The influence of power, supplier independence and trust on a buyer’s preferred customer status

(1)

The influence of power, supplier independence and trust on a

buyer’s preferred customer status

Thesis MSc SCM

MSc Supply Chain Management

University of Groningen, Faculty of Economics & Business

Duisenberg Building, Nettelbosje 2

9747 AE Groningen, The Netherlands

Supervisor: dr. J. Veldman

Second supervisor: prof. dr. J. Wijngaard

(2)

2

ABSTRACT

This research represents an explorative study on power, trust and supplier independence on their relation on supplier allocation of resources. Earlier research has shown that power and trust can be seen as key strategies for buyers to influence supplier resource allocation. This study is a combination of a partly replication of an earlier study and an extension on this by adding some extra variables to the conceptual framework. In this study, survey data is analyzed to draw conclusions. The sample size exists of 41 cases which are composed from surveys from buyers and surveys from their suppliers. Structural equation modeling is used to analyze the data. The analyses yield some outcomes that are interesting. Findings show that coercive power influences physical innovation allocation positively and thereby shows that coercive power not always has a negative influence on resource allocation. Reward power has a negative effect on the allocation of resources. Also, expert power and the presence of goodwill- and competence trust are good mechanisms to acquire better resources. The use of competence trust only has a negative effect on allocation of innovation resources and therefore should be avoided by buyers when innovation resources need to be acquired.

Keywords: preferred customer, coercive power, reward power, expert power, supplier

(3)

3

1. INTRODUCTION

Many industrial business-to-business situations are characterized by a low supplier to buyer ratio. This results in many buyers that source from the same few suppliers (i.e. shared supplier). These buying firms are in competition with each other to get the best resources from these suppliers. Ellram, Tate & Feitzinger (2013) stated that supply chain resources are essential to achieve a competitive advantage. When a firm can acquire these resources, it can stay competitive and successful by creating a better market position. To get the better resources from shared suppliers, a buying firm should try to influence these suppliers to favor them above competing buying firms and therefore acquire a preferred status above other buyers.

“Successful relationships with suppliers can create benefits that extend beyond the actual product or service exchange (Lindgreen & Wynstra, 2005)” (Pulles et al., 2014, p. 18). One of these benefits can be a preferred customer status over other supplier's customers, which will have an influence when suppliers decide on whom to source or who to source first. A preferred customer is a buyer to whom the supplier allocates better resources in comparison with less preferred buyers (Pulles et al., 2016). According to Pulles et al. (2016) this is because of the supplier that favors the buyer's behaviors, practices, business values, or some combination of them. Ellis, Henke and Kull (2012) found that the behavior of a buyer affects their customer status. Their study demonstrates that it is important for purchasing managers to adopt supply chain management practices that strengthen their firm’s image as a high value customer, and

by this try to maximize the value of the relationship with that supplier. By strengthening this image, a buyer can become a preferred customer and in that way create a preferential resource allocation position.

‘‘In the supply chain management (SCM) literature, power and trust are seen as the two main strategies used by buying firms to influence suppliers (Terpend & Ashenbaum, 2012). Both power and trust can be seen as mechanisms to control the dynamics of social business relationships (Bachmann, 2001)’’ (Pulles et al., 2014, p. 18). Narasimhan et al. (2009) builds their

definition of power upon

(5)

5

physical resource allocation for buyers accounting for a large share in the supplier’s turnover.

Research has highlighted the need for further exploration of variables that influence preferential resource allocation (Pulles et al., 2016). Based on this need for further exploration the following research question is conducted: What is the effect of

power, trust and supplier independence on supplier resource allocation in a buyer-supplier relationship, and does trust have a moderating effect in this relationship?

The aim of this research from a theoretical perspective is to extend the prior research of Pulles et al. (2014). First of all, their research will be extended with the variables expert power and supplier (in)dependence. Supplier dependence is measured in their paper based on buyer’s share in supplier turnover. In this study supplier independence will be measured from the supplier’s perspective. Secondly, their study is based on data from the supplier’s perspective. In contrast, this study will collect data from both a supplier's and a buyer’s perspective. In their study, power and trust are addressed as independent variables measured from supplier’s perspective. On the contrary, this study will collect data on coercive and reward power from buyer’s perspective. This will strengthen the research field that is overrepresented by research measured from supplier perspectives. Thirdly, trust will also be analyzed as a moderator variable. This will be measured from supplier's perspective. In this way, it can be studied if trust moderates the relationship between reward power and supplier independence and the dependent variable preferential resource allocation. From a

practical point of view this study provide insights about purchasing strategy mechanisms that a buyer can use for suppliers to increase their competitive advantage.

Most of the studies that are about the use of power in supply chain relationships just measure one side of the relationship dyad, which limits the evaluation of perceptual congruence between partners (Anderson & Weitz, 1992). Only six studies between 1986 and 2005 about buyer-supplier relationships gathered data from both a buyer’s and a supplier’s perspective (Terpend et al., 2008) and researchers acknowledge that there is a lack of this kind of dyadic response. Perceptions of a partner about another partner are often different compared with reality. This leads to outcomes that need caution in drawing conclusions. By combining both data from supplier and buyers, results based on perceptions are reduced and thereby strengthen the outcomes of this study. This research is based on survey data from the manufacturing industry to answer the research question. This industry is very suitable for this research because companies operating in this industry have often a lot of suppliers. These suppliers often also serve possible competitors of the buying firm. Data is gathered by multiple buying firms, and a sample of these buyer’s suppliers. Several suppliers of this buying firm filled in a survey about the buying firm, and the buying firms representatives filled in a survey about the suppliers.

(6)

6

hypotheses will be shown. After that, the methodology section will provide an overview of the research design, sample and data collection, used measurements, data validity and data analysis. Thereafter the results will be presented followed by the discussion, conclusion, limitations and further research.

2. THEORETICAL

BACKGROUND

2.1 Preferential resource allocation from suppliers

Between firms there are exchanges of resources. Within these exchange relationships both firms allocate resources to each other from which the firms derive advantages (Pulles & Schiele, 2013). As mentioned above there is a lot of competition between buyers to get the best resources form shared suppliers. To win this competition a buyer must try to get a better customer position than other competing buyers. A better customer position can be described as a preferred customer position. Resource allocation from a supplier towards a customer is defined by Pulles et al. (2014, p.26) as ‘‘the extent to which the supplier allocates to the particular buyer better resources than to the buyer's competitors’’. “The allocation of supplier resources to relationships with buying firms is a selective process in which competing customers may be treated unequally (Mitsuhashi & Greve, 2009)” (Pulles et al., 2016, p.2). This means that suppliers differentiate between their customers and decide to treat one buyer better with prime resources than another. This differentiation fits with the preferred customer concept that is mentioned by Baxter (2012). When

a buyer has such a preferred customer position, a supplier will prefer this buyer to allocate their better resources, or allocate their resources faster to this buyer.

Resources can be divided in ‘‘the tangible or intangible assets a firm possesses or has access to’’ (Newbert, 2008, p.766). These resources can further be subdivided into financial, human, intellectual, organizational/relational, legal and physical resources (Hunt & Davis, 2008; Newbert, 2008). Resources in the case of this study are divided into 1) physical- and 2) innovation resources.

(7)

7

lose the ability to differentiate from competitors and thereby lower their competitive advantage (Gnyawali & Madhavan, 2001; Takeishi, 2001)’’. To stay competitive a buyer should try to continuously retain and improve their relationship with suppliers that are crucial for buyer’s performance to stay ahead of their competitors.

2.2 Social exchange theory (SET)

‘‘The concept of social exchange directs attention to the emergent properties in interpersonal relations and social interaction. A person for whom another has done a service is expected to express his gratitude and return a service when the occasion arises’’ (Blau, 1964, p. 4). Within an interpersonal exchange there are different resources that can be exchanged. Foa and Foa (1980) characterized these resources as love, status, information, money, goods and services. Next to interpersonal exchanges, these transactions can also occur between firms. So, “Social exchange theory (SET) argues that individuals or corporate groups interact for a reward or with the expectation of a reward from their interaction with others (Homans, 1958; Thibaut & Kelley, 1959; Emerson, 1976)’’ (Narasimhan, et al., 2009 p.2). Instead of only tangible resources by economic exchanges, social exchanges are not limited to material goods but also include intangible value (Pulles et al., 2014). Examples of these intangible values are social amenities or friendship, emotional satisfaction, spiritual values, pursuit of personal advantages, and sharing ideas (Lambe, Wittmann & Spekman, 2001). From the SET perspective, an exchange as a social behavior may result in both economic and social outcomes (Lambe et al., 2001). So, SET extends

beyond previous economic theories like the resource based view theory. Within a buyer-supplier relationship, a social exchange means that one of the parties makes a contribution, for example in the form of resources, to the other party. In return, that party expects something back from the other party at a later moment. When there is a well relationship between two firms, reciprocity will often occur whereby actions are performed back and forth. Within SET there are six basic assumptions (Narasimhan et al., 2009, p. 2). These assumptions are listed below: 1) People are rational and calculate the

best possible means to engage in an interaction and look to maximize the profit/returns.

2) Most gratification is centered in others. 3) Individuals have access to information

about social, economic, and

psychological dimensions that allows them to assess alternatives, more profitable situations relative to their present condition.

4) People are goal oriented.

5) Building social ‘‘credit’’ is preferred to social ‘‘indebtedness’’.

6) SET operates within the confines of cultural context (i.e. norms and behaviors being defined by others.

2.3 Power

(8)

8

can occur for a single exchange between a buyer and supplier as well as for repeatable exchanges between the firms in a buyer-supplier relationship. Within such a relationship, a buyer can use power mechanisms to control the dynamics of a trans-organizational relationship (Bachmann, 2001).

Some conceptualizations of power in a buyer-supplier relationship come forth from the SET perspective. From these conceptualizations the following definition of power come forth: ‘‘the ability of one member of a supply chain to influence or control the decisions and behavior of other persons, groups, or organizations (Beier & Stern, 1969; Gaski, 1984; Payan & McFarland, 2005)’’ (Narasimhan et al., 2009, p. 3). There are different sources of power that can be used in a relationship to influence another party described in literature. French & Raven (1959) distinguish between the major types of power: 1) reward power, 2) coercive power, 3) legitimate power, 4) referent power and 5) expert power. These different forms of power can be classified into different dichotomizations like mediated/non-mediated, coercive/non-coercive and economic/non-economic power (Maloni & Benton, 2000). In this research the dichotomy mediated/non-mediated will be used. According to (Maloni & Benton, 2000) mediated power includes reward, coercive and legal legitimate power, and non-mediated power includes expert, referent and traditional legitimate power. Forms of mediated power are used by the power source as influence efforts to guide the target response by being deliberately engaged or threatening (Maloni & Benton, 2000). On the other hand, non-mediated power

sources are often not exercised to manipulate the target (Maloni & Benton, 2000). By being polite or threatening by punishments a buyer can try to force his suppliers to get the responses they desire. When a power mechanism is successful used, the resulting supplier’s behavior will be optimal for the buyer (Ramsay, 1996). Using power can also have a negative effect on the relationship with suppliers. As stated by Pulles et al. (2014, p. 19): ‘‘failed attempts may not result in any changes in the supplier’s behavior or even in undesirable behavior’’. To be successful in getting the best resources from a supplier, a buyer need to know in what way forms of power have an influence on the relationship with a supplier. Therefore, the focus within this research will be on coercive, reward and expert power and their influence on supplier’s resource allocation.

2.4 Trust

Like power, trust can also been seen as an important mechanism that is used by buying firms to influence their suppliers

(Terpend & Ashenbaum, 2012). Trust is an important variable in SET for successful relationships (Lambe et al., 2001). The creation of trust is an important aspect of social exchange because social exchanges are governed to a large degree by social obligations rather than contracts (Blau, 1968, p. 454).

(9)

9

possibility is there that trust will arise. In a buyer-supplier relationship trust can be beneficial with actions creating positive outcomes. As stated in the paper of Pulles et al. (2014, p.19): ''Trust between firms creates an atmosphere where partners willingly exceed the minimal requirements of an exchange relationship, based on the belief that the partner will take actions that will results in positive outcomes for the firm and will not perform actions that result in negative outcomes (Anderson & Narus, 1990; Ireland & Webb, 2007)''. So, positive economic and social outcomes over time increase the partners’ trust of each other and commitment to maintaining the exchange relationship (Lambe et al., 2001) and can result in positive outcomes for both firms in the relationship.

Trust is a multilevel phenomenon that can exists at personal, organizational, inter

-organizational, and international levels (Das & Teng, 2001). There are different forms of trust described in literature. Das and Teng (2001) distinguish between two dimensions of trust that can exist in inter-organizational relationship. These forms are goodwill trust and competence trust. These concepts are based on the paper of Barber (1983). Goodwill trust is defined as ''the expectation that some others in our social relationship have moral obligations and responsibility to demonstrate a special concern for other's interest above their own'' (Barber, 1983, p.14). When an actor within a relationship has goodwill trust they feel that the other party will make sacrifices for them and will do everything that is possible to satisfy them. Competence trust is defined by Barber (1983, p. 14) as ''the expectation of technically competent role performance''. When competence trust is in play in a

relationship, one party feels that the other party has the necessary expertise and knowledge to do what is needed. In a buyer-supplier relationship it is important that these forms of trust exist. When there is trust between the actors in this relationship it will be beneficial for both parties. It is possible that this trust results in better supplier's resource allocation for the buyer.

2.5 Supplier dependence

(10)

10

where either the supplier or the buyer is dependent on its partner, and by symmetric dependence that both parties are dependent on each other. These forms of dependence can vary in their intensity ranging from slightly dependent relationships to heavily dependent lock-in situations (Narasimhan et al., 2009).

3. HYPOTHESES

In this section the hypotheses will be introduced based on existing literature and SET reasoning. Hypothesis 1a to 4b will link power and trust to supplier allocation of physical- and innovation resources. Herewith, the study of Pulles et al. (2014) will be replicated. Hypotheses 5a to 6b will link expert power and supplier independence to both forms of resource allocation. Hypothesis 7 to 8b will link trust as a moderator on the relations between reward power and supplier independence on resource allocation.

3.1 Coercive power

Coercive power can be used to manipulate the attainment of valences (French & Raven, 1959). Coercive power stems from the expectation that one of the actors in a relationship will be punished by the other when agreements will not be honored. So, in a buyer-supplier relationship a buyer can use this kind of power to pressure the supplier. Pulles et al. (2014) stated that the aim of buying firms to use coercive power is to pressure a supplier into complying with their requirements (Pulles et al., 2014). There is a drawback of the use of coercive power in a relationship between a buyer and supplier. There is the possibility that by the use of buyer’s coercive power a supplier decide to favor other buyers that do not use coercive power against them.

Taking a look at the results of Pulles et al. (2014), that investigated the relationship between coercive power and resource allocation, their results show a positive effect that was not significant between coercive power and supplier allocation of physical resources. But when they looked at buyer’s share in supplier’s turnover, they found that coercive power did have a significant positive effect on physical resources. This result is a different from previous study outcomes. ‘‘Many studies mainly point to the negative effects of coercive power (e.g., Brown et al., 1995; Maloni & Benton, 2000)’’ (Pulles et al., 2014, p. 30). Instead of a negative effect, Pulles et al. (2014) found that there can be a positive effect of coercive power on the allocation of physical resources. Based on this the following hypothesis is conducted:

H1a: Coercive power is positively related

(11)

11

Ashenbaum (2012) stated that it is possible that when the buyer-supplier relationship is characterized by coercion, the supplier is disinterested in sharing new innovations or in providing new ideas. Then, a supplier will allocate their innovation resources to other buyers that do not use coercive power as a relation mechanism. Pulles et al. (2014) found a negative, not significant, effect between coercive power and the supplier allocation of innovation resources. For replication of their study, the following hypothesis is conducted:

H1b: Coercive power is negatively related

to supplier allocation of innovation resources.

3.2 Reward power

Reward power is defined by French and Raven (1959) as power used with the intention to reward. They also stated that reward power depends on the ability to administer positive valences and to remove or decrease negative valences (French & Raven 1959, p.253). Compared with coercive power that can be seen as a form of punishment, in a buyer-supplier relationship, buyer’s reward power thus results in benefits for the supplier or decreases supplier’s negativity towards the buyer. The use of reward power has a positive effect to the collaborative and adaptive behavior of supplier (Nyaga et al., 2013). Then, this use of reward power can result in a buyer’s preferential resource allocation position compared to other supplier’s buyers. Reciprocity is important for reward power in a buyer-supplier relationship. When a buyer offers consecutive rewards in time, a supplier can feel committed towards this relationship. Then, ‘‘suppliers can be more inclined to allocate their physical resources to buying

firms that offer them benefits’’ (Pulles et al., 2014, p.20). This commitment can become beneficial for the buyer in forms of a preferred customer status. As stated by Pulles et al. (2014, p. 20): ‘‘if a buyer offers benefits to a supplier who shares ideas and new innovations, this supplier can be expected to be more willing to offer future innovations to this firm’’. In SET literature, it is stated that when a supplier can choose between multiple buyers to allocate their resources, they will choose for the buyer that offers the highest reward (Narasimhan et al., 2009). So high rewards given by a buying firm could increase the resource allocation form a supplier that receives this reward towards the buying firm. The results of Pulles et al. (2014) shows that reward power used by buyers has significant positive effects on both the physical- and innovation resource allocation. The hypotheses used in their study will be replicated in this study:

H2a: Reward power is positively related to

supplier allocation of physical resources.

H2b: Reward power is positively related to

supplier allocation of innovation resources.

3.3 Goodwill trust

(12)

12

resources towards this buyer. This can lead to higher competitive advantage for a buyer compared to competitors. Then this allocation of better products will lead to lower production costs for the buyer (Li et al., 2007). A trustworthy partner is a more likely partner for other parties in a relationship (Tsai & Ghoshal, 1998). In terms of innovation resources, when supplier’s goodwill trust in a buyer is high they would be more willing to share their technology and knowledge with this buyer (Roy, Sivakumar & Wilkinson, 2004). When looking at Pulles et al. (2014), goodwill trust was found to be not a significant effect on the supplier allocation of physical and innovation resources. But they found a significant effect for supplier allocation for physical and innovation resources when the buyer accounts for a large share in supplier’s turnover. For the replication of the study of Pulles et al. (2014), the following hypotheses are conducted:

H3a: Goodwill trust is positively related to

H3b: Goodwill trust is positively related to

supplier allocation of innovation resources.

3.4 Competence trust

As before, competence trust is defined as ‘‘the expectation of technically competent role performance’’ (Barber, 1983, p. 14). Hereby one party has the confidence that the other party has the necessary competence to do what is necessary. Without a certain level of competence trust, it will be more difficult for a firm to share resources to a buyer. Ellram and Edis (1996) found empirical results that indicated that trust plays an important role in achieving a successful relationship with

a partner. When firms work together as a team to come up with new products, a team member must feel confident that the other member is able to solve new problems when they emerge (Madhavan & Grover, 1998). In a buyer-supplier relationship there is need for mutual confidence in each other. A mismatch in competence trust can lower chances of innovation performance in the relationship (Roy et al., 2004). So, for both buyer and supplier it is important to create competence trust in their relationship. The trust that a supplier has in the competence of a buying firm can stimulate future reciprocity, and thereby is expected to satisfy buyer’s need by allocating resources (Pulles et al., 2014). Pulles et al. (2014) found that competence trust has a significant positive effect on the allocation of physical and innovation resources and even have a higher effect when firms account for a small share in supplier’s turnover. Based on above and for replication of the study of Pulles et al. (2014) the following hypotheses are conducted:

H4a: Competence trust is positively

related to supplier allocation of physical resources.

H4b: Competence trust is positively

related to supplier allocation of innovation resources.

3.5 Expert power

(13)

13

the use of non-mediated power increases commitment. Also, a positive relation between non-mediated power and cooperation was found by Hunt et al. (1987). This shows that expert power can be important in a relationship between two actors. So, this can also be the case in a relationship between a buyer and a supplier. The target firm, supplier, decides how it will be influenced by the firm, buyer, exerting the power (Zhao et al., 2008). Palmatier et al. (2006) mentioned that when a firm interacts with a partner firm that is competent, they will receive increased value, their relation becomes more important, and they will invest more in that relation to strengthen and maintain it. When a buyer exerts expert power, it is possible that a supplier is influenced, and that this will lead to higher influence on the supplier resulting in better resources allocated. Jonsson and Zineldin (2003) argue that non-mediated power sources tend to increase the value of relationships because they increase the level of cooperation. It is expected that a weaker firm, less expertise and reverence, will put more effort in enhancing the relationship with a stronger partner. So, when a supplier has a high perception of the buyer’s expertise it is likely that the supplier will allocate better resources to that buyer to enhance the relation. Based on above the following hypotheses are established:

H5a: Expert power is positively related to

H5b: Expert power is positively related to

supplier allocation of innovation resources.

3.6 Supplier (in)dependence

‘‘When a significant proportion of a supplier’s sales is to a buyer, its dependence towards this buyer increases (Crook & Combs, 2007; Hallén et al., 1991; Laamanen, 2005; Provan, 1993; Sambharya & Banerji, 2006)’’ (Carr et al., 2008, p. 901). The supplier will then be more dependent on the specific buyer and assume that the supplier will put effort in their relationship to maintain this relationship. When a supplier can easily shift its sales to another buyer the supplier has independence over their buyer. As stated in the paper of Touboulic, Chicksand and Walker (2014, p.581):

‘‘In buyer-supplier relationships with power imbalances, the dominant organization is likely to exercise its influence over the other party and act to maintain its power, whereas the weaker organization is more likely to comply to continue accessing resources (Kumar, Scheer & steenkamps, 1995; Gulati & Sytch, 2007; Zhu et al., 2008)’’

(14)

14

automotive industry shows that supplier dependence is also related to more integration between the supplier and buyer within the product development process. This willingness to cooperate with a buyer caused by the dependence can lead to that supplier will prefer this buyer above others to share with the supplier’s key technologies, best ideas, and innovation resources. When a supplier is not dependent on a buyer, this buyer will not be preferred above other buyers when the supplier has to decide to whom serve their best resources. Based on above the following hypotheses are conducted:

H6a: Supplier independence is negatively

related to supplier allocation of physical resources.

H6b: Supplier independence is negatively

related to supplier allocation of innovation resources.

3.7 Trust – moderating effect

Goodwill trust can moderate the relationship between buyer’s reward power and innovation resource allocation. When there is high goodwill trust, the firm perceives positive feelings of the reward power influence exercised by the powerful partner (Ke & Wei, 2008). When a supplier has high goodwill trust in a buying firm, the supplier then will think that the buying firm will keep their promises and will treat them fairly. Ke and Wei (2008) propose that goodwill trust moderates the relation between reward power and the dependent firm’s predisposition to share knowledge and know-how to the more powerful firm. Thus, when the supplier’s goodwill trust in the buyer is high, the buyer’s reward power will have a bigger effect on the supplier’s innovation resource allocation.

Based on above the following hypothesis is established:

H7: High supplier’s goodwill trust

strengthens the relation between reward power and supplier allocation of innovation resources.

When a supplier is dependent on a specific buyer, they probably give them a preferred treatment compared to other buyers. The supplier can fear that when they do not operate as the buyer expects the supplier to do, the buyer decides no longer to work together with them. Then, the supplier will lose a large part of his sales. Because of the fear, they will cooperate with the buyer to satisfy them. But when a supplier has independence towards a specific buyer, the supplier probably will not give the buyer a preferred treatment compared with other buyers. Ke and Wei (2008) propose that firm’s goodwill trust can mitigate negative feelings that are created by a state of asymmetric dependence. Then, this supplier does not prefer the buyer when they have to allocate their resources. Supplier independence has then a negative effect on supplier resource allocation. Goodwill trust can weaken this negative effect by mitigating the negative feelings towards the buyer. Then, the buyer will be a more interesting partner to cooperate with. This can result in better supplier resource allocation. The supplier will then prefer the buyer faster when they have to decide who to source. Based on above the following hypotheses are established:

H8a: High goodwill trust weakens the

(15)

15 H8b: High goodwill trust weakens the

negative effect of supplier independence on supplier allocation of innovation resources.

In Figure 1 the conceptual model of this study is shown.

3. METHODOLOGY

3.1 Research design

This study is about resource allocation measured from both buyer's and supplier's perspective. Hypotheses H1a to H4b are a replication of the study of Pulles et al. (2014) about power and trust, and their relation on preferential resource allocation. In their study the variables were measured from a supplier’s perspective. The answers given by their respondents are for some variables a perception. In this study these variables are measured from the buyer’s perspective and suggest real answers. Next to this replication, their study is extended with the variables expert power and supplier independence (H5a – H6b). Hereby, from supplier’s perspective, feelings are measured about how dependent on the buyer a supplier feel and if the supplier thinks that their buyer is an expert in their field. Also, the moderating function of trust on specific relations is tested (H7 – H8b). Concepts like power, trust, and resource allocation have been studied previously in literature. However, trust as a moderator variable has not been researched before within this context. These hypotheses will give an answer to the research question of this study: What is

the effect of power, trust and supplier independence on supplier resource allocation in a buyer-supplier relationship, and does trust have a moderating effect in this relationship?

3.2 Sample and data collection

The data in this study is collected in collaboration with three manufacturing companies. These are operating in the metal-, façade building-, and packaging industry. The first two companies are mainly operating in The Netherlands and have sometimes contacts with international suppliers. The third company is an international company with many international suppliers. Quantitative data is gathered by using the online survey tool Qualtrics. The questionnaire was sent to suppliers of these companies. Employees of the buying firms filled in a questionnaire about all the suppliers that cooperated in this research. Hereby the scope of the data collection is limited to buyer-supplier relationships in business-to-business settings.

Before the distribution, the survey was pretested by two master graduates and one pilot that did his training and education in England and the United States for two years. Their feedback resulted in some minor changes to the survey. Beforehand the companies supported the research by sending an email to a group of their suppliers with the question to cooperate in this research.

(16)

16

respondent. There was explained that there were no ‘good’ and no ‘bad’ answers and they were asked to fill in an answer that fits best with the relationship between them and the buying firm. Respondents were motivated by the possibility to get a management summary of the research. Two weeks after the first invitation for cooperating in the study, a reminder was sent to the respondents that did not yet filled in, or partially filled in, the survey. After respondents of the supplying companies filled in a survey, the contact persons of the buying firms were informed and asked to fill in a survey about the specific supplying firms.

The total sample size exists of 41 cases. Each case is a combination of a survey filled in by a supplier and a survey filled in by an employee of a buying firm of the specific supplier. In total 104 suppliers were invited to cooperate in the study. This leads to a response rate of 39.42%. The biggest parts of the respondents’ firms are located in The Netherlands. An Independent Samples Test was executed to test if there were differences between the data from Dutch firms and firms from other countries. There were 2 measurement items whereby the test reveals significant differences between the response of Dutch firms and firms from other countries. In table 1 below the characteristics of the respondents that participated in the study are provided. Most of the respondents are employees of companies located in The Netherlands, 80.48%. Most of these companies supplying the buying firms for more than 10 years, 64.85%. One of the companies changed their name after a business acquisition in February 2012. This can be a reason that multiple suppliers

filled in 5 years at the question how long they supplying the buyer, but supplying the same company under the old name before. Thereby the actual percentage could be higher.

(17)

17 FIGURE 1

Conceptual model

(18)

18 Profile of the sample

Frequency Frequency

Country of supplier Relationship with supplier

The Netherlands 80.48% ≤ 5 years 9.76%

Germany 2.44% > 5 years, ≤ 10 years 24.39%

Czech Republic 2.44% > 10 years, ≤ 15 years 19.51%

Romania 2.44% > 15 years, ≤ 20 years 9.76%

Italy 2.44% > 20 years 30.70%

United Kingdom 4.88% Unknown 4.88%

Austria 2.44%

Denmark 2.44% Annual turnover

0-5 million 21.95%

Respondent function 5.1-10 million 19.51%

Executive 4.88% 10.1-20 million 17.07%

Strategic 14.63% 20.1-100 million 12.20%

Tactical 26.83% >100.1 million 7.32%

Operational 53.66% Unknown 21.95%

3.3 Measures

In table 2 the measures of this study are shown. All constructs are measured on a five-point Likert-scale. These scales range from 1 (No, strongly disagree) to 5 (Yes, strongly agree).

The dependent variables supplier allocation of physical resources and supplier allocation of innovation resources will measure how the supplier allocates their resources better to a particular buyer than to this buyer’s competitors. These variables will be measured based on the scales used by Pulles et al. (2014). They developed these items based on the resource-based studies of Newbert (2008), Hunt and Davis (2008), and Surroca, Tribó and Waddock (2010). These constructs emphasize the extent to which a supplying firm will or will not prefer the specific buying firm.

Coercive power and reward power will be measured based on scales of Pulles et al. (2014). Pulles et al. (2014) based their

items of reward- and coercive power on scales of Terpend and Ashenbaum (2012) and Maloni and Benton (2000). The coercive power construct measures the extent to which the buying firm punishes the respondent’s firm if their firm does not conform to the buyer’s influence attempt (Pulles et al., 2014). The reward power construct emphasized the extent to which the buying firm aims to influence the respondent’s firm by offering benefits (Pulles et al., 2014).

(19)

19

measures the supplier’s believe that a buyer is capable to perform how is expected.

Expert power will be measured on scales adopted from Nyaga et al. (2013). These items measure the supplier’s opinion about the buyer’s business expertise. Supplier independence will be measured based on the scale of Krause et al. (2007). These

measurement items measure how

dependent on a buyer the supplier feels they are.

3.4 Data validity and common method bias

To check the reliability and validity for the measurement items used in this study several tests were conducted. First, SPSS 24.0 was used to check if the items factor

well. Herewith, the multi-item construct is measured for internal consistency reliability. The Kaiser-Meyer-Olkin Measure of Sampling Adequacy test (KMO) was used to check if the data factors well. This test is based on correlation and partial correlation. Additionally, the Bartlett’s test of sphericity tested if the measurement items correlate significant. All measurement items of this study were included for these tests. The KMO test has a value of 0.686 and thereby exceeds the threshold of 0.5 (Malhotra, 2008). The Bartlett’s test gave a significance of p=0.000. This means that the items correlate significant. After these tests, the factor loadings of the items were analyzed in SmartPLS. The factor loadings of the construct items are shown in table 2.

TABLE 2 Measurement Items

Constructs Measurement items Factor

loadings

Coercive power

(Terpend & Ashenbaum, 2012; Maloni & Benton, 2000) (Cronbach’s alpha = 0.901; composite reliability = 0.900; average variance extracted = 0.820)

1 (“no, strongly disagree”) to 5 (“yes, strongly agree”)

- We made it clear to this supplier that failing to comply with our requests will result in penalties against them - If this supplier did not do as asked, they did not receive

very good treatment from us

- If they do not go along with us, we threatened to withdraw certain services

0.999 Removed 0.800

Reward power

(Terpend & Ashenbaum, 2012; Maloni & Benton, 2000) (Cronbach’s alpha = 0.904; composite reliability = 0.924; average variance extracted = 0.805)

- We offer this supplier rewards so that we will go along with our wishes

- If this supplier did not do as asked, they did not receive the award offered by us

- If this supplier agrees with our requests, we offer them rewards.

0.954 0.732 0.984

Expert power

(Based on: Nyaga et al., 2013) (Cronbach’s alpha = 0.823; composite reliability = 0.895; average variance extracted = 0.740)

- This buyer is an expert in the industry

- We respect the judgment of buyer ’s representatives - This buyer has business expertise that makes them

likely

(20)

20 TABLE 2 (continued)

Constructs Measurement items Factor

loadings

Supplier independence (Based on: Krause et al., 2007) (Cronbach’s alpha = 0.694; composite reliability = 0.862; average variance extracted = 0.759)

- Our firm could easily replace this customer’s volume with sales to some other buyer.

- It would be relatively easy for us to find another buyer for the components we sell to this customer

- If the relationship with this customer was terminated, it would not hurt our firm’s operations.

0.814 0.925 Removed

Goodwill Trust (Sako & Helper, 1998;

Miyamoto & Rexha, 2004; also based on: Roy et al., 2004; Ireland & Webb, 2007) (Cronbach’s alpha = 0.666; composite reliability = 0.815; average variance extracted = 0.596)

- We can rely on this customer to help us in ways not required by our agreement with them.

- We can depend on this customer to always treat us fairly.

- This customer takes initiatives for mutual benefits that exceed the contractual agreements.

- We believe that this customer would make sacrifices for us to support our firm.

0.845 Removed 0.737 0.729

Competence trust (Sako & Helper, 1998;

Miyamoto & Rexha, 2004; also based on: Roy et al., 2004; Ireland & Webb, 2007) (Cronbach’s alpha = 0.723; composite reliability = 0.870; average variance extracted = 0.771)

- We feel that this customer is a highly capable partner. - This customer is very capable of providing value to our

firm.

- We trust that this customer has the managerial and technical capabilities to do what it says it will do.

0.807 0.944 removed

Physical resources (Based on: Hunt & Davis, 2008; Newbert, 2008; Surroca et al., 2010)

(Cronbach’s alpha = 0.912; composite reliability = 0.944; average variance extracted = 0.849)

Compared to our other customers. . .

- We grant this customer better utilization of our production facilities.

- we choose to give this customer priority in the

allocation of our products in the case of extreme events (e.g., natural disasters)

- We allocate our scarce materials to this customer in case of capacity bottlenecks.

0.913 0.937

0.914

Innovation resources (Based on: Hunt & Davis, 2008; Newbert, 2008; Surroca et al., 2010)

(Cronbach’s alpha = 0.898; composite reliability = 0.936; average variance extracted = 0.830)

Compared to our other customers. . .

- We are more willing to share key technological information with this customer.

- We share our best ideas with this customer first. - We dedicate more innovation resources to the

relationship with this customer. .

(21)

21

Next, the average variance extracted (AVE) was examined to test for convergent validity. Bagozzi and Yi (1988) suggests that the AVE should be higher than 0.5. As shown in table 2, all the constructs exceeded the threshold of 0.5. After this the internal consistency was tested for all measurement items by the Cronbach’s alpha. The Cronbach’s alpha for the different constructs range between 0.666 and 0.912. Two of the constructs do not exceed the threshold of 0.7 (Nunnally, 1978), but still have acceptable values. The composite reliability values range between 0.815 and 0.944. These all exceed the threshold of 0.7 (Bagozzi & Yi, 1988). These two outcomes are enough for indication internal consistency. To check for discriminant validity, the AVE numbers are compared with the correlations between latent variables. Hereby, the square root of AVE should be greater than the correlations between the latent variables beneath and on the left of the square root of AVE (Fornell & Larcker, 1981). When looking at table 3, it shows that the square root of the AVE is higher than the correlations with the other constructs in the same row and column.

This indicates discriminant validity for the constructs.

3.5 Data analysis

To test the hypotheses in this study, PLS path modelling is used. For this, the software tool SmartPLS 3.0 is used. Partial least-squares (PLS) structural equation modelling (SEM) is a regression-based approach. Hereby, SmartPLS tries to maximize the relations between the independent variable(s) and the dependent variable(s). PLS-SEM is often used for the developing of theories in exploratory research; it does this by explaining variance in the dependent variables by testing the model (Hair et al., 2016).

PLS is an appropriate tool when the sample size is rather small, consists of multiple explanatory variables, and variables are not normally distributed (Reinartz, Haenlein & Henseler, 2009). Also, PLS is preferred when formatively measured constructs are part of the structural model and the complexity is high with many constructs and many indicators (Hair et al., 2016). Because the sample size of this study is small, has high complexity, has formative

TABLE 3

Means, Standard Deviations, and Correlations of the Constructs

(22)

22

constructs and exists of multiple exploratory variables, PLS is more appropriate than methods like multiple linear regression for analyzing this study its hypotheses.

4. RESULTS

SPSS version 24 is used to test the proposed model. SmartPLS version 3.0 was used to test the robustness and the interaction effects in the model. First, the effects of power, trust and supplier independence on preferential resource allocation were tested (H1a – H6b). Then the moderating effect of goodwill trust on the relations between reward power and innovation resource allocation, and supplier independence and both physical and innovation resource allocation (H7 – H8b), were tested.

4.1 Results of structural model - Replication study

Figure 2 beneath shows the results of the replication study. As shown, coercive power is positively related to physical

resource allocation (β = 0.133, nonsignificant (NS)), but negatively related to innovation resource allocation (β = -0.084, NS). Both effects were as hypothesized, but these hypotheses are not supported because they are not significant. Reward power is negatively related to physical resource allocation (β = -0.146, NS) and innovation resource allocation (β = -0.112, NS). Therefore, H3a and H3b are rejected.

Goodwill trust is positively related to physical resource allocation (β = 0.092, NS) and innovation resource allocation (β = 0.049, NS). Therefore, both hypothesis H3a and hypothesis H3b are rejected. Competence trust has a positive relation on physical resource allocation (β = 0.049, NS). Thus, hypothesis H4a is rejected. The relation between competence trust and innovation resource allocation is negative (β = -0.025, NS). This is the opposite of what was expected, and therefore hypothesis H4b is rejected.

TABLE 4

Replication study of Pulles et al. (2014)

Hypothesis Outcomes of Pulles et

al. (2014)

Current outcomes

H1a: Coercive power is positively related to supplier allocation of physical resources.

β = 0.17 nonsignificant β = 0.133 nonsignificant H1b: Coercive power is negatively related to supplier

allocation of innovation resources.

β = -0.18 nonsignificant β = -0.084 nonsignificant H2a: Reward power is positively related to supplier

allocation of physical resources.

β = 0.34 significant at p<0.05

β = -0.146 nonsignificant H2b: Reward power is positively related to supplier

β = -0.112 nonsignificant H3a: Goodwill trust is positively related to supplier

β = 0.18 nonsignificant β = 0.092 nonsignificant H3b: Goodwill trust is positively related to supplier

β = -0.02 nonsignificant β = 0.049 nonsignificant H4a: Competence trust is positively related to supplier

β = 0.081 nonsignificant H4b: Competence trust is positively related to supplier

(23)

23

Supplier resource allocation

Supplier resource allocation Coercive Power Reward Power Goodwill Trust Physical Resource Allocation Innovation Resource Allocation

+

-+

Competence Trust

+

β = 0.1₃₃ β = -0 .08 4 β = -0.1₁₂ β = 0.08 1 β = -0.146 β = 0. 092 β =0.049 β = -0 .025 4.2 Comparison

When take a look at the results of Pulles et al. (2014) and this study, there are some similarities and some differences between the results of both studies. Results of both studies are shown in table 4. The effects of coercive power on both physical- and innovation resource allocation are the same in both studies. The effect of both goodwill- and competence trust on physical resource allocation is in both studies positive. In Pulles et al. (2014), reward power has a positive effect on both physical- and innovation resource allocation. The results of this study have reversed outcomes. In this study reward power has a negative effect on physical- (β

= -0.146, NS) and innovation resource allocation (β = -0.112, NS). In their study, goodwill trust has a small negative effect on innovation resource allocation (β = -0.02). The outcomes of this study shows a positive effect (β = 0.049). Lastly, competence trust has a positive effect on innovation resource allocation (β = 0.32) in the paper of Pulles et al. (2014). In this study, this effect is negative (β = -0.025).

4.3 Results of structural model - Study extension

(24)

24

both physical- and innovation resource allocation will be analyzed. Then the moderating effects will be discussed. The results are shown in figure 3 and table 5. Expert power has a positive effect on both physical- (β = 0.347) and innovation (β = 0.456) resource allocation. Both effects are significant (physical: p=0.044, innovation: p=0.023). Therefore, both hypothesis H5a and H5b are supported. Supplier independence has a positive effect on physical resource allocation (β=0.026, NS), and a negative effect on innovation resource allocation (β = -0.098, NS). Therefore, hypotheses H6a and H6b are

rejected. Goodwill trust has a decreasing effect on the impact of the relation between reward power and supplier allocation of innovation resources (β = -0.136 NS). Secondly, goodwill trust has a decreasing effect on the impact of the relation between supplier independence and supplier allocation of physical resources (β = -0.010). Thirdly, goodwill trust has a decreasing effect on the impact of the relation between supplier independence and supplier allocation of innovation resources (β = -0.028).

FIGURE 3

Results of Structural model – Extension

Reward Power Reward Power Supplier Independence Supplier Independence

Supplier resource allocation

Physical Resource Allocation Physical Resource Allocation Innovation Resource Allocation Innovation Resource Allocation Goodwill trust Goodwill trust Expert Power Expert Power β = 0.347 * β = 0 .456* β = -0 .0 1 0 β = -0.098 β = -0.112 β = -0 .0 2 8 β = 0.0 26 β = -0 .1 3 6

(25)

25 Extension on Pulles et al. (2014)

H5a: Expert power is positively related to supplier allocation of physical resources.

β = 0.347 significant at p= 0.044

H5b: Expert power is positively related to supplier allocation of innovation resources.

β = 0.456 significant at p= 0.023

H6a: Supplier independence is negatively related to supplier allocation of physical resources.

β = 0.026 nonsignificant H6b: Supplier independence is negatively related to supplier allocation of

innovation resources.

β = -0.098 nonsignificant

Moderating effects

H7: High goodwill trust strengthens the effect of reward power on supplier allocation of innovation resources.

β = -0.136 nonsignificant H8a: High goodwill trust weakens the negative effect of supplier independence on

β = -0.010 nonsignificant H8b: High goodwill trust weakens the negative effect of supplier independence on

supplier allocation of innovation resources.

β = -0.028 nonsignificant

4.4 Discussion of p-values

The use of p-values for determining if an hypothesis is ‘true’ or ‘false’ is been under discussion the last years. The statistical community has his concerns about reproducibility and replicability of scientific conclusions (Wasserstein & Lazar, 2016). Editors of Basic and Applied

Social Psychology came up with a radical

change based on doubts about the use of p-values. They decided to ban p-values in their journal (Trafimow & Marks, 2015). The Strategic Management Journal also made some changes in their publishing policy (Bettis et al., 2016). They will now also publish submissions of replication studies and studies with nonresults. Also, they will no longer accept studies that use cut-off levels of significance (Bettis et al., 2016). They now expect exact p-values and standard errors, and a discussion of effect sizes.

A group of scientists came together commissioned by the American Statistical

(26)

26

focus will be on effect sizes and robustness.

5. DISCUSSION

The purpose of this study was to replicate and extent the study of Pulles et al. (2014). Pulles et al. (2014) aimed to identify practices that help firms to acquire resources helping in achieving a competitive advantage. They did this by examining the impact of power and trust as relation mechanisms on the resource allocation by suppliers. By replication, the use of power and trust and their effect on supplier allocation of resources in a buyer-supplier relationship is studied. This topic is extended with the variables expert power, supplier independence and the moderating effect of trust. None, except two, of the measurements has significant support in the analysis. Therefore the discussion is based on the effects (β) of the relations in the structural model.

5.1 Replication study

There are similarities and differences between the outcomes of Pulles et al. (2014) and the outcomes of this study. Half of their hypotheses are confirmed in this study. Coercive power is positive related to the allocation of physical resources. This is consistent with their outcome, and contradicts with earlier studies that mostly conclude that coercive power has a negative effect (Brown et al., 1995; Maloni & Benton, 2000; Terpend & Ashenbaum, 2012). Contrary, as hypothesized is coercive power negatively related to allocation of innovation resources. This is consistent with earlier literature. The effect of coercive power is reversed with respect to physical innovations when intangible resources are in play (Pulles et al., 2014),

and suppliers are then disinterested in sharing new innovations or providing new ideas (Terpend & Ashenbaum, 2012). Reward power has opposite outcomes compared to Pulles et al. (2014). Their outcomes show positive impact of reward power on physical- and innovation resource allocation. In this study reward power has a negative effect on physical- and innovation resource allocation. This does not match with existing literature (Nyaga et al., 2013; Pulles et al., 2014). In SET literature is stated that suppliers choose for the highest reward when they have to choose between multiple buyers (Narasimhan et al., 2009). A possible explanation can be that the buying firms in these cases have competing buyers that offer higher or better rewards. These contrary results can also be explained by the perception of supplying firms. As stated in Pulles et al. (2014, p.20): ‘‘reward power, in contrast, is likely to encourage positive perceptions by the supplier (Nyaga et al., 2013)’’. In this study reward power is measured from buyer’s perspective and measures if the buyer offers or does not offer rewards to the specific supplier. But the feeling that a supplier has about received rewards will probably have influence on their allocation decision. If a buyer uses a kind of reward, it is not self-evident that a supplier sees this as a reward or is positively motivated towards the buyer.

(27)

27

allocation. Because in this study there is no distinction between small and large shares in supplier’s turnover, it can be said that the result of this study is consistent with their result and literature about goodwill trust (Tsai & Ghoshal, 1998). Goodwill trust is hypothesized to have a positive effect on supplier allocation of innovation resources. Were Pulles et al. (2014) found a small negative effect overall and a positive effect for the large share group, this study found a positive effect of goodwill trust on the allocation of innovation resources. This confirms literature stating that when goodwill trust is in play between suppliers and buyer, the supplier is more willingly to share technologies and knowledge to a buyer (Roy et al., 2004).

Competence trust has a positive effect on the supplier allocation of physical resources and is consistent with the study of Pulles et al. (2014). Goodwill trust was hypothesized as positively related to supplier allocation of innovation resources. Contrary to the outcome of Pulles et al. (2014), in this study a negative effect is found. This outcome does not fit with earlier literature that suggests that competence trust is important for collaboration on innovation (Madhaven & Grover, 1998). This can be explained by the statement of Roy et al. (2004) that states that a mismatch in competence trust can lower the chances of innovation performance in a relationship. Hereby mutual competence trust between a buyer and supplier is important. The supplier needs confidence that the buyer is able to use the service ordered and the buyer needs confidence that the seller can deliver the innovation resources (Frazier, 1999).

5.2 effect of expert power and supplier independence

First, unless there is less value attached to p-values in this study than normally due the fact that the sample size is relatively small, two significant relations are found. Expert power has a significant positive effect on the supplier allocation of physical and innovation resources. When supplier’s perception of buyer expertise is high this has a positive influence on their resource allocation decision. Were French and Raven (1959) argued that coercive and expert power are the least effective power instruments, results in this study show expert power as the most effective relation mechanism.

Supplier independence is negatively hypothesized to supplier allocation of physical and innovation resources. Results show an opposite outcome for supplier allocation of physical resources. Supplier independence has a positive effect on supplier allocation of physical resources in this study. Different antecedents can underlie this. This opens a path for future research. For the relation of supplier independence on supplier allocation of innovation resources a negative effect is found. This is consistent with the reasoning behind the hypothesis.

5.3 moderating effect of goodwill trust

(28)

28

positive. Results of this study show the opposite. Goodwill trust as a moderator even leads to a more negative effect on the relation between reward power and supplier allocation of innovation resources. Dapiran and Hogarth-Scott (2003) propose that the use of reward power bases in category management relations, like sharing information, is likely to lead to a desire to exit the existing relationship. This contradicts with other literature about the use of reward power in partner relationships. Therefore, further research on the use of reward power, its influences in relationships, and antecedents that can have an influence on reward power is necessary.

High supplier goodwill trust was expected to weaken the negative relation between supplier independence and the supplier allocation of both physical- and innovation resources. Were a negative effect was expected between supplier independence and resource allocation of physical resources that would be weakened by supplier goodwill trust, the opposite is going on. As mentioned above, supplier independence is positively related to physical resource allocation. Supplier goodwill trust weakens the positive relation of hypothesis h6a instead of strengthen it. Supplier goodwill was expected to weaken the negative effect of supplier independence on innovation resources and thereby creates a better resource allocation position. However, results show that supplier’s goodwill trust strengthen the negative effect of supplier independence on supplier allocation of innovation resources. This can be explained from SET literature. When a supplier is not that dependent on a buyer, they will probably not have a very high

personal relation and the supplier will prefer a relation with a buyer whereby more social exchanges occur.

These outcomes have some interesting insight for managers in manufacturing industries. When normal and standard products need to be bought, it is recommended to use coercive power when needed. The presence of goodwill- and competence trust is also good for acquiring better resources. When innovative resources like knowledge, human capital or new ideas need to be acquired the use of coercive power should be avoided. Expert power and goodwill trust will then be better relation mechanisms to use. When a supplier sees a buyer as an expert in the industry and respect the judgment of buyer’s representatives, they will likely see this buyer as a partner to work with. When suppliers have goodwill trust in a buyer, this will have positive effects for the buyer. This can let the supplier decides to give a preferential treatment to a buyer. A buyer should invest in a relation with a supplier develop supplier goodwill trust. To stay ahead of competitors in their field of business where supplier resources are shared with multiple buyers, buyers should try to continuously retain and improve their relationship with suppliers that are crucial for their performance. Suppliers that are important for competitive advantage should keep tight and investments should be made to develop and maintain the relation with these suppliers.

The influence of power, supplier independence and trust on a buyer’s preferred customer status