Asset specificity's impact on outsourcing relationship ...

Journal of Business Research 63 (2010) 657–666

Contents lists available at ScienceDirect

Journal of Business Research

Asset specificity's impact on outsourcing relationship performance: A disaggregated analysis by buyer–supplier asset specificity dimensions Glauco De Vita a,⁎, Arafet Tekaya a, Catherine L. Wang b a b

Oxford Brookes University Business School, Wheatley Campus, Oxford OX33 1HX, UK School of Management, Royal Holloway, University of London, Egham Hill, Egham, Surrey TW20 0EX, UK

a r t i c l e

i n f o

Article history: Accepted 27 April 2009 Keywords: Asset specificity Outsourcing relationship performance

a b s t r a c t Using hierarchical regression analysis on a sample of UK service firms, this study tests the impact of asset specificity on outsourcing relationship performance within a disaggregated methodological framework that allows to discern the specific effects of various buyers and suppliers' individual dimensions of asset specific investments. The results indicate that the impact of asset specific investments on outsourcing relationship performance varies according to the particular specificity dimension examined. While all statistically significant dimensions of buyers' asset specificity have a negative impact on relationship satisfaction, suppliers' human and dedicated asset specific investments exert a positive and significant influence. The results also show that, in three interaction instances, reciprocal specific investments are positively associated with outsourcing relationship performance. These findings have profound theoretical and methodological implications. © 2009 Elsevier Inc. All rights reserved.

1. Introduction According to transaction cost theory (TCT), under buyer–supplier relationship conditions of asset specificity (i.e., relationships involving non re-deployable investments specifically dedicated to the relationship), having to incur transaction costs to safeguard against costly opportunism makes vertical integration, rather than outsourcing, the most efficient and, hence, preferable governance structure. The literature widely investigates the transaction cost explanation for companies' boundary choice versus market governance under asset specificity conditions, and provides this explanation with considerable empirical support (Anderson and Coughlan, 1987; Klein et al., 1990; Levy, 1985; Masten, 1984; Monteverde and Teece, 1982). However, the literature overlooks the TCT's implication for outsourcing relationship performance in the presence of asset specific investments. This gap is particularly striking when considering that although factors influencing the make-or-buy decision are of great significance, of no less importance and, possibly, of greater relevance is the question of what happens to those companies that do choose to outsource under asset specificity conditions. Given the considerable measurement and contracting costs to safeguard against the hazards of asset specificity, the reported difficulties in crafting complex contracts and the well-known problems of

⁎ Corresponding author. E-mail addresses: [email protected] (G. De Vita), [email protected] (A. Tekaya), [email protected] (C.L. Wang). 0148-2963/$ – see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.jbusres.2009.04.019

outsourcing relationships to realize expected benefits (PA Consulting Group, 1993, 1996), this line of inquiry is particularly timely. The few studies that consider the question of the extent to which asset specific investments affect the performance of buyer–supplier relationships (Artz, 1999; Heide and Stump, 1995; Leiblein et al., 2002; Lui et al., 2009; Poppo and Zenger, 1998; Rodriguez and Padilla, 2005; Wang, 2002) produce mixed results from which discerning a conventional wisdom is difficult. These studies also bring to the fore a number of methodological and measurement issues. Aiming to address these issues, this study adds to this literature in a number of ways. First, studies that test hypotheses from TCT often use a loose definition of the buyer–supplier transactional relationship, thus failing to distinguish between transactions that involve the mere procurement of raw materials and/or intermediary inputs, and those which actually entail the external relocation of previously vertically integrated functions (which is how this article defines the outsourcing transaction). This distinction is important in that although there is no reason to assume that previously integrated functions should operate differently from all other external activities, only the operationalization of the former can reveal the extent to which the buyer obtains benefits compared to in-house production, which is the conceptual premise for the scale development of outsourcing relationship performance in this study. The only studies that specifically investigate the effects of asset specificity on the performance of such a buyer– supplier relationship (Poppo and Zenger, 1998, 2002; Wang, 2002) do not differentiate between buyers and suppliers' investments nor do they disaggregate data according to asset specificity dimensions.

658

G. De Vita et al. / Journal of Business Research 63 (2010) 657–666

Although both parties can undertake non re-deployable investments, with few notable exceptions (Artz, 1999; Buvik and Reve, 2001; Heide and John, 1990; Heide and Stump, 1995; Rokkan et al., 2003), previous empirical work only measures the asset specificity content of investments by one side of the buyer–supplier dyad. This study investigates the impact of asset specificity on outsourcing relationship performance while distinguishing between the effects pertaining to specific investments by both buyers and suppliers. Prior work that tests the effect of asset specificity on any measure of performance also raises the critical issue of the actual operationalization of the asset specificity construct. Morill and Morill (2003) argue that such a construct is not directly observable, requiring the use of multiple indicators. In their systematic assessment of TCT-related literature, David and Han (2004, p.54) conclude that the term asset specificity means “many different things to different people”. Yet, as early as 1985, Anderson (1985) calls for a more consistent and comprehensive scale development in an effort to reach a better approximation of the construct's multi-dimensional nature. Though efforts to inform the asset specificity scale from items relating to more than a single dimension of specificity are not uncommon (Geyskens et al., 2006; Klein et al., 1990; Levy, 1985; Zaheer and Venkatraman, 1995), with the exception of Masten et al. (1989) and Maltz (1993) (who do not test effects on relationship performance), they all still end up with the estimation of a single, albeit composite, asset specificity coefficient. The present study goes a step further, by testing the impact of asset specificity on the basis of a wide menu of distinct dimensions of specific investments by both buyers and suppliers. The study also examines the effect of reciprocal specific investments (i.e. the interaction effects of buyers' and suppliers' asset specificity dimensions) while controlling for the role of firm size, length of the relationship, type of function being outsourced, and (service-sector) industry type. To sum up, this is the first attempt that sets out to investigate empirically the direct impact of various dimensions of buyers and suppliers' asset specific investments on outsourcing relationship performance in relationships characterized by the externalization of previously integrated functions. 2. Theory and hypotheses Williamson's (1971, 1975, 1985) TCT identifies particular transaction characteristics upon which the discriminating choice of governance (i.e., hierarchy, hybrid or market) can be optimized subject to rational, transaction-cost-minimizing constraints. The main tenets of TCT (from which the most popularly testable hypotheses derive) relate to the effect of governance choice on firm performance given the influence these transaction characteristics exert on the transactioncost-minimizing tendency of the firm. These transaction characteristics are asset specificity (henceforth, AS), uncertainty, and frequency. Williamson (1989, p.142) defines AS in terms of “the degree to which an asset can be redeployed to alternative uses by alternative users without sacrifice of productive value”. The presence of asset specific investments gives rise to a safeguarding problem against opportunistic behavior. TCT, therefore, predicts that transactions characterized by high AS will be performed under a hierarchy, given the firm's ability to economize on contractual and monitoring costs through its more efficient internal control and conflict resolution mechanisms (Williamson, 1971, 1999). Uncertainty captures the degree to which ex-ante contractual costs and ex-post monitoring and enforcing costs are augmented by environmental and behavioral unpredictability, respectively. Conditional upon the presence of AS, the effect of uncertainty is to push transactions away from the market and towards a hierarchy since uncertainty incentivizes expropriation when a party's specific investment is exposed. Finally, the greater the frequency of transactions, the greater the monitoring costs. Transaction-cost-mini-

mization, therefore, makes governance through firm organization the preferable structure. Between the two extremes, market and hierarchy, lie hybrid relational structures which firms should favor in the absence of uncertainty and when frequent transactions occur at the intermediate range of AS (Williamson, 1985, 1991; Powell, 1990). Hybrid structures should allow firms to reap some of the benefits of vertical integration (lower transaction costs) alongside the economic gains that accrue from market transactions (in the case of outsourcing, cost savings and value adding). For a comprehensive assessment of the typical hypotheses developed from TCT's core propositions, this article refers the reader to the important reviews by Leiblein (2003), David and Han (2004) and Geyskens et al. (2006). On the other hand, the present study's interest centers on interrogating TCT on a different (though not unrelated) and still under-researched question. Specifically, this study asks whether TCT has anything to say on what happens to relational performance in the presence of deviations from the optimal form of governance predicted by the theory's core tenets, that is, what happens to relational performance in outsourcing relationships characterized by asset specific investments. TCT postulates that the hostage of asset specific investments increases the risk of opportunistic expropriation that stems from newly acquired (post-contracting) bargaining power and the threat of contract termination (Klein et al., 1978). This hazard creates what Williamson (1985) describes as a monopoly relationship whereby the disadvantaged party faces the unpleasant choice of continuing to work with its opportunistic partner or forgo the expected value of its specific investment (Anderson and Coughlan, 1987). Failure to safeguard against costly opportunism through adequate contracts essentially means that the party undertaking specific investments will be locked into the transaction, and will be vulnerable to opportunistic expropriation in the form of lower quality of the product/service and/or price/cost losses to the detriment of outsourcing performance (Klein et al., 1978). Against this backdrop, the answer of TCT to the research question that this study poses appears unambiguous. Under the hazards of AS, a high degree of uncertainty makes market governance “subject to costly haggling and maladaptiveness” (Williamson, 1985, p.89). Hence, TCT not only predicts that a hierarchy is the most effective governance mode because it reduces the costs of drafting and negotiating contracts necessary to safeguard the productive value of exposed assets. TCT also implies that should this transaction be performed through the market, unilateral specific investments will have negative economic and qualitative consequences on the performance of the outsourcing relationship. Exposure to opportunism is evidently more pronounced under conditions where the supplier unilaterally employs specific assets (Buvik and Reve, 2001). Faced with the buyer's opportunism while being locked into the relationship, the supplier's only option to save on the costs of the relationship is to cut back on the operational resources devoted to the focal product or service with a consequent negative impact on delivery performance as well as buyer's satisfaction. Under conditions where the buyer unilaterally employs specific investments, as eloquently put by Artz (1999, p.117): “TCE arguments predict that these assets can also have a negative effect on the OEM [Original Equipment Manufacturer]. Since specialized assets are worth little outside the present relationship, the OEM is dependent on the good-faith behaviour of the supplier to realize the value of its investment. Consequently, OEM control over that supplier is reduced (Heide and John, 1992). As control declines, the OEM is forced to expend more time and effort negotiating and monitoring contracts to safeguard its investment.. […]..Furthermore, since the supplier knows the OEM is at least somewhat ‘locked in’ to the relationship, its incentive to provide superior delivery performance is reduced. The increased transac-


tion costs and less favorable delivery performance likely result in lower satisfaction”. Our reading of this aspect of TCT, which highlights the potential for both buyers and suppliers' asset specific investments to affect negatively outsourcing relationship performance, also calls for reinterpretation of the role of AS itself. Whilst the implications of the core tenets of TCT ordinarily lead to the formulation of hypotheses in which AS moderates the governance-performance relationship, the specific research question of this article calls for testing whether different dimensions of AS (on both the buyers and suppliers' side) have any direct effect on the performance of outsourcing relationships. This narrower theoretical focus, in turn, demands even greater precision in the justification for the empirical disaggregation of the AS construct into distinct AS dimensions. Initially, Williamson (1983, p. 526) differentiates between four types of specific investments: i) human AS; ii) physical AS; iii) site specificity; and iv) dedicated AS. Williamson (1985, 1988) then adds brand capital and temporal specificity, resulting in a total of six dimensions. This study takes Williamson's classification of different AS dimensions as the conceptual rationale for its disaggregated analytical approach. As Lohtia et al. (1994, p. 264) suggest, in testing the predictions of TCT “the theoretical ramifications of each of the dimensions and types of transaction-specific assets also requires research attention. Research results based on TCE may be dependent on the specific type or dimension of the transaction-specific asset used in the research setting”. On the basis of the above, this article proposes the following hypotheses: H1a. An increase in buyers' asset specificity across different dimensions of non re-deployable investments negatively affects outsourcing relationship performance. H1b. An increase in suppliers' asset specificity across different dimensions of non re-deployable investments negatively affects outsourcing relationship performance. In a significant extension of Williamson's original framework, Klein and Leffler (1981), and Williamson (1983) point to relationships characterized by reciprocal specific investments as an additional safeguard device against opportunism. They argue that reciprocal investments can signal a credible commitment by both parties in an exchange relationship and, hence, reduce the trading hazard that can arise from AS through the creation of a symmetrical “mutual reliance relation” (Williamson, 1983, p. 528). As Williamson (1983, p.530) explains: “The offer of hostages poses a hazard of expropriation. One way to deter this is to expand the contracting relationship from one of unilateral to bilateral exchange…Reciprocity in these circumstances is thus a device by which the continuity of a specific trading relation is promoted with risk attenuation effects”. Although risk attenuation may also stem from a complementary effect due to the extended use of contractual safeguarding (Buvik and Reve, 2001), purely on the basis of the TCT notion of reciprocal exposure, this study proposes the following hypothesis: H2. Reciprocal non re-deployable investments (i.e. the interaction effects of the buyers and suppliers' AS dimensions) positively affect outsourcing relationship performance.

3. Methodology 3.1. Data The population consists of all companies operating in the UK and belonging to one of the following service-related industries: (i) banking

659

and finance (UK SIC code 651); (ii) hotel industry (UK SIC code 551); (iii) IT industry (UK SIC code 722); and (iv) telecommunication industry (UK SIC code 642). Studies that examine the impact of AS on governance choice or outsourcing relationship performance usually focus on firms operating in the manufacturing sector, leaving a glaring gap in the knowledge of this relationship in the context of service-related firms. The sample consists of 2286 companies randomly selected from the Financial Analysis Made Easy database using the stratified sampling method. With the objective of minimizing key-informant bias, we sought to identify respondents highly knowledgeable about their firms' outsourcing activities. Consultation with the UK National Outsourcing Association (NOA) indicated that, unlike manufacturing firms where a procurement senior-level manager is usually in place, in UK servicesector firms, HR directors are usually the most knowledgeable informants about their companies' outsourcing activities (which often involve restructuring and/or redundancies). Furthermore, HR related functions generally rank among the top activities being outsourced (Wahrenburg et al., 2006). Pre-survey phone conversations with a randomly selected number of firms confirmed that HR directors fitted the profile. Given their knowledge of outsourcing activities in their respective companies, HR directors were also invited (as the first contacts) to identify a more suitable respondent should one exist. To motivate company directors to participate in the study, an incentive in the form of a summary report of key findings was also offered. The research design follows Dillman's (1978) total design method. The pre-test was conducted with knowledgeable academics and twenty-five representative company directors who reviewed, filled-in and critiqued initial versions of the questionnaire for both content and clarity. The revised mail questionnaire consisted of a cover letter, a tearoff detachable participant information sheet, and the set of questions. Respondents were asked at the beginning of the questionnaire to identify one of the most significant functions being outsourced at the time of the questionnaire. This function then served as the referent for all remaining questions. The pilot-testing process involved the administration of the questionnaire to a further sample of 50 companies. Questions that did not provide useful data were discarded and ambiguous areas were clarified. Cronbach's alpha tests demonstrate an acceptable level of reliability. Data collection was accomplished by means of a mail questionnaire and a follow-up mail survey of non-respondents. Out of 2286 questionnaires dispatched, a total of 286 were returned. Among the 286 respondents, 20 refused to take part due to lack of time and companies' policy. In addition, 118 respondents of the remaining 266 reported that they had not been engaged in any outsourcing projects, resulting in a 6.4% response rate. Because of incomplete information, only 137 out of the remaining 148 responses were usable. Although this study acknowledges potential concerns with a fairly low rate of return, for a mail survey of this kind, response rates of less than 10% are not uncommon (Lepack et al., 2003; Perez-Nordtvedt et al., 2008). Since the data come from a single survey instrument and a single informant per outsourcing project, susceptibility to common method bias is a potentially serious issue. This concern is particularly important in the context of the present study since unilateral specific investments by the buyer increase its dependency to the supplier and this dependency trap might influence the way the buyer perceives the outsourcing relationship. To address such concerns, this study follows Podsakoff et al.'s (2003) procedural remedies (the path of their ‘Scenario 7’). Accordingly, at the questionnaire design stage, the study applied Malhotra and Grover's (1998) Ideal Survey Attributes. It also employed Parkhe's technique of separation of scale item (1993) and guaranteed response anonymity. Finally, Harman's (1967) one-factor test was used. The analysis involved entering all the dependent and independent variables in an exploratory factor analysis to examine the unrotated factor solution so as to identify whether a single or major factor would emerge. If a substantial amount of common method is present in the data, a single or general

660


factor accounting for the majority of the variance is expected (Podsakoff and Organ, 1986). An unrotated principal components factor analysis on all continuous variables revealed that no single factor emerged and the largest factor only accounted for 23.1% of the variance. These results suggest that common method bias is not a major issue. To test for non-response bias, respondents were divided into two groups: (i) early respondents; and (ii) respondents to the follow-up mailing (late respondents). In line with the extrapolation method over successive waves, it was assumed that late respondents were “less readily” to take part in the survey and, consequently, were considered most similar to non-respondents (Armstrong and Overton, 1977). Chi-square tests showed no significant differences between both groups with respect to key categorical variables (i.e. firm size and industry type). In addition, using the one-way (between groups) ANOVA test, the two groups were also compared on all continuous variables. The results confirmed that there was no significant difference (p b .05) between the two groups, indicating the absence of significant response bias. We also compared the firm size of companies that responded with that of a sample of 50 non-respondents. No significant differences were identified. Lastly, we checked for the reliability of the self-reported data (with respect to size and industry type) through public data and telephone contact. The 10 companies approached by phone were randomly chosen and the information was cross-checked with the responses of the questionnaire. The remaining 15 were chosen because of the availability of public data. No biases with the data provided by informants were detected. All survey items were adopted from the literature. Most of the measurement items are tested items with strong precedents in prior studies though, in a few necessary instances, new items are developed drawing from theoretical insights. All the measurement scales are rated on a seven-point Likert scale, from 1 = strongly disagree to 7 = strongly agree (Appendix A reports the full list of items and detailed results of the exploratory factor analysis). 3.2. Dependent variable Outsourcing relationship performance is the dependent variable in this study. Scales describing the buyer's performance evaluation of a given buyer–supplier relationship fall into two broad categories. First, scales which, aiming to capture the performance outcome of the alignment of organizational form with transaction characteristics, place significant emphasis on the direct measurement of strictly defined transaction costs to quantify governance efficiency (Artz, 1999; Artz and Brush, 2000). The second category includes scales which, being more concerned with the overall satisfaction of relational performance, do not measure governance costs explicitly, focusing instead on overall cost savings, realization of expected benefits and various indicators of relationship satisfaction such as the quality of the product/service (e.g., Poppo and Zenger, 1998, 2002). The narrower theoretical scope of the present study (and its definition of outsourcing as the externalization of previously integrated functions), dictates the adoption of the latter operationalization of the construct. This approach is particularly appropriate in our context since it is reasonable to assume that in answering the question of whether the firm has or “has NOT achieved the target level of cost savings expected by outsourcing this function”, the firm's benchmarking calculation of net benefits, factors in any contracting, negotiating and safeguarding costs incurred in outsourcing the function vis-à-vis in-house production. Accordingly, this study treats outsourcing relationship performance as a construct based both on the buyer's realization of contextdependent expected benefits, and on overall satisfaction of relationship performance. As noted in Lui et al. (2009), previous research shows that a manager's subjective satisfaction with a partnership is

a good indicator of the partnership objective performance. Outsourcing relationship performance (α = 0.93) is measured by nine items. In addition to overall cost savings, items intend to capture additional qualitative factors that typically underpin relational performance (as perceived by the buyer), including suppliers' reliability (Grover et al., 1996), suppliers' responsiveness (Goodman et al., 1995; Lacity et al., 1996) and the degree of assistance in information sharing (Ghani and Khan, 2004). 3.3. Independent variables This study disaggregates the measurement of AS into various buyers and suppliers' dimensions. However, the analysis that follows excludes some dimensions since, in practice, they only apply to one side of the buyer–supplier dyad. For example, buyers' site specific investments are, at best, extremely rare. Similarly, the buyer's temporal AS is a construct devoid of a practical counterpart on the supplier side since the supplier's investment in the relationship is never compromised by the reputation costs brought about by the buyer's actions. Since the study targets service-related industries which, by their very nature, are less likely to require physical investments, physical AS is also dropped to be replaced by procedural AS. This decision is consistent with that of Zaheer and Venkatraman (1995), who develop the procedural AS dimension to better capture physical AS in the service sector, where considerable investments in physical components and tools are less likely. Given the above, from the buyer's side, this study covers five AS dimensions, namely: (i) human AS; (ii) dedicated AS; (iii) temporal AS; (iv) procedural AS; and (v) brand capital. On the basis of the earlier considerations, the study considers three suppliers' AS dimensions: (i) human AS; (ii) dedicated AS; and (iii) site specificity. 3.3.1. Human asset specificity Human AS relates to knowledge-specific assets (Dibbern et al., 2005) that arise from “learning-by-doing” (Williamson, 1996, p. 105) and which have limited transferability to other work settings (Lamminmaki, 2005). According to Ruchala (1997), human AS involves not only the expertise that is required for carrying out a particular function but also the costs of training and development to support interactions with the other party. Buyers' human AS (α = 0.83) is measured by three items, including: “Your company has acquired new knowledge in order to adapt to the specific technological norms of your supplier”; and “Your company has recruited additional staff for the sole purpose of managing the outsourcing relationship”. These items come from Bucklin and Sengupta (1993) and Heide and John (1992), and intend to capture the degree to which skills, knowledge and experience of buyers' personnel, are specific to the requirements of dealing with the supplier (Zaheer and Venkatraman, 1995). Suppliers' human AS (α = 0.77) is a four-item scale designed to capture buyers' perception of the extent of non re-deployable, knowledge-related investments made by the supplier in terms of customization of workflows and routines (Rodriguez and Padilla, 2005), and of the level of adaptation made through training (Brouthers and Brouthers, 2003). 3.3.2. Dedicated asset specificity Dedicated AS refers to investments that are of general purpose as opposed to specialized uses (e.g., physical AS) but which result from a particular transactional agreement likely to entail a longterm relationship. Should this relationship end prematurely, excess capacity will, however, be created (Williamson, 1983; Joskow, 1987; Lamminmaki, 2005). For example, a production contract with one large customer may cause a company to expand its capacity to meet demand, which may result in significant overcapacity and important financial disruption if the customer in question chooses not to renew the contract (Ruchala, 1997). Under


certain circumstances, this dimension also applies to buyers (e.g. additional investment in laboratory accessories that help the company assess the quality of a bigger proportion of goods or services). Drawing from the theoretical insights of Williamson (1983), Joskow (1987) and Lamminmaki (2005), this study develops three items to measure buyers' dedicated AS (α = 0.82). Example items are: “For the purpose of the outsourcing relationship, your company has made additional investments in quality control facilities which are likely to result in excess capacity in the event of contract termination”; and “In order to cope with the ‘weight’ of the relationship with this supplier, your company has made additional investments (e.g. in communication facilities) that are likely to result in excess capacity in the event of contract termination”. Suppliers' dedicated AS (α = 0.89) is measured by four items. With the exception of one item adopted from Dyer (1996), given the lack of studies offering tested items, this study develops three items on the basis of the conceptual discussions in relevant literature. For example: “Should your outsourcing relationship cease, your supplier would be left with substantial unsold output or excess capacity”. 3.3.3. Site specificity This dimension refers to a relationship that requires close proximity for reducing inventory and/or other processing costs. Once in place, however, the relocated assets are highly immobile (Williamson, 1983; Joskow, 1987; Morill and Morill, 2003; Lamminmaki, 2005). Commenting on the importance of this dimension, Ruchala (1997, p. 21) observes that “without this site, a very inexpensive [transaction] would become very expensive”. Most studies that try to operationalize the site specificity construct tend to measure the physical proximity between the two parties using the distance between the parties' premises as a proxy (Joskow, 1987; Ghani and Khan, 2004). However, this approach overlooks whether or not physical proximity is specifically due to the transactional relationship. Since site specific investments (involving relocation) tend to be incurred by the supplier with the aim of securing a long-term relationship with its customer, this study measures buyers' perception of how important close proximity is to the supplier, and the extent to which the supplier's relocation decision is specific to the outsourcing relationship. The suppliers' site specificity (α = 0.91) scale is measured by four items developed by Joskow (1987) and Nishiguchi (1994). 3.3.4. Temporal asset specificity Temporal AS refers to transactional relationships where timing and coordination are of high importance (Lamminmaki, 2005). Here, criticality lies on a well-timed response from on-site human assets (Lohtia et al., 1994). According to Malone et al. (1987, p. 486) “an asset is time specific if its value is highly dependent on its reaching the user within a specified, relatively limited period of time”. Due to its link to the timing of a service delivery, most risks emerging from

661

temporal AS are on the buyer side. Since the buyer's investment in the relationship itself can be compromised by the temporal costs brought about by the supplier's actions, the latter can make opportunistic demands and/or enhance individual profit. As with all other dimensions of AS, the buyer is then left with the option to remain in the relationship and tolerate opportunistic actions or leave the relationship and incur the associated switching costs. Buyers' temporal AS (α = 0.91) is a three-item scale that draws from the items proposed by Brown and Potoski (2005) and Masten et al. (1991). 3.3.5. Brand capital This AS dimension relates to reputation investment. An outsourcing relationship involving functions that have a direct impact on the overall firm's reputation is one of high brand capital specificity. In such a case, the outsourcing supplier could find itself in a position enabling it to intentionally or unintentionally cause damage to the buyer's reputation (Gatignon and Anderson, 1988; Lohtia et al., 1994). A typical example is the outsourcing of restaurants within the hotel industry, where a bad reputation could prove very costly to the overall hotel business (Lamminmaki, 2005). Evidently, this AS dimension is more likely to be a concern for buyers. The buyers' brand capital scale (α = 0.78) is measured by three items taken from Gatignon and Anderson (1988), Levy (1985) and Lohtia et al. (1994). 3.3.6. Procedural asset specificity Zaheer and Venkatraman (1995) develop this AS construct as a substitute for the physical AS dimension in the service industry. The term refers to organizational routines and workflows tailored to a particular relationship, which are difficult to modify once created, or to re-deploy to other purposes without value loss. Buyers' procedural AS (α = 0.70) is measured by the three items used in Barthelemy and Quelin (2002), Zaheer and Venkatraman (1995) and Heide and John (1990). 3.4. Interaction effects Following Jaccard et al. (1990), to test H2, reciprocal specific investments are assessed by examining the symmetrical asset specificity dependence stemming from the resulting (15) interactions between all the buyers and suppliers' AS dimensions considered (for a similar approach see Heide, 1994). 3.5. Control variables Although here interest lies in developing a parsimonious model, to account for structural power-dependence (e.g. the possibility that large firms may exert undue influence on partner behavior by virtue of their superior bargaining position), this study

Table 1 Descriptive statistics and correlations. Variable

Mean

S.D

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(1) (2) (3) (4) (5) (6) (7) (8) (9)

5.08 2.26 1.90 5.13 5.22 3.39 4.07 2.76 1.94

1.12 1.46 1.20 1.74 1.50 1.47 1.48 1.61 1.65

1.00 − .16 − .16 −.48** − .43** −.28** .03 .04 −.14

1.00 .35** .09 .11 .34** .28** .26** .01

1.00 .05 .18* .22** .23** .36** .13

1.00 .56** .21* .16 .11 − .12

1.00 .30** .22* .15 .06

1.00 .42** .30** .06

1.00 .23** .03

1.00 .25**

1.00

Outsourcing performance Buyers' human AS Buyers dedicated AS Buyers' temporal AS Buyers' brand AS Buyers' procedural AS Suppliers' human AS Suppliers dedicated AS Suppliers site specificity

n = 137. * p b .05; ** p b .01.

662


Table 2 Results of regression analysesa. Variables

Hypothesis prediction

Intercept Control variables Firm size Relationship length Industry type Banking and finance Hotel IT Type of activity being outsourced HR-related IT-related Housekeeping Asset specificity (AS) Buyers' asset specificity Human AS Dedicated AS Temporal AS Procedural AS Brand AS Suppliers' asset specificity Human AS Dedicated AS Site specificity Reciprocal investment b Buyers' temporal X Suppliers' site Buyers' dedicated X Suppliers' human Buyers' temporal X suppliers' dedicated R2 Δ R2 ΔF

Model 1

Model 2

Model 4b

Model 3

1.86***

(0.49)

3.83***

(0.61)

4.19***

(0.60)

4.85***

(0.59)

−0.14 2.24***

(0.09) (0.24)

− 0.07 1.75***

(0.09) (0.24)

− 0.10 1.48***

(0.09) (0.24)

− 0.07 1.11***

(0.08) (0.23)

−0.20 − 0.30 0.02

(0.22) (0.24) (0.22)

− 0.15 − 0.20 0.03

(0.21) (0.23) (0.21)

− 0.22 0.04 0.10

(0.19) (0.23) (0.19)

− 0.13 0.18 0.18

(0.18) (0.21) (0.18)

0.14 − 0.05 0.26

(0.26) (0.29) (0.35)

0.29 0.10 0.33

(0.24) (0.28) (0.33)

0.25 0.07 0.36

(0.23) (0.26) (0.31)

0.19 − 0.06 0.14

(0.21) (0.24) (0.29)

− 0.06 − 0.03 − 0.11* − 0.06 − 0.11†

(0.05) (0.06) (0.05) (0.05) (0.06)

− 0.03 −0.08 − 0.15** − 0.15** − 0.14*

(0.05) (0.06) (0.05) (0.05) (0.06)

− 0.02 − 0.10† − 0.20*** − 0.09* −0.15**

(0.04) (0.06) (0.04) (0.05) (0.05)

0.18** 0.12* − 0.11*

(0.05) (0.05) (0.04)

0.14** 0.13** − 0.07†

(0.05) (0.04) (0.04)

H1a (−)

H1b (−)

H2 (+)

.41

.50 .09 9.81***

.58 .08 10.57***

0.06** 0.10* 0.04* .65 .07 11.64***

(0.02) (0.04) (0.02)

a

n = 137. The coefficients are unstandardized estimates, with standard errors in parentheses. † p b .10; * p b .05; ** p b .01; *** p b .001. Model 4 estimates all 15 interaction terms though, to conserve space, we only report the results obtained by including the three significant interactions. The full results are available from the authors upon request. b

includes firm size as a control variable (without imposing any a priori restrictions on the parameter). Following Poppo and Zenger (1998) and Lazzarini et al. (2008), firm size is the log of total number of employees in buyers' companies. This measure constitutes a reliable proxy, devoid of the subjectivity inherent in relying on the buyer's perception of relative size (which would inevitably be based on the buyer's guess of the supplier's size). In addition, this study controls for the type of function being outsourced (HR-related activities, IT maintenance and development, housekeeping, and payroll) and for service sector industrytype (banking and finance, IT, hotel and telecommunication) using dummy variables. Since the duration of time that outsourcing partners have worked with one another may affect the performance of the relationship, we also control for relationship length measured by the log of the total number of months that the buyer has been involved with the supplier. 3.6. Estimation method This study employs hierarchical regression analysis. The virtue of this estimation technique is that it enables to assess changes, and the statistical significance of changes, in the proportion of variance explained (R2) as additional variables are progressively introduced into the model. 4. Results Table 1 reports descriptive statistics and correlations. Although some of the correlations between the variables are relatively high

(0.55 for buyers' temporal AS x brand capital was the highest), inspection of variance inflation factors (VIF) does not indicate a problematic level of multicollinearity (VIF values being significantly lower than 10) (Cohen et al., 2003). Examination of both the residual scatter plot and the normal probability plot confirms a normal distribution. Also, prior to calculating the interaction terms to test H2, the variables were mean-centered so as to minimize multicollinearity (Aiken and West, 1991). Following the purification of the multi-item scale using item-tototal correlations and factor analysis, all constructs exhibit satisfactory levels of reliability, with Cronbach's alphas ranging from 0.70 to 0.93 (Appendix A). Convergent validity can be judged by examining item loadings. Each item is significantly related to its underlying factor, and all item loadings are well above the cut-off value of 0.50 (Hair et al., 1998) while their cross loadings are low, hence the results support convergent validity. Table 2 reports the estimation results. Model 1 includes the control variables. Model 2 adds the buyers' specificity dimensions. Model 3 introduces the suppliers' AS dimensions, and Model 4 tests for the effect of reciprocal investments by adding the interaction terms (to conserve space, Table 2 only reports the three significant interactions out of the 15 estimated). The hierarchical regression framework confirms that, as the analysis adds variables, and moves from Models 1 to 4, the incremental variance in outsourcing relationship performance is significant (R2 change is 0.09 (p b .001) in Model 2; 0.08 (p b .001) in Model 3; and 0.07 (p b .001) in Model 4). Model 4 explains 65% of variance. In support of H1a, Models 2, 3 and 4 show that the coefficients of all buyers' AS dimensions are consistently negative though, in some


instances (e.g. buyers' human and dedicated AS), they are insignificant. The estimates in relation to suppliers' AS dimensions portray a more mixed picture and hence reject H1b. The estimated coefficient of suppliers' site specificity has the expected negative sign but is statistically insignificant in Model 4. On the other hand, suppliers' human AS (p b .01; p b .01) and dedicated AS (p b .05; p b .01) have a significant and positive effect in both Models 3 and 4. With respect to reciprocal specific investments, computations of all the 15 interaction terms among buyers and suppliers' AS dimensions support H2 in three interaction instances, whereby each of the three suppliers' AS dimensions significantly and positively interacts with one buyers' AS dimension. Looking at Model 4, suppliers' site AS (earlier negative, and significant in Model 3) now, when interacting with buyers' temporal AS, positively affects relational performance (p b .01). The interactions between suppliers' human AS and buyers' dedicated AS (pb .01), and between suppliers' dedicated AS and buyers' temporal AS (p b .01) also have positive effects on outsourcing relationship performance. None of the control variables are statistically significant except relationship length. 5. Discussion and implications The lack of empirical studies that adopt a disaggregated analytical framework for the treatment of AS as well as differences in the operationalization of several constructs, including “outsourcing transaction” and “performance” makes the comparison with previous results, at best, difficult. Nevertheless, it is worth pointing out that the results in relation to H1a are consistent with those by Artz (1999), who in two of his three models finds that buyer-specific assets increase “transaction costs” and decrease “buyer satisfaction”, and those by Heide and Stump (1995), who find that buyer-specific assets have negative effects on “relationship performance”. The results of the present study also validate, and are corroborated by, those of Leiblein et al. (2002), who find that deviation from the optimal form of governance as dictated by transactional characteristics, has a negative effect on “technological performance” when contractual safeguards are inadequate for the hazards present in a given relationship. The results partially support H2, since only some dimensions of reciprocal investments positively affect relationship performance. The statistical insignificance of most of the interaction terms may relate to the value of the mutually sacrificed hostages. Since low values in our estimated interaction terms also represent reciprocity by mutually low AS, it is reasonable to suppose that while symmetrical specific investments of low value may signal sufficiently credible commitments to reduce the problem of moral hazard, such commitments may be insufficient to generate a positive effect on relationship performance. These results complement those by Artz (1999), who uses data from the manufacturing sector, though the disaggregated framework employed in the present study is more informative by additionally revealing which specific AS dimensions display significant interactions. None of the control variables is significant except relationship length, suggesting that the longevity of outsourcing relationships helps such relationships to become more cooperative. Moreover, a long prior history of collaboration between two firms can reduce the level of contractual governance and hence allow the outsourcing partners to save on transaction costs (Buvik and Haugland, 2005). The lack of empirical support for H1b is the most important result stemming from the present analysis. While suppliers' site specific investments have a negative effect on outsourcing relationship performance (though only significantly so at p b .10 in Model 4), both suppliers' human and dedicated AS dimensions consistently display positive and significant coefficients. It would be convenient at this stage to invoke alternative paradigms, particularly the competence/resource or the relational exchange

663

perspectives, to explain the mixed effects of buyer- and supplierheld specific assets. When considering these competing theoretical frameworks, it becomes evident that unilateral specific investments have the potential to both reduce as well as promote opportunism. For example, Rokkan et al. (2003) suggest that the presence of strong solidarity norms cause a shift in the effect of specific investments, from negative (due to opportunism-driven expropriation), to positive, due to bonding. Furthermore, Lui et al. (2009) find that AS positively correlates with trust which, in turn, promotes cooperative behavior and partnership satisfaction, an important finding that supports the premise of relational exchange theory and highlights the inadequacy of TCT as the sole, dominant framework. However, remaining within TCT reasoning, which defines the theoretical boundaries of the scope of this investigation, the explanation of why buyer-held and supplierheld specific assets lead to different effects may lie in the assumption of the inadequacy of contractual safeguards. Indeed, it is possible that when the supplier unilaterally employs specific assets, the outsourcing relationship will rely more on formal contracts to offer additional safeguards than it is the case when buyer-held specific assets are exposed. Previous empirical evidence supports this supposition. For example, Buvik and Reve (2001) demonstrate that the level of formalized purchase contracting is significantly lower when the buyer unilaterally deploys specific assets than when the suppliers' specific investments are at risk. One further explanation consistent with TCT that this article puts forward to rationalize the result of the mixed effects of buyers and suppliers' AS dimensions is that although all asset specific investments are, by their very nature, dedicated to the particular requirements of the relationship, the threshold at which they begin to trigger opportunistic expropriation may vary across AS dimensions due to the value (cost) and non re-deployability content embedded in them. Given their intrinsically limited value and/or transfer cost vis-à-vis other AS dimensions (e.g. site specificity), relatively low levels of suppliers' human and dedicated AS may represent insufficient hostages to trigger expropriation, leading instead to a more effective and efficient relationship. Only very sizeable investments in such assets would, by taking AS beyond the hostage threshold, invert the sign of the relationship between AS and outsourcing relationship performance as the buyer's individual gain from opportunistic expropriation is perceived to outweigh the shared benefits from an improved outsourcing relationship. On the other hand, given their intrinsically high value and non re-deployability content, the threshold at which suppliers' site specific investments lead to opportunistic behavior to the detriment of the relationship is likely to be low. Hence, even very low levels of suppliers' site specificity may take the relationship beyond the hostage threshold, resulting in a negative effect. This hostage threshold rationalization of the findings not only implies that the point at which opportunistic expropriation is triggered varies according to the particular type of investment across both buyers and suppliers' AS dimensions, it also suggests that the relationship between each dimension of AS and relational performance may not be a constant function but one which starts with a positive effect that reduces to insignificance as it approaches the hostage threshold and, only passed that stage, assumes negative connotations. Although this intuitive explanation as to why suppliers' investment in both human and dedicated AS appear not to be reciprocated by opportunistic expropriation from buyers, may seem as promising as other alternative explanations, from a theoretical perspective, these findings make a significant contribution to the existing knowledge of the predictive power of TCT when applied to the question of the impact of AS on outsourcing relationship performance. From a methodological perspective, an implication of this study concerns the approach by which future empirical work should operationalize the AS construct. Since outsourcing relationship

664


performance responds differently to different types of buyers and suppliers' specific investments, these findings confirm the empirical necessity to treat AS as more than a mere composite construct. Indeed, even when the measurement scale of AS draws from items relating to different AS dimensions, estimation of a single AS coefficient would at best mask the individual effects of the different types of specific investments that characterize the buyer–supplier relationships examined. By shedding light on the various effects of each particular AS dimension from both the buyer and the supplier's side, the findings also have implications that should be of interest to managers. To frame it boldly, the question for managers now becomes “why should buyers unilaterally employ specific assets if it is shown that, in the absence of adequate contractual safeguards, this endeavor results in poor outsourcing relationship performance?” Drawing purely from TCT-reasoning, this article proposes a careful assessment of how the value and non re-deployability content of specific investments ranks against the (AS dimension-specific) hostage threshold point. However, it is obvious that the question this article poses to managers is best answered by also bringing into consideration other factors that affect relationship performance, factors that go beyond the ones the single TCT paradigm accounts for. 6. Limitations and avenues for future research Notwithstanding the contribution of these findings, several limitations must be acknowledged. First, this study limits the data to outsourcing relationships in four UK service-related industries. This constraint implies caution in generalizing the findings. Accordingly, the article encourages replication studies across different countries and sectors. Second, although this study's disaggregated approach allows to differentiate between buyers and suppliers' specific investments, the data solely focus on the buyers' perception of the relationship. While such an approach may represent a potential source of

Appendix A. Reliability measures and factor loadings

response bias, there is sufficient evidence in the extant empirical literature to support the contentions of the validity of reliance on a single, senior-level informant (Lui et al., 2009; Zaheer and Venkatraman, 1995), and of the unproblematic nature of the assumption that buyers and suppliers hold consistent perceptions of the relationship (Artz, 1999; Poppo and Zenger, 2002; Provan and Skinner, 1989; Saxton, 1997). Here interest exclusively centers upon the relationship between AS and outsourcing relationship performance implied by TCT. Although the value of incorporating concepts such as collaborative ties, relational norms, trust, reputation and managerial skills into the transaction cost reasoning remains debatable, many studies find that these factors have explanatory power in the determination of various measures of outsourcing performance (Conner and Prahalad, 1996; Chiles and McMackin,1996; Hill,1990; Klein and Leffler,1981; Madhok, 2002; Murray, 2001; Rokkan et al., 2003; Zaheer et al., 1998; Zajac and Olsen, 1993). Going beyond TCT, future work could, therefore, control for the role these variables play in moderating the relationship in question. Finally, although this study estimates the impact of AS on relational performance within a straightforward linear regression framework, one cannot exclude that the functional form of the relationship be of a non-linear nature, nor can one rule out the possibility that relational performance itself may, under certain conditions, and over time, influence (induce or deter) asset specific investments. In these circumstances, the bias stemming from the linear approximation of the true model's nonlinearities as well as endogeneity bias due to feedback effects from relational performance to AS may seriously affect the reliability of the estimates. To address these issues, a profitable avenue for future research entails constructing a longitudinal dataset to test a structural, inter-temporal model that would allow establishing both the dynamic structure and the reverse causality properties of the relationship in question. This extension, however, awaits more data.

a

Factor and contents Buyers' temporal asset specificity 1. The product or service provided by your supplier requires timely delivery. 2. In the relationship with your supplier, precise scheduling is very important. 3. Punctual delivery from your supplier is crucial; hence any delay will result in a significant cost to your company (e.g. loss of clients). Buyers' human asset specificity 1. Your company has invested considerably in the training of personnel for the purpose of the relationship. 2. Your company has acquired new knowledge in order to adapt to the specific technological norms of your supplier. 3. Your company has recruited additional staff for the sole purpose of managing the outsourcing relationship. Buyers' dedicated asset specificity 1. For the purpose of the outsourcing relationship, your company has made additional investments in quality control facilities which are likely to result in excess capacity in the event of contract termination. 2. In order to cope with the ‘weight’ of the relationship with this supplier, your company has made additional investments (e.g. in communication facilities) that are likely to result in excess capacity in the event of contract termination. 3. For the purpose of the outsourcing relationship, your company has made additional investments that would result in excess capacity in the event of contract termination. Buyers' brand capital 1. Any underperformance from your supplier will result in a highly negative effect on your company's reputation. 2. In the industry in which your company operates, you cannot afford receiving a low quality product or service from your supplier since this will negatively affect your reputation. 3. Given the importance of your company in the market, your supplier must do its utmost to maintain the quality of service provided to your company.

Loading Alpha Percentage of variance explained 0.91

18.4

0.83

15.5

0.82

15.0

0.78

13.9

0.92 0.89 0.83

0.85 0.84 0.82

0.88 0.82 0.81

0.87 0.81 0.64


665

Appendix (continued) A (continued) Factor and contents

Loading Alpha Percentage of variance explained

Buyers' procedural asset specificity 1. The outsourcing relationship has entailed significant changes for the overall operations of your company. 2. The outsourcing relationship has entailed significant restructuring and downsizing (e.g. redundancies) in your company. 3. The outsourcing relationship has entailed NO changes for your employees. Total percentage of variance Overall Cronbach's alpha value for buyers'

12.6

75.4 asset specificity

0.82

Factor and contents


Suppliers' site asset specificity 1. Your supplier has relocated the whole or part of its operations for the purpose of being nearer to your company since close proximity is important to the outsourcing relationship. 2. Your supplier has relocated the whole or part of its operations for the sole purpose of the outsourcing relationship with your company and, hence, this relocation has little value outside this relationship. 3. Your supplier has relocated some of its operations or assets in order to improve its services towards your company. 4. The outsourcing relationship requires your supplier to be located near your company. Suppliers' dedicated asset specificity 1. Your supplier has expanded its production capacity in the hope of a long-term relationship with your company. 2. Your supplier has made extra investments in order to expand its production capacity and be able to meet your needs. 3. Should your outsourcing relationship cease, your supplier would be left with substantial unsold output or excess capacity (e.g. extra staff). 4. Your supplier's sales to your company represent an important share of your supplier's total sales. Suppliers' human asset specificity 1. If you were to change your supplier, it would take a long time for a new supplier to serve you as well as the current one. 2. Your supplier has customised its own workflows and routines to the peculiarities of your company. 3. Your supplier has made a high degree of adaptation (e.g. via training) in order to provide the customised service required by your company. 4. Your supplier faced initial difficulties in learning and adapting to your company's way of doing things. Total percentage of variance Overall Cronbach's alpha value for supplier's

0.70 0.81 0.74 0.73

0.91

27.1

0.89

25.6

0.77

20.1

0.94 0.92 0.84 0.83

0.89 0.88 0.87 0.78

0.82 0.80 0.74 0.70 72.8

asset specificity

0.80

Outsourcing relationship performance⁎


1. Your company is very satisfied with the quality of the service received in terms of consistency, timeless and accuracy. 2. Your company is very satisfied with this supplier's responsiveness to problems or queries. 3. Your company is very satisfied with the overall benefits obtained from outsourcing this function. 4. Overall, the objectives set by your company in relation to the outsourcing project have been met. 5. The service level received from this supplier has exceeded your company's expectations. 6. By outsourcing the function your company — via your supplier — has benefited from better access to skilled personnel. 7. By outsourcing the function your company has benefited from higher quality. 8. Outsourcing the function of reference has allowed your company to concentrate own resources (e.g. staff) on core activities. 9. Our company has NOT achieved the target level of cost savings expected by outsourcing this function.

0.90 0.89 0.88 0.85 0.84 0.78 0.77 0.72 0.66

0.93

67

a

The Bartlett's sphericity tests are all significant at p b .001. The Kaiser–Meyer–Olkin (KMO) sample adequacy test statistics are all above 0.75. ⁎ For outsourcing relationship performance, all ‘alpha if item deleted’ values are below or equal to 0.93.

References

Aiken LS, West SG. Multiple regression: testing and interpreting interactions. Thousand Oaks, CA: Sage; 1991. Anderson E. The sales person as outside agent or employee: a transaction cost analysis. Mark Sci 1985;4:234–54. Anderson EL, Coughlan AT. International market entry and expansion via independent or integrated channels of distribution. J Mark 1987;51:71–82. Armstrong JS, Overton TS. Estimating nonresponse bias in mail surveys. J Mark Res 1977;14:396–402. Artz KW. Buyer–supplier performance: the role of asset specificity, reciprocal investments and relational exchange. Br J Manage 1999;10:113–26. Artz KW, Brush T. Asset specificity, uncertainty and relational norms: an examination of coordination costs in collaborative strategic alliances. J Econ Behav Organ 2000;41: 337–62. Barthelemy J, Quelin BV. Competence, specificity and outsourcing: impact on the complexity of the contract. Academy of Management Conference; 2002. Brouthers KD, Brouthers LE. Why service and manufacturing entry mode choices differ: the influence of transaction cost factors, risk and trust. J Manag Stud 2003;40: 1179–204. Brown T, Potoski M. Transaction costs and contracting: the practitioner perspective. Public Perform Manage Rev 2005;28:326–51.

Bucklin LP, Sengupta S. Organizing successful co-marketing alliances. J Mark 1993;57:32–46. Buvik A, Haugland SA. The allocation of specific assets, relationship duration, and contractual coordination in buyer–seller relationships. Scand J Manag 2005;21: 41–60. Buvik A, Reve T. Asymmetrical deployment of specific assets and contractual safeguarding in industrial purchasing relationships. J Bus Res 2001;51:101–13. Chiles TH, McMackin JF. Integrating variable risk preferences, trust, and transaction cost economics. Acad Manage Rev 1996;21:73–99. Cohen J, Cohen P, West SG, Aiken LS. Applied multiple regression/correlation analysis for the behavioral sciences. Mahwah, NJ: Lawrence Erlbaum Associates; 2003. Conner KR, Prahalad CK. A resource-based theory of the firm: knowledge versus opportunism. Organ Sci 1996;7:477–501. David RJ, Han SK. A systematic assessment of the empirical support for transaction cost economics. Strateg Manage J 2004;25:39–58. Dibbern J, Wynne WC, Heinzl A. The impact of human asset specificity on the sourcing of application services. The 13th European Conference of Information Systems: Regensburg (ECIS paper); 2005. http://is.lse.ac.uk/asp/aspecis/20050144.pdf. Dillman D. Mail and telephone surveys: the total design method. New York: Wiley; 1978. Dyer JH. Specialized supplier networks as a source of competitive advantage: evidence from the auto industry. Strateg Manage J 1996;17:271–92. Gatignon H, Anderson E. The multinational corporation's degree of control over foreign subsidiaries: an empirical test of a transaction cost explanation. J Law Econ Organ 1988;4:305–36.

666


Geyskens I, Steenkamp JBEM, Kumar N. Make, buy, or ally: a transaction cost theory meta-analysis. Acad Manage J 2006;49:519–43. Ghani FA, Khan JH. Network relationships and asset specificity in Pakistan automotive industry. J Asia Pac Econ 2004;9:85-100. Goodman PS, Fichman M, Lerch F, Snyder PR. Customer–firm relationships, involvement, and customer satisfaction. Acad Manage J 1995;38:1310–24. Grover V, Cheon MJ, Teng JTC. The effect of service quality and partnership on the outsourcing of information systems functions. J Manage Inf Syst 1996;12:89-116. Hair JF, Anderson RE, Tatham RL, Black WC. Multivariate data analysis. New Jersey: Prentice Hall; 1998. Harman HH. Modern factor analysis. Chicago: University of Chicago Press; 1967. Heide JB. Inter-organizational governance in marketing channels. Theoretical perspectives on forms and antecedents. J Mark 1994;58:71–85. Heide JB, John G. Alliances in industrial purchasing: the determinants of joint action in buyer–supplier relationships. J Mark Res 1990;27:24–36. Heide JB, John G. Do norms matter in marketing relationships? J Mark 1992;56:32–44. Heide JB, Stump RL. Performance implications of buyer–supplier relationships in industrial markets: a transaction cost explanation. J Bus Res 1995;32:57–66. Hill C. Cooperation, opportunism, and the invisible hand: implications for transaction cost theory. Acad Manage Rev 1990;15:500–13. Jaccard JR, Turrisi R, Wan CK. Interaction effects in multiple regression. Newbury Park, CA: Sage; 1990. Joskow PL. Contract duration and relationship-specific investments: empirical evidence from coal markets. Am Econ Rev 1987;77:168–85. Klein B, Leffler KB. The role of market forces in assuring contractual performance. J Polit Econ 1981;89:615–41. Klein B, Crawford R, Alchian A. Vertical integration, appropriable rents, and the competitive contracting process. J Law Econ 1978;21:297–326. Klein S, Frazier GL, Roth VJ. A transaction cost analysis model of channel integration in international markets. J Mark Res 1990;27:196–208. Lacity MC, Willcocks LP, Feeny DF. The value of selective IT sourcing. Sloan Manage Rev 1996;37:13–25. Lamminmaki D. Why do hotels outsource? An investigation using asset specificity. Int J Contemp Hosp Manag 2005;17:516–28. Lazzarini SG, Claro DP, Mesquita LF. Buyer–supplier and supplier–supplier alliances: do they reinforce or undermine one another? J Manag Stud 2008;45:561–84. Leiblein MJ. The choice of organizational governance form and performance: predictions from transaction cost, resourced-based, and real option theories. J Manag 2003;29: 937–61. Leiblein MJ, Reuer JJ, Dalsace F. Do make or buy decisions matter? The influence of governance on technological performance. Strateg Manage J 2002;23:817–33. Lepack DP, Takeuchi R, Snell SA. Employment flexibility and firm performance: examining the interaction effects of employment mode, environmental dynamism, and technological intensity. J Manag 2003;29:681–703. Levy DT. The transaction cost approach to vertical integration: an empirical examination. Rev Econ Stat 1985;67:438–45. Lohtia R, Brooks CM, Krapfel RE. What constitutes a transaction-specific asset: an examination of the dimensions and types. J Bus Res 1994;30:261–70. Lui SS, Wong Y-Y, Liu W. Asset specificity roles in interfirm cooperation: reducing opportunistic behavior or increasing cooperative behavior? Journal of Business Research 2009;62:1214–9. Madhok A. Reassessing the fundamentals and beyond: Ronald Coase, the transaction cost and resource-based theories of the firm and the institutional structure of production. Strateg Manage J 2002;23:535–50. Malhotra MK, Grover V. An assessment of survey research in POM: from constructs to theory. J Oper Manag 1998;16:407–25. Malone TW, Yates J, Benjamin RI. Electronic markets and electronic hierarchies. Commun ACM 1987;30:484–97. Maltz A. Private fleet use: a transaction cost approach. Transp J 1993;32:46–53. Masten SE. The organization of production: evidence from the aerospace industry. J Law Econ 1984;27:403–17. Masten SE, Meehan JW, Snyder EA. Vertical integration in the U.S. automobile industry: a note on the influence of transaction specific assets. J Econ Behav Organ 1989;12:265–73. Masten SE, Meehan JW, Snyder EA. The costs of organization. J Law Econ Organ 1991;7:1-25. Monteverde K, Teece DJ. Supplier switching costs and vertical integration in the automobile industry. Bell J Econ 1982;13:206–13.

Morill C, Morill J. Internal auditors and the external audit: a transaction cost perspective. Manag Audit J 2003;18:490–504. Murray JY. Strategic alliance-based global sourcing strategy for competitive advantage: a conceptual framework and research propositions. J Int Mark 2001;9:30–58. Nishiguchi T. Strategic industrial sourcing: the Japanese advantage. Oxford: Oxford University Press; 1994. PA Consulting Group. IT outsourcing survey. London: PA Consulting Group; 1993. PA Consulting Group. Riding the wave of channel substitution: international strategic sourcing survey. London: PA Consulting Group; 1996. Parkhe A. Strategic alliances structuring: a game theoretic and transaction cost examination of inter-firm co-operation. Acad Manage J 1993;36:794–829. Perez-Nordtvedt L, Kedia BL, Datta DK, Rasheed AA. Effectiveness and efficiency of cross-border knowledge transfer: an empirical examination. J Manag Stud 2008;45: 714–44. Podsakoff PM, Organ DW. Self-reports in organizational research: problems and prospects. J Manag 1986;12:531–44. Podsakoff PM, MacKenzie SB, Lee J-Y, Podsakoff NP. Common method biases in behavioural research: a critical review of the literature and recommended remedies. J Appl Psychol 2003;88:879–903. Poppo L, Zenger T. Testing alternative theories of the firm: transaction cost, knowledgebased, and measurement explanations for make-or-buy decisions in information services. Strateg Manage J 1998;19:853–77. Poppo L, Zenger T. Do formal contracts and relational governance function as substitutes or complements? Strateg Manage J 2002;23:707–25. Powell WW. Neither market nor hierarchy: network forms of organization. Res Organ Behav 1990;12:295–336. Provan KG, Skinner SJ. Interorganizational dependence and control predictors of opportunism in dealer–supplier relations. Acad Manage J 1989;32:202–12. Rodriguez TFE, Padilla AMG. The relationship between leisure outsourcing and specificity: performance and management perception in hotels in the Canary Islands. J Hosp Tour Res 2005;29:396–418. Rokkan A, Heide JB, Wathne KH. Specific investments in marketing relationships: expropriation and bonding effects. J Mark Res 2003;40:210–24. Ruchala LV. Managing and controlling specialized assets, vol. 79. Managing Accounting: Official Magazine of Institute of Management Accountants; 1997. p. 20–4. Saxton T. The effects of partner and relationship characteristics on alliance outcomes. Acad Manage J 1997;40:443–61. Wahrenburg M, Hackethal A, Friedrich L, Gellrich T. Strategic decisions regarding the vertical integration of human resource organizations: evidence from an integrated HR model for the financial services and non-financial services industry in Germany, Austria and Switzerland. Int J Hum Resour Manag 2006;17:1726–71. Wang ET. Transaction attributes and software outsourcing success: an empirical investigation of transaction cost theory. Inf Syst J 2002;12:153–81. Williamson OE. The vertical integration of production: market failure considerations. Am Econ Rev 1971;61:112–23. Williamson OE. Markets and hierarchies: analysis and antitrust implications. New York: Free Press; 1975. Williamson OE. Credible commitments: using hostages to support exchange. Am Econ Rev 1983;73:519–40. Williamson OE. The economic institutions of capitalism. New York: Free Press; 1985. Williamson OE. The logic of economic organization. J Law Econ Organ 1988;4:65–93. Williamson OE. Transaction cost economics. In: Schnalen R, Willing R, editors. Handbook of Industrial Organization. Amsterdam: Elsevier Science; 1989. p. 136–82. Williamson OE. Comparative economic organization: the analysis of discrete structural alternatives. Adm Sci Q 1991;36:269–96. Williamson OE. The mechanisms of governance. Oxford: Oxford University Press; 1996. Williamson OE. Public and private bureaucracies. J Law Econ Organ 1999;15:306–42. Zaheer A, Venkatraman N. Relational governance as an interorganizational strategy: an empirical test of the role of trust in economic exchange. Strateg Manage J 1995;16: 373–92. Zaheer A, McEvily B, Perrone V. Does trust matter? Exploring the effects of interorganizational and interpersonal trust on performance. Organ Sci 1998;9:141–59. Zajac EJ, Olsen CP. From transaction cost to transactional value analysis: implications for the study of interorganizational strategies. J Manag Stud 1993;30:131–45.