCE ignores most of the social mechanisms of governance, while this seems to be of importance in an IOR and therefore should be taken into account (Dekker, 2004). Due to these critiques, the distinction of control mechanisms in formal and social control by Smith et al. (1995) is used because this distinction puts more emphasis on taking a completer view on control mechanisms.
The second critique has to do with how TCE explains governance of IOR’s, because TCE does not take relational aspects of change into account (Carson et al., 2006). Therefore, according to Carson, Madhok and Wu (2006), does the RCT put more emphasis on giving enough attention to individual transactions within larger systems of economic, and the RCT puts more emphasis on social interaction. The RCT as well as the RET will be elaborated upon in the next paragraph.
The third critique is on how TCE may be too narrow as a theory to explain the control mechanisms. Li, Xie, Teo & Peng (2010) argue that varying findings of previous studies lead to the chance that TCE is not a wide enough approach to sufficiently account for how formal and social control mechanisms are chosen in an IOR. Lin, Peng, Yang and Sun (2009) argue that social and institutional perspectives can lead to more cohesive results investigating how control mechanisms are chosen in an IOR, therefore Li et al. include the social and institutional
perspective in their study.
9 2.2.2 Relational theory and opportunistic behavior
An important aspect of formal control mechanisms is the contract that is established between partners. However, as the definition of formal control by Dekker (2004) indicated, formal control mechanisms do not only exist out of contracts, but also out of formal
organizational mechanisms. Therefore are RET and RCT discussed next, as they enhance TCE with a more social and relational approach.
RET is created by Macneil (1978, 1980), who made a distinction between discrete and relational exchange. Discrete exchange is based upon the assumption that each transaction is independent of past and future relations between contracting partners, and that the transactions are just about transferring the ownership of a product or service (Goldberg, 1976). Furthermore, discrete exchanges are between autonomous individual parties which have their own interest. The other distinction is relational exchange, which is the contrast of discrete exchange. In relational exchange are the past and future relations taken into account, even as the social context (Heide, 1994). Furthermore, the execution of responsibilities are seen as logic because of the mutual interest between the parties (Heide, 1994). The way of looking at relational exchanges captures the view of a two-sided system, in which individuals utility functions are included in the overall comprehensive utility of the system (Bonoma, 1976). Furthermore, Bonoma (1976) states that the goals of individuals in such a two-sided system are achieved through joint accomplishments, and opportunistic behavior is reduced because of the long-term benefits of the system.
RET complies with previous critiques on TCE because of its two-sided aspect, including more social structure of exchanges (Granovetter, 1985), and RET does not rely on hierarchical mechanisms to manage relations between partners (Heide, 1994). RET provides a way to analyze the expected behaviors in an IOR (Cao & Lumineau, 2015). RET takes relational norms into account, like flexibility, information exchange and solidarity, as they are important to remind partners in an IOR that their relationship is not only about a contract and that both partners are expected to act in an appropriate way because of shared relational norms (Cao & Lumineau, 2015). Furthermore, these relational norms are considered to be effective to manage an IOR (Cannon, Achrol & Gundlach, 2000).
10 Lee (2014), a contract can be divided into different dimensional components which can be
measured. These dimensions are: “role specification, planning, adjustment process, monitoring system, incentive system and explicit enforcement” (Heide, 1994 p. 75), which align with the previous explained distinction between outcome and behavior control mechanisms. The RCT suggests that each conducted contract lies on a line between discrete and relational (which is discussed before), and depending on the IOR the relationship fits within the spectrum of discrete and relational, and based on where it fits the spectrum, the content and composition of a contract will differ (Macneil, 1978). Next to that, Heide (1994) discusses that the established relationship between partners in an IOR can lead to different distinctions in the form of governance. So, based on RCT, contracts of formal control mechanisms can differ in IOR’s due to the kind of
relationship that exists.
Both RET and RCT take more social aspects into account trying to restrain opportunism, which is discussed next. According to Carson et al. (2006) a lot of attention has been given to the phenomenon of opportunism as well as how to structure the governance in an IOR to safeguard against opportunism. In an IOR there is room for opportunistic behavior, because the contracts in an IOR are based upon the activities covered in the IOR, but also if, and how many times, the partners worked together (Gulati, 1995). Contracts can differ due to previous activities, existence and prior ties (Gulati, 1995). This raises the opportunity of looser contracts as partners build more confidence in each other (Gulati, 1995). Cao and Lumineau (2015) argue that putting emphasis on relational norms in an IOR leads to less occurrence of opportunistic behavior. Based on the RET is opportunistic behavior governed because the assumption that both partners are expected to act in an appropriate way and have shared relational norms (Cao & Lumineau, 2015). Based on the RCT opportunistic behavior can also be restrained within a contract, by inserting the
11 deal with unforeseeable problems (Li et al., 2010).
Another aspect that influences the occurrence opportunistic behavior in an IOR is market uncertainty (Rhee et al., 2014). Market uncertainty is about unpredictable changes in external environments (Poppo et al., 2015) and requires adaptive abilities in the actions of managers and firms (Carson et al., 2006). Adaptive abilities are required because market uncertainty leads to changes in an environment over time, which leads to uncertainties about future conditions (Carson et al., 2006). Market uncertainty increases the likelihood of opportunistic behavior to occur, because the need to reconsider agreements to avoid inadequate adaptation is raised (Carson et al., 2006).
2.2.3 Social Exchange Theory
The Social Exchange Theory (SET) provides explanations for decisions taken in an IOR regarding governance. The assumption of SET is that trust is crucial for establishing a stable social relation (Cao & Lumineau, 2015). As indicated by Lee et al. (2010), TCE may be too narrow to explain governance mechanisms in an IOR. SET is not only an economic theory but also sociological and therefore has a wider reach than TCE and may be better able to explain governance mechanisms (Cao & Lumineau, 2015). According to Cao and Lumineau (2015) “SET focuses on the role of social exchange which is defined as voluntary actions of exchange parties that are motivated by the returns they are expected to obtain” (p. 17). Rhee et al. (2014) use SET to explain the social aspect of governance in an IOR. Rhee et al. argue that social control
mechanisms contribute to partners acting in a desirable way, which leads to joint problem solving and participatory decision making by exchanging information and fulfillment of promises, when firms develop a trust relationship. This view has been criticized by Jeffries and Reed (2000) because of the extremely positive point of view this has. However, Rhee et al. provide more insight into SET, elaborating upon that a successful ongoing IOR increases the amount of trust partners have in each other. Partners of an IOR want to prove that they are trustworthy, therefore the partners in an IOR will put emphasis on creating a good relationship with the partner, and showing their commitment to the partnership (Lambe, Wittmann & Spekman, 2001).
Furthermore, a successful ongoing IOR enhances commitment, and leads to better performance due to recurring interactions improving social norms (Rhee et al., 2014). Next to this, the social control mechanism can lead to replacing formal control mechanisms, due to the recurrent
12 the monitoring and incentive system (Rhee et al., 2014). The routinization of actions is also acknowledged by Jap and Ganesan (2000) who argue that the longer a relationship exists, the higher the chance that routinizing of actions has been established. Li et al. (2010) argue, just like Rhee et al., that the social aspect of an IOR can influence the use of control mechanisms in an IOR. This influence is based on the assumption that social relation restrain partner’s behavior, and that partners do not focus on short-term gains but want to build trust and create long-term economic value (Uzzi, 1997). To conclude, SET provides insight into the social interaction between partners in an IOR and provides explanations for how to govern an IOR (Cao & Lumineau, 2015).
2.2.4 Complements versus substitutes perspective
The previous paragraphs provide arguments to support the complements or substitutes perspective based on theories and literature. Complying with previous critiques, RET and RCT provide bases to support the complements perspective, by taking more relational norms into account. On the other hand, SET provides the basis for arguments who follow the substitutes perspective. First are arguments provided for the substitutes perspective, thereafter for the complements perspective. Secondly is the influence of market uncertainty on the complements versus substitutes perspective addressed.
Followers of the substitute perspective firstly argue that the use of both social and formal control mechanisms together can hardly be economical once sufficient trust is built to only use social control mechanisms, because of the costs of creating a comprehensive contract (Dyer & Singh, 1998; Gulati, 1995). Secondly, according to Rhee et al. (2014), the recurrence of actions between partners can lead to ‘routinizing’ of formal control mechanisms like the transaction role, monitoring system and incentive system. The routinizing of this could lead to replacing the formal control mechanisms by social control mechanisms. According to Jap and Ganesan (2000) is the likelihood of substituting formal control mechanisms with social control mechanisms dependent on how long the relationship exists. Thirdly, based on SET, contractual obligations can lead to distrust, because in a longer existing IOR the partners get a distrusted feeling when a very extensive contract is used (Jap & Ganesan, 2000). Fourthly, Rhee et al. have a similar view as the first and second argument: the followers of the substitutes perspective conclude that the
13 Followers of the complements perspectives firstly argue that relational norms like
flexibility, information exchange and solidarity should be taken into account (Cao & Lumineau, 2015). Secondly, based on the RCT, a comprehensive contract provides guidelines for improved cooperation, and that trust has the benefit that when something unexpected occurs in contract details, trust between the partners in an IOR can help solving this unexpected problem (Luo, 2002; Poppo & Zenger, 2002). Thirdly, when only social control mechanisms are used a lawsuit is more complicated, because in a contract the rights and duties of each partner are worked out, and a contract can be used as a legal framework which shows how the partners in the IOR want to cooperate (Luo, 2002; Williamson, 2008). Fourthly, according to Uzzi (1997), social control mechanisms can cover for limitations of formal controls, because not everything can be specified in a contract, and when unanticipated problems occur, the social control mechanisms based on trust can make sure that the cooperation continues. Lastly, Macneil (1978) notes that contracts have adaptive limits and Li et al. (2010) point the flexibility aspect of social control mechanisms out, and that social control mechanisms may foster bilateralism, therefore the interactive use of formal and social control may be positive on performance.
Market uncertainty has the ability to influence control mechanisms used by the increased occurrence of opportunistic behavior (Rhee et al., 2014). Besides that, market uncertainty has several other possibilities to influence the control mechanisms used. Firstly, in an uncertain market should the ability to change the transaction conditions be safeguarded by the partners (Rhee et al., 2014). This implies more flexibility in the relationship with the partner, and therefore to put more emphasis on social control mechanisms (Li et al., 2010). Secondly, Dyer and Chu (2003) argue that when a sufficient trust relationship is built, this does however not decrease ex-ante contract costs in more uncertain markets and circumstances: explicit contracts provide clarity and guidance in uncertain circumstances. Thirdly, in a more uncertain market the partners may communicate more extensively, leading to an increased importance of social control mechanisms because of the routinizing of actions and decision making (Rhee et al., 2014). 2.4 Hypotheses development
Interaction effect of formal and social control mechanisms
14 the nonsocial aspects of formal control mechanisms and comply with critiques on TCE. Based on SET arguments can be given to also use social control mechanisms, or substitute the formal control mechanisms.
To recall, formal control mechanisms focus on contractual obligations and formal organizational mechanisms to enhance collaboration between partners in the IOR, and can be subdivided into outcome and behavior control mechanisms (Dekker, 2004). Social control mechanisms are about creating a trust relationship which leads to joint problem solving, comprehensive and extensive information exchange, fulfillment of promises and participatory decision making (Li et al., 2010).
RET and RCT provide rationale to support the complements perspective, thus using both formal and social control mechanisms. Taking more social control mechanisms into account has the benefit of its safeguarding ability against opportunistic behavior (Carson et al., 2006). Opportunistic behavior can occur because the control mechanisms are influenced by existence and prior ties of the IOR (Gulati, 1995). Taking more social control mechanisms into account, based on the RCT and RET, safeguard for opportunistic behavior by enhancing flexibility, information exchange and solidarity (Cao & Lumineau, 2015), and the embeddedness of
individual transactions within the whole system, which also protects against opportunism because the exchanges and transactions between the partners are embedded in the IOR (Carson et al., 2006). Furthermore, using both formal and social control mechanisms has several advantages. Firstly, using both control mechanisms enhances cooperation because a contract is used Luo, 2002; Poppo & Zenger, 2002). Secondly, when only social control mechanisms would be used, a lawsuit would be more complicated (Luo, 2002; Williamson, 2008).Thirdly, not everything can be specified in a contract; in uncertain situations the partners can rely on their trust relationship based on social control mechanisms (Uzzi, 1997). Fourthly, social control mechanisms are more flexible, and cover adaptive limits of formal control mechanisms (Macneil, 1978; Li et al. (2010). Furthermore, according to Rhee et al. (2014) is the complements perspective somewhat more emphasized in recent studies, which leads to the following hypotheses:
H1a: The interaction between outcome and social control mechanisms is positive on partner performance.
15 Moderating effect of market uncertainty on the interaction between outcome, behavior and social control mechanisms
The structure of control mechanisms used in an IOR can be influenced by how uncertain the market is (Rhee et al., 2014). As discussed earlier is market uncertainty about unpredictable changes in external environments (Poppo et al., 2015) and requires adaptive abilities in the actions of managers and firms (Carson et al., 2006). Furthermore, adaptive abilities are required because market uncertainty leads to changes in an environment over time, which leads to uncertainties about future conditions (Carson et al., 2006).
Market uncertainty has the possibility to influence control mechanisms used. Two reasons on how market uncertainty influences control mechanisms used are discussed previously. Firstly, due to market uncertainty is the likelihood of the occurrence of opportunistic behavior increased (Carson et al., 2006). Opportunism can occur due to market uncertainty raising the need to reconsider agreements to avoid inadequate adaptation (Carson et al., 2006). Secondly, partners in an IOR might want to structure their control mechanisms differently under more market
uncertainty. Three arguments for this can be given: Firstly, in an uncertain market should the ability to change the transaction conditions be safeguarded by the partners (Rhee et al., 2014). This implies more flexibility in the relationship with the partner, which is also supported by the RCT, and therefore to put more emphasis on social control mechanisms (Li et al., 2010). Secondly, Dyer and Chu (2003) argue that when a sufficient trust relationship is built, this does however not decrease ex-ante contract costs in more uncertain markets and circumstances: explicit contracts provide clarity and guidance in uncertain circumstances. Thirdly, in a more uncertain market the partners may communicate more extensively, leading to an increased importance of social control mechanisms because of the routinizing of actions and decision making (Rhee et al., 2014). This leads to the following hypotheses:
H2a: Market uncertainty positively affects the interaction between outcome and social control mechanisms on partner performance.
16 The conceptual model in figure 1 shows developed hypotheses and their relations.
Figure 1 Conceptual model
3. Research Method
The research method of this study is theory testing. This approach is appropriate since the inconclusive findings in previous literature on the complements versus substitute perspective (Aken, Berends & Van der Bij, 2012). This chapter discusses the methodology of this study. At first, the procedure for sampling and data collection is explained. Secondly, the variables and measurements are described. The last paragraph entails the process of cleaning the data, and are the prerequisites before analyzing the data explained and performed, which enhance the validity and reliability.
3.1 Sample and data collection
17 made to conduct the survey in a face-to-face setting. Conducting the surveys face-to-face
enhanced the understandability of the survey for the respondents because respondents could ask questions when something was unclear. Furthermore, the face-to-face setting made sure that the manager or director himself filled in the survey. In the end, my own response rate was 26%. I contacted 75 business professionals, mainly via telephone and LinkedIn, which led to conducting 20 face-to-face surveys. The total obtained sample which I started with before cleaning the data contained a total of 131 surveys.
3.2 Variables and measurement
The variables of this study are measured by 7-point Likert scales ranging from
“completely disagree” to “completely degree”, “very less” to very much” and “totally not” to “to a large extend” provided by the supervisor. The control variables of this study are firm size compared to the partner firm, the amount of employees and interdependence. The control variables used are adopted from previous studies. First, firm size compared to partner is adopted from Li et al. (2010), who also study the interaction effect. Differences in firm size has the possibility to influence the control mechanisms used, since it can indicate how reliant a partner is on the other. For example, an IOR can have a huge importance for a small firm, whilst it can be of lesser importance for a larger firm. Therefore, firm size might influence the control
mechanisms used and partner performance. The second control variable is interdependence, which is also adopted from Li et al. (2010). Interdependence is used, because a high degree of interdependence may foster using both formal and social control mechanisms (Li et al., 2010). This is a similar control variable as firm size compared to the other partner in the sense that when a partner is too dependent on the other, it can influence the control mechanisms used. The third control variable used is amount of employees, adopted from Rhee et al. (2014). This control variable measures how many employees work at the firm. This control variable is used because larger firms may have more resources, and therefore may be less reliant on the partner.
3.3 Data cleaning, validity and reliability
The process of data cleaning, and increasing validity and reliability contains six steps which are discussed next. Assessing validity and reliability is important, because it increases the strength and interpretability of the results calculated in chapter 4.
18 The first step after obtaining the whole sample is cleaning the data. From the original sample of 131, 20 surveys are deleted because they were to a very large extend not completed. The second step is to investigate common method bias (CMB) by examining Herman’s single factor test (Podsakoff, MacKenzie, Lee & Podsakoff, 2013). The Herman’s single factor test shows an explained variance of 26.1%, which is below the threshold of 50%, indicating that CMB is not present. This means that no single factor is responsible for the majority of the variance in the constructs (Mat Roni, 2014).
The third step is performing the prerequisites of principal component analysis (PCA), which are the Bartlett’s test of sphericity and the Kaiser-Meyer-Olkin (KMO) test. Performing a PCA is chosen because of two benefits, which are described by Abdi and Williams (2010): 1) using PCA leads to only using the most important variables in the data set and 2) using PCA reduces the size of the data set while only keeping the most important variables. The Bartlett’s test of sphericity test the overall significance of the correlations, and shows a p-value of 0.00, which is below 0.05, indicating that it is appropriate to perform a factor analysis on this data (Williams, Onsman & Brown, 2010). The KMO test measures if the variables show a strong relationship between each other. This test shows a result of 0.724 which is sufficient because it is higher than 0.50, which is the score it should exceed (Williams et al., 2010). Furthermore, for the PCA, the covariance matrix is used, because the variables are all measured by using a 7-point Likert scale. Besides the covariance matrix, direct Oblimin and rotated solution are used. Additionally, small coefficients below 0.40 are suppressed, a fixed number of 6 factors are chosen and cases are excluded listwise. The results of the PCA are presented in appendix A. In the end are the following four items excluded: one item of interdependence (interdependence1), one item of outcome control mechanisms (outcomecontrol4) and two items of behavior control mechanisms (behaviorcontrol 1 and 3).
The fourth step is ensuring the reliability of the constructs indicated by the PCA, by calculating Cronbach’s Alpha. The calculated Cronbach’s Alpha range from 0.760 to 0.822 all exceeding 0.70 which is acceptable (Nunnally, 1978).
The fifth step is creating the logarithm of the variable ‘how many employees work at your organization’, which is one of the control variables. The normal variable is very positively
19 The final step is the creation four models in which the data is tested. Model 1 contains the control variables. Model 2 contains the control variables plus the study variables: outcome control mechanisms, behavior control mechanisms, social control mechanisms and market uncertainty. In this model the main effects of the independent variables on the dependent variables are tested. Mean centered variables are created to increase the estimation of the interaction effect between the moderator and the independent variables (Grace-Martin, 2015). Model 3 contains the control variables, the main study variables and the interaction variables and tests the interaction effects (H1) of social control mechanisms with outcome and behavior control mechanisms. Model 4 contains the control variables plus the mean centered study variables and the moderator variable market uncertainty times the interaction variables. This model test hypothesis 2.
4. Results
The previous chapters provide grounds to start testing data. However, before testing the data, the assumptions necessary for the multiple regression analysis are examined in the first paragraph. The second paragraph explains the sample characteristics. Lastly, the third paragraph presents the results of the multiple regression analysis.
4.1 Assumptions before regression
Before the indicated models are tested, several assumptions are considered: normal distribution, linear relationship between the independent and dependent variables,
20 each other. The VIF’s of the interdependent variables are calculated and show a result of smaller than 10, indicating that the variables are not multi-correlated (O’Brien, 2007).
4.2 Sample characteristics
The sample characteristics are calculated based on 110 observations. 20 observations were deleted due to large incompleteness, and one observation is deleted because it was an outlier. In table 1 are the descriptive statistics provided. Furthermore are the means, standard deviations and Pearson correlations presented.
Table 1 Descriptive statistics and correlations
Firm size compared to the other partner has a mean of 3.04, which indicates that most of the partners are comparable in size. Looking at the Pearson correlations, an interesting result can be seen in the correlation between partner performance, outcome control mechanisms (r = .228, p < .05) and social control mechanisms (r = .212, p < .05) as they are important variables in this study. Furthermore, decent correlation exists between outcome control mechanisms and behavior control mechanisms (r =.476, p < .01). This is quite expected, since they have a lot in common and both are a subdivision of formal control mechanisms. Also behavior control mechanisms and social control mechanisms show decent correlation (r =.0.441, p < .01). This is also not a
surprising result since behavior control mechanisms are about assessing behaviors and social control mechanisms about social interactions, which both have to do with human interactions. 4.3 Regression analysis
21 total sample is reduced to 94 observations instead of the earlier 111. The outlier detected earlier is also not included. The results are presented in table 2.
Table 2 Results multiple regression
Model 1, consisting of the control variables, explains partner performance for 0.8% (R² = .008). The F-test is not significant for model 1 (F = .254, ns). However, model 2 (F = 2.329, p < .05), model 3 (F = 2.347, p < .05) and model 4 (F = 2.832, p < .01) are significant, which increase the reliability of the results for the significant models. Furthermore, the control variables are insignificant in every model, indicating that they do not influence the control mechanisms and partner performance. When market uncertainty is taken into account the control mechanisms remain insignificant.
Model 2, which test to which extend the study variables explain partner performance, explains partner performance for 15.9% (R² = .159). The change in R² is 15.1% (∆R² = .151). Outcome control mechanisms (β = .198, p < .01), behavior control mechanisms (β = -.147, p < .1) and social control mechanisms (β = .225, p < .05) are significant and market uncertainty is
22 between the formal control mechanisms and social control mechanisms. This hypothesis is tested in model 3. The results show that the variables explain partner performance for 20.1% (R² = .201). The change in R² is 4.2% (∆R² = .042), which indicates that the interaction effect increases the explaining ability of the control mechanisms on partner performance. The same three main variables as in model 2 are significant. There is an increase in the significance of behavior control mechanisms (p < .05) and an increase of the significance of social control mechanisms (p < .01).Testing hypotheses H1a shows a significant interaction between outcome and social control mechanisms (β = .107, p < .1), thus supporting H1a. H1b is not supported due to insignificance (β = -.033, ns). Thus, outcome and social control mechanisms are complements in explaining
partner performance under normal market conditions. In model 4 the interaction effect is also tested including the moderating effect of market uncertainty. The normal interaction effect is different from the results in model 3. In model 4 is the interaction between outcome and social control mechanisms insignificant (β = -.019, ns) and the interaction between behavior and social control mechanisms significant (β = -.119, p <.10). However, under these circumstances is H1b still not supported, since an positive interaction was expected.
23 5. Discussion
This study investigates the interaction effect of formal control mechanisms, and the effect of market uncertainty on partner performance. It hypothesizes that outcome and behavioral control mechanisms positively influence partner performance, and that when the IORs perform in an uncertain market, the partners in the IOR put more emphasis on the interaction effect to deal with the uncertainty. The empirical results indicate that outcome and social control mechanisms complement each other in explaining partner performance. For the other subdivision of formal control mechanisms, namely the behavior control mechanisms, the complement perspective is not supported. Furthermore does market uncertainty negatively influence the interaction between outcome and social control mechanisms. The results contribute to literature and theory which will be discussed below.
First, the results indicate that the interaction effect between outcome and social control mechanisms explains partner performance, showing support for hypothesis H1a. The direct effects of outcome and social control mechanisms are also positive and significant, implying that outcome and social control mechanisms can explain and improve partner performance. The support for the interaction effect was expected, based on the arguments of RCT, RET and SET which provide rationale for this result. Next to that, previous work on the complements
perspective versus the substitutes perspective shows a moderate amount of support for the
complements perspective in recent studies (Rhee et al., 2014), which is also the case in my study. Although, the study of Rhee et al. (2014) itself does not find support for the complements
24 needed or already covered for.
Second, this study shows that under the circumstance of market uncertainty, the
interaction between behavioral and social control mechanisms has a negative influence on partner performance. The non-supported hypothesis just discussed may provide rationale for the negative influence on partner performance. Under normal market conditions hypothesis H1b does not hold, and when market uncertainty is taken into account, the interaction between behavior and social control mechanisms negatively influence partner performance. The direct effect of behavior control mechanisms is also tested and has a negative influence on partner performance every time. Thus, the interaction of behavior and social control mechanisms does not increase partner performance, and should be carefully looked into by partners in the IOR.
Third, this study shows that market uncertainty negatively influences the interaction between outcome and social control mechanisms, although expected otherwise. Literature provided rationale to expect otherwise, because under market uncertainty managers are required to be able to adapt (Carson et al., 2006) and because market uncertainty increases the possibility of opportunistic behavior to occur (Carson et al., 2006). Therefore it was expected that under market uncertainty firms would put more emphasis on using both formal and social control mechanisms, to safeguard against opportunistic behavior and to have more flexibility in their control mechanisms.
Fourth, the control variables used in this study are insignificant and do not explain partner performance. This result does not change when measured in different models. When the
interaction effect is measured, the control variables remain insignificant. Next to that, when market uncertainty as moderating variable is taken into account, the control variables remain insignificant. This was unexpected, since the control variables positively or negatively influenced performance in the research of Li et al. (2010) and Rhee et al. (2014). Thus, my results suggest that firm size compared to the partner, the amount of employees and interdependence are not determinants of partner performance.
25 2006; Li et al., 2010). The critiques are tested in this study by investigating upon the effect of social control mechanisms. The results indicate that the direct effect social control mechanisms are very important, with the highest positive significant beta. Outcome control mechanisms are also important with a positive and significant beta. However, behavior control mechanism are negative significant, indicating that they do not increase partner performance. Thus, the results acknowledge the critiques on TCE and suggests that social control mechanisms should be taken into account. This aligns with the assumptions of the RCT, RET and SET, which argue that relational and social aspects should be taken into account. RET argues that relational norms such as flexibility, information exchange and solidarity are effective means to manage an IOR
(Cannon et al., 2000). This view is acknowledged by testing the data, leading to support for the interaction between outcome and social control mechanisms. The RCT, which has similar assumptions as the RET, also argues that social aspects should be taken into account (Carson et al., 2006), which is also acknowledged by this study. RET and RCT take more social aspects into account to restrain opportunism (Cao & Lumineau, 2015). However, when market uncertainty is taken into account, a positive influence on the interaction effect between formal and social
control mechanisms is not supported. This indicates that market uncertainty negatively influences formal and social control mechanisms, meaning that partners in an IOR do not use control
mechanisms to restrain opportunistic behavior. This contradicts the predictions based on the RET and RCT. Lastly, SET provides arguments to use social control mechanisms which can
complement or substitute formal control mechanisms. This study shows that outcome control mechanisms positively influence partner performance, therefore indicating that social control mechanisms do not replace the outcome control mechanisms. Thus, the routinizing ability of social control mechanisms, to routinize the transaction role, and the monitoring and incentive system (Rhee et al., 2014), is not supported in this study. However, the arguments to complement formal control mechanisms are supported.
6. Conclusion
26 “Does the interaction of formal and social control mechanisms lead to higher firm performance in an inter-organizational relationship, and what is the effect of market uncertainty?“
This question can now be answered after testing the data obtained from business professionals with knowledge of the relationship between the partners.
This research provides two contributions by answering the research question. First, this study contributes to narrowing down the gap between the complements versus substitutes perspective that is present within the literature (Li et al., 2010; Rhee et al., 2014). The results suggest that under normal market circumstances the interaction between outcome and social control mechanisms complement each other and improve partner performance. Furthermore, the interaction between behavior and social control mechanisms negatively influences partner performance. Lastly, when market uncertainty is taken into account, it negatively influences the interaction between outcome and social control mechanisms.
Second, answering the research question contributes to the theories used within the study. TCE is contributed to by investigating critiques that have arisen upon this theory, namely that TCE is unable to fully predict the governance structure of an IOR (Dekker, 2004; Carson et al., 2006; Li et al., 2010). The results show that although outcome control mechanism are important, they are complemented by relational and social aspects. RET and RCT are contributed to by acknowledging their arguments based on the results. The results show that the relational arguments are important for partners in an IOR, by the support of the interaction between outcome and social control mechanisms. Furthermore, SET is enhanced by looking into social control mechanisms, and the interaction effect of social and formal control mechanisms. The results show that although social control mechanisms are important, they do not substitute the formal control mechanisms. Furthermore, an interaction between outcome and social control mechanisms is supported within the study.
6.1 Managerial implications
According Mamavi et al. (2015), the determining factors of an IOR are still an unsolved question for managers and researchers. This study investigated the structure of control
mechanisms that improve partner performance, therefore enhancing the knowledge of
27 takes form of joint problem solving, participatory decision making, thorough information
exchange and fulfillment of promises (Li et al., 2010). However, based on the results outcome control mechanism are also able to increase partner performance. This suggests that an
interaction between both leads to the most optimal form, which is also supported in this study. Using both outcome and social control mechanisms leads to taking relational and social aspects into account, as well as using contracts to enhance cooperation, safeguarding against opportunism and when a lawsuit is required the contract can be used as legal framework (Luo, 2002; Poppo & Zenger, 2002; Williamson, 2008).
6.2 Limitations and future research
Besides the contributions this study has, limitations and future opportunities for research exist. Firstly is the sample size a limitation, since it is quite low. Higher sample size could improve the results of the tests. The sample size could be increased by increasing the length of the data collection process. Seven students had to collect all the data within one month. The time to conduct an interview contains one hour plus respondents had to be found. The results are hard to generalize due to the sample size and further characteristics of the sample. No clear sector is chosen were the partners should perform their business in, and different relations could exist between the analyzed IORs, for example buyer supplier or outsource relation. Future research could aim to investigate specific relations, like buyer-supplier, joint ventures and outsource relations.
Secondly, the results indicate that behavioral control mechanisms negatively influence partner performance, which was not expected based on the literature. The interaction between outcome and social control mechanisms was negative and insignificant under normal market conditions. However, when market uncertainty was taken into account, the interaction between behavioral and social control mechanisms became significant, although still negative. Based on the literature a positive relation between behavioral control mechanisms and partner performance was expected, since they ensure that partners perform desirable behavior when goals are not aligned and performance ambiguity occurs (Das & Teng, 1998). Future research could investigate behavior control mechanisms specifically and how they influence partner performance in
different types of IORs.
28 performance based on the interaction between formal and social control mechanisms or when market uncertainty is taken into account. Therefore, other control variables should be looked into. One example of a control variable that could be used is prior experience, which is positively related to performance in supplier-buyer relationships (Poppo et al., 2015). Future research could investigate prior experience in other relationships, like outsourcing or joint ventures.
Fourthly, other theories exist in the literature which are able to explain and predict the governance structure of IORs. In this study are TCE, RCT, RET and SET used. However, Li et al. (2010) used the institutional theory besides SET. Although the theories used in this study are able to explain the governance structure of IORs, Li et al. (2010) found significant results with measuring institutionalization. Li et al. argue that institutionalization positively influences formal and social control mechanisms used because institutions constrict behavior, and “institutionalized beliefs and behavioral norms may help secure interfirm trust in ‘‘perpetuity’’ (p. 336). Thus, future research could take the institutionalized perspective into account by adopting the
29 References
Abdi, H., & Williams, L. J. (2010). Principal component analysis. Wiley interdisciplinary
reviews: computational statistics, 2(4), 433-459.
Anderson, E., & Oliver, R. L. (1987). Perspectives on behavior-based versus outcome-based salesforce control systems. The Journal of Marketing, 76-88.
Berry, W. D., & Feldman, S. (1985). Multiple regression in practice: Quantitative applications in the social sciences.
Bonoma, T. V. (1976). Conflict, cooperation and trust in three power systems. Behavioral
Science, 21(6), 499-514.
Cannon, J. P., Achrol, R. S., & Gundlach, G. T. (2000). Contracts, norms, and plural form governance. Journal of the Academy of Marketing Science, 28(2), 180-194.
Cao, Z., & Lumineau, F. (2015). Revisiting the interplay between contractual and relational governance: A qualitative and meta-analytic investigation. Journal of Operations
Management, 33, 15-42.
Carson, S.J., Madhok, A., Wu, T., 2006. Uncertainty, opportunism, and governance: the effects of volatility and ambiguity on formal and relational contracting. Academy of Management
Journal, 49 (5), 1058–1077.
Chiles, T. H., & McMackin, J. F. (1996). Integrating variable risk preferences, trust, and transaction cost economics. Academy of Management Review, 21(1), 73-99. Cohen, J. (81). Cohen. P.(1983). Applied multiple regression/correlation analysis for the
behavioral sciences. Hillsdale, NJ: Erlbaum.
Das, T. K., & Teng, B. S. (1998). Between trust and control: Developing confidence in partner cooperation in alliances. Academy of management review, 23(3), 491-512.
30 Dekker, H. C. (2004). Control of inter-organizational relationships: Evidence on appropriation
concerns and coordination requirements. Accounting, Organizations and Society, 29(1), 27-49.
Dyer, J. H., & Singh, H. (1998). The relational view: Cooperative strategy and sources of inter-organizational competitive advantage. Academy of Management Review, 23(4), 660-679. Dyer, J. H., & Chu, W. (2003). The role of trustworthiness in reducing transaction costs and
improving performance: Empirical evidence from the United States, Japan, and Korea.
Organization Science, 14(1), 57–68.
Goldberg, V. P. (1976). Regulation and administered contracts. The Bell journal of economics, 426-448.
Grace-Martin, K. (2015). When NOT to Center a Predictor Variable in Regression. Retrieved from http://www.theanalysisfactor.com/when-not-to-center-a-predictor-variable-in-regression/
Granovetter, M. (1985). Economic action and social structure: The problem of embeddedness. American journal of sociology, 91(3), 481-510.
Gulati, R. (1995). Does familiarity breed trust? The implications of repeated ties for contractual choice in alliances. Academy of management journal, 38(1), 85-112.
Heide, J. B. (1994). Inter-organizational governance in marketing channels. The Journal of
Marketing, 71-85.
Jap, S. D., & Ganesan, S. (2000). Control mechanisms and the relationship life cycle:
Implications for safeguarding specific investments and developing commitment. Journal
of Marketing Research, 37(2), 227–245.
Jeffries, F. L., & Reed, R. (2000). Trust and adaptation in relational contracting. Academy of
31 Lambe, C. J., Wittmann, C. M., & Spekman, R. E. (2001). Social exchange theory and research
on business-to-business relational exchange.Journal of Business-to-Business Marketing,8(3), 1-36.
Lew, Y. K., & Sinkovics, R. R. (2013). Crossing borders and industry sectors: Behavioral
governance in strategic alliances and product innovation for competitive advantage. Long
Range Planning, 46(1), 13-38.
Li, Y., Xie, E., Teo, H., & Peng, M. W. (2010). Formal control and social control in domestic and international buyer–supplier relationships. Journal of Operations Management, 28(4), 333-344.
Lin, Z. J., Peng, M. W., Yang, H., & Sun, S. L. (2009). How do networks and learning drive M&As? an institutional comparison between china and the united states. Strategic
Management Journal, 30(10), 1113-1132.
Luo, Y. (2002). Contract, cooperation, and performance in international joint ventures. Strategic
Management Journal, 23(10), 903-919.
Macneil, I. R. (1978). Contracts: Adjustment of long-term economic relations under classical, neoclassical and relational contract law. Northwestern University Law Review, 72, 854– 905.
Macneil, I. R. (1980). The new social contract. New Haven: Yale University Press.
Mamavi, O., Meier, O., & Zerbib, R. (2015). Alliance management capability: the roles of alliance control and strength of ties. Management Decision, 53(10), 2250-2267. Mat Roni, S. (2014). Introduction to SPSS. Edith Cowan University, SOAR Centre, Australia. Nunnally, J. (1978). Psychometric theory. New York: McGraw Hill.
O’brien, R. M. (2007). A caution regarding rules of thumb for variance inflation factors. Quality
&Quantity, 41(5), 673-690.
32 Palmatier, R. W., Dant, R. P., & Grewal, D. (2007). A comparative longitudinal analysis of
theoretical perspectives of inter-organizational relationship performance. Journal of
marketing, 71(4), 172-194.
Pedhazur, E. J. (1997). Multiple regression in behavioral research: Explanation and prediction. Podsakoff, P. M., MacKenzie, S. B., Lee, J.-Y., & Podsakoff, N. P. (2003). Common method
biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88(5), 879-903
Poppo, L., & Zenger, T. (2002). Do formal contracts and relational governance function as substitutes or complements? Strategic Management Journal, 23(8), 707-725. Poppo, L., Zhou, K. Z., & Li, J. J. (2015). When can you trust “trust”? Calculative trust,
relational trust, and supplier performance. Strategic Management Journal.
Rhee, J. H., Kim, J. W., & Lee, J. (2014). Interaction effects of formal and social controls on business-to-business performance. Journal of Business Research, 67(10), 2123-2131. Rossi, P. H., Wright, J. D., & Anderson, A. B. (Eds.). (2013). Handbook of survey research.
Academic Press.
Smith, K. G., Carroll, S. J., & Ashford, S. J. (1995). Intra-and inter-organizational cooperation: Toward a research agenda. Academy of Management Journal, 38(1), 7-23.
Uzzi, B., 1997. Social structure and competition in interfirm networks: the paradox of embeddedness. Administrative Science Quarterly, 42 (1), 37–69.
Van Aken, J., Berends, H., & Van der Bij, H. (2012). Student Projects. In Problem solving in organizations: A methodological handbook for business and management students.
Cambridge University Press.
Williamson, O. E. (2008). Outsourcing: Transaction cost economics and supply chain management. Journal of supply chain management, 44(2), 5-16.
33 Appendix A:
