Filling a gap in traditional transaction cost economics ...

Int. J. Production Economics 79 (2002) 1}14

Filling a gap in traditional transaction cost economics: Towards transaction bene"ts-based analysis K. Blomqvist , K. KylaK heiko, V.-M. Virolainen* Telecom Business Research Center, Lappeenranta University of Technology, P.O. Box 20, FIN- 53851 Lappeenranta, Finland Department of Business Administration, Lappeenranta University of Technology, P.O. Box 20, FIN-53851 Lappeenranta, Finland Received 12 April 2000; accepted 10 August 2000

Abstract This paper analyzes dyadic partnership formation between asymmetric buyers and specialized suppliers. In the "rst part of the paper di!erent economics of organization-based approaches are evaluated. Their basic implications concerning the rise of partnerships are derived. In the second part a dynamized transaction cost and bene"t model is introduced to analyze the most critical elements of a typical partnership decision. The "nal part is based on insights from practice and in-depth interviews among 12 specialized suppliers and their four large incumbent partners in the Information and Telecommunications Industry. 2002 Elsevier Science B.V. All rights reserved. Keywords: Transaction costs and bene"ts; Theory of the "rm; Partnership; Supply management; Trust

1. Posing the issue According to the founders of transaction cost economics (TCE), Coase [1] and Williamson [2], markets and vertical integration (or hierarchies) are the two main governance structures, out of which a "rm may choose the most e$cient one. Coase did not even mention the intermediate governance structure between markets and hierarchies, called hybrid by Williamson. We call these hybrid governance structures partnerships and interpret them as individual contracts between

* Corresponding author. Tel.: #358-5-6212612; fax: #3585-621-2699. E-mail addresses: kirsimarja.blomqvist@lut." (K. Blomqvist), kalevi.kylaheiko@lut." (K. KylaK heiko), veli-matti.virolainen@ lut." (V.M. Virolainen).

parties. The aim of the contract is, of course, to create the joint surplus through cooperation and share it in a way, which bene"ts both (all) the parties. Rapid changes in business environments are increasingly driving the formation of strategic partnerships between companies in the world economy. Di!erent types of partnerships are a logical and timely response to intense and rapid changes in economic activities, technologies, and globalization of world markets [3]. Partnerships have gained much theoretical interest in strategic literature during the last 10 years. Despite the fact that Coase skipped them altogether modern economics of organization-related approaches have managed to shed light on some factors behind the rise of partnership-based governance structures. A recent stream of resource-based view, the knowledge-based view, [4}6] analyzes organizational

0925-5273/02/$ - see front matter 2002 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 5 - 5 2 7 3 ( 0 0 ) 0 0 0 9 5 - 5

2

K. Blomqvist et al. / Int. J. Production Economics 79 (2002) 1}14

Table 1 Typology of di!erent explanations of inter-"rm partnerships [14,17] Focus

Production & Cognition-based-' Transaction & management benexts Exchange-based-' Transaction & managements costs

Temporal dimension Static

Dynamic

A: Resource-based view based on static capabilities and given cognitive frames [15,38] added with [18] innovation-related transaction costs C: Static Coasean [1] & Williamsonian [19] transaction cost economics

B: Dynamic capability view based on learning and changing cognitive frames [8,5] D: Dynamic transaction cost economics (embryos launched by [20,9,39]

capabilities and knowledge as a source for competitive advantage. The key organizational issue is whether to develop the needed competencies, capabilities and knowledge internally or whether it is rational to exploit and integrate external knowledge. Knowledge-based competition leads us to the classic questions of the modern theories of the "rm concerning the crucial role of organizational boundaries. Unfortunately, these explanations are typically static and focused primarily on cost-based comparisons. Our contribution to this discussion will be the explication of the main sources of the governance bene"ts, which can be obtained through di!erent kinds of inter-"rm contractual arrangements.

2. On partnership explanations We start with a survey of some theoretical results obtained so far in economics-based literature. The rise of partnerships has been studied at least from two di!erent angles. The "rst one can be characterized as the economics of organization-inspired contractual or governance approach, which in its explanations emphasizes either the Coasean transaction costs or the ownership of non-contractible assets (i.e. the property rights school introduced by Grossman and Hart [7]). The alternative, more evolutionarily inspired perspective is based on the dynamic capability or knowledge-based approach [8,9], which, in turn, emphasizes the role of "rm-speci"c, rare, and hard-to-imitate routines, capabilities, learning and socially embedded tacit

knowledge when creating and sustaining joint surpluses through partnerships. For those who want to get acquainted well with the basic premises and di!erences between these two approaches there are many quite recent metatheoretical comparisons available (cf. [10}12]). Here we would like to pick up only one important position which arised from recent discussions, the integrationistic metatheoretical point of view strongly propagated by Foss [11] (cf. also [13]). According to this view, both the seemingly rival approaches can (should) be interpreted as the complements but not as substitutes. This view that we share has important implications as to the analysis of inter-"rm arrangements. Following the integrationist research strategy we take seriously some parts of the knowledge-based criticism of the proponents of the dynamic capability view (e.g. [14}16]). To put it simply, these critics maintain that the governance approach at least partly neglects the production and cognition-related issues, such as genuinely bounded rationality and imperfect and disperse knowledge of agents, collective tacit know-how, radical uncertainty, and overemphasizes the exchange-related issues, such as market (in)e$ciency and incentives. We interpret the tone of this critical message so that the gap between production and exchange has to be bridged by launching some additional explanatory items into the standard TCE framework. Table 1 outlines the main di!erences between the two main approaches mentioned above from the explanatory point of view. This typology is based on


the insightful ideas introduced "rst by Winter [14] and worked out by KylaK heiko and Miettinen [17]. The distinctions used are between (i) staticness and dynamics on the one hand and (ii) production and cognition-based and exchange-based categories on the other hand. In terms of these distinctions we can characterize the seemingly rival but actually complementary domains of theoretical explananda (i.e. the set of explained items) of di!erent approaches. Our main thesis will be that every comprehensive explanation of the emergence of various inter-xrm relationships has to be based on all the aspects represented in Table 1. The largely neglected theoretical bridge between the production and the exchangebased focal points of Table 1 can be built by introducing relevant concepts such as transaction benexts or transaction values (cf. [21]).

3. Towards dynamized and extended governance cost minimizing model Next, we will introduce some basic concepts of the dynamic capability view (cell B in Table 1) in order to make sense of our forthcoming analysis which can be characterized as an attempt to combine all cells A}D and to use them as a common explanans set. We begin with learning mechanisms, the hallmarks of the dynamic capability view. They can best be analyzed in terms of static and dynamic routines introduced by Nelson and Winter [22]. Static routines replicate existing organizational and technological competencies. Through partial replication there is room for adaptational adjustments by learning. Dynamic routines are routines through which the "rm can `learn by learninga and di!use generic scienti"c and engineering knowledge from other "rms. The dynamic capability concept rests on dynamic routines and can be de"ned as `the capacity of a xrm to renew, augment, and adapt its core competencies over timea [23]. The dynamic capability view regards the "rm as an organization, which combines partly tacit and cumulative know-how (`technoa) with generic information (`logya). Heterogeneity of capabilities implies that the boundaries of the "rm have to be interpreted as strategic devices when outlining

3

a "rm's competitive strategy. To put it brie#y, it is a question of how to generate more value by using internally produced/ externally acquired capabilities and resources in the most e$cient way. In our opinion, this message has to be imported into the static Coasean framework in order to make it more comprehensive for coping with the inter-xrm arrangement issues. Fig. 1 summarizes the basic #ows and determinants in our extended and dynamized governance framework. The "rm is viewed as a value chain consisting of many activities. They are all based on partly tacit and partly generic routines/capabilities. Some internal and external capabilities already exist (i.e. they are static), whereas some have to be developed through learning or knowledge transferring or created through knowledge integration (these not yet existing ones are called dynamic). Some activities can be bought from other "rms (i.e. outsourced capabilities), whereas some are based upon internal capabilities. Outsourcing costs are called transaction costs (relating to search, planning, negotiating, monitoring, and enforcement) and the insourcing costs are called management costs (relating to administration, control and monitoring as well as the costs of using low-powered bureaucratic incentives). Following the lead of Coase and Williamson we conclude that the main issue in our model is to "nd out such a governance structure, i.e. a combination of outsourced, networked and insourced transactions which economizes on the sum of production, transaction and management costs at the same time when the surplus value obtained through transaction and management bene"ts is maximized as well. Next, we will launch the main determinants of di!erent cost and bene"t categories shown in Fig. 1. Williamson [2,19] explicated the following determinants that give rise to (static) transaction costs: (i) bounded rationality, (ii) opportunism, (iii) information impactedness, (iv) frequency of transactions, and (v) asset speci"city. His most paradigmatic case was (and is) the so-called ex post hold-up problem caused by the dangerous triad of uncertainty, frequency and asset speci"city. When the assets are very speci"c it means that their value is much lower (or even zero), if they cannot be used in the joint transaction between the partners [10].

4


Fig. 1. Dynamic transaction and management costs and bene"ts.

In such a situation there opens up an opportunity for opportunistic ex post behavior for the non-investing partner, who may threaten to stop the cooperation, unless he/she/they cannot have a greater share of the joint surplus [13]. A paradigmatic solution for this case is vertical integration. Later, Teece [18] introduced such concepts as complementary capabilities and appropriability regime. The former consists of external capabilities needed to complete a "rm's internal capabilities. The complementary external capabilities, which have to be outsourced, a!ect the bargaining situation the more, the more ine$cient are their markets. This implies higher transaction costs and, consequently, more integrated solutions. The appropriability criterion determines how e!ectively a "rm can protect its strategic knowledge from free rider imitators. The tightness of this regime depends upon legal protection and tacitness. The more tacit knowledge is, the lower are transaction costs and vice versa. Next, we have to introduce dynamics into the framework. The dynamic TC problem can be for-

mulated as follows: the "rm has to decide, whether it is more e$cient (i) to generate new and develop old internal capabilities through continuous learning and R&D investments (à conglomerate strategya) or (ii) to acquire external capabilities from the market (à hollow "rm strategya) or (iii) to exploit economies of scale and scope through networking (à partnership strategya). Henceforward, the costs of transferring capabilities over the "rm's boundaries are called dynamic governance costs. They can further be divided into dynamic transaction costs (i.e. persuading, negotiating and teaching with the providers of external capabilities) and dynamic management costs, which consist of the costs of persuading, negotiating and teaching within the "rm of own when trying to create/develop a capability internally or persuading, negotiating and teaching external partners when a "rm-made activity is tried to sell (cf. [20]). As an example of how to use these dynamic concepts we can take the "rm operating at the turbulent emerging phase of a new technological


trajectory (e.g. telecommunications) where the role of tacit knowledge is high. This means that the appropriability regime is tighter than in the case of well-established mature trajectory (e.g. the paper and pulp industry). Following this logic, one can predict a tendency towards vertically more integrated inter-"rm arrangements as the technological trajectory becomes more stabilized because of decreased technological uncertainty and increased standardization. Now, it is time to go brie#y into the governance benexts problematics (more about this in Section 5). A bit paradoxically traditional TCE literature totally neglects the issue of transaction, management or partnership benexts. This means that they implicitly assume that these bene"ts are somehow independent of governance structures. Of course, this argument is not valid. There are always bene"ts, which basically depend on the very nature of the governance structure. Fig. 1 allows us now to introduce our "rst characterizations concerning them. When the "rm utilizes its own resources and capabilities it can e!ectively build on cumulative learning and exploit the economies of scope through learning. It can also utilize competenceenhancing innovations or exploit monopoly power over other "rms. They all are benexts related to the vertical integration strategy. We call them management or xrm-internal benexts. On the other hand, when the "rm uses the market option it can exploit high-power incentives through "erce competition, economies of scale through specialization, and utilize #exibility and variation generated through many alternative partners operating in open markets. From the property rights perspective this originally evolutionary view can be defended by referring to the fact that the market-related bene"ts arise basically, since `markets are identixed with the right to bargain and, when necessary, to exit with the assets owned. This... provides entrepreneurial incentivesa [10]. Through the market a "rm can also better cope with radical uncertainty and competence-destroying innovations. We call these market-related mechanisms transaction benexts. The usual distinction between staticness and dynamics is not necessary here, since bene"ts typically are generated only through dynamic processes. Hence, they are called dynamic.

5

In order to clarify all the transaction and management-related categories we extend and modify Dietrich's [24] approach, which is based on two crucial distinctions. The "rst one is between static transaction costs (Cm) and static management costs (Cf), where C denotes `costsa and `m, fa denote `marketa and `"rm's internal organizationa, respectively. The second one deals with the distinction between dynamic transaction benexts (Bm) and dynamic management benexts (Bf). Our own extensions dynamic transaction and management costs can be di!erentiated by using the asterisks (*:s). Our dynamized governance cost formulations are as follows: (1) Use pro-market option iw Bm!Cm!Cm*'Bf!Cf!Cf* or Bm!Bf'Cm#Cm*!Cf!Cf* (2) Use pro-integration option iw Bm!Cm!Cm*(Bf!Cf!Cf* or Bm!Bf(Cm#Cm*!Cf!Cf*, where Bm* dynamic transaction benexts, which are positively correlated with the ability to exploit economies of scale, high-powered incentives and #exibility. They also allow to generate more ideas through the market variation and to cope with competence-destroying innovations. Cm static transaction costs, which are positively correlated with opportunism, few partners available, asset or capability speci"city, inability to cope with parametric uncertainty, dependence on complementary assets holders, systemic nature of innovation, low appropriability, and organizational inertia against newcomers. Cm* dynamic transaction costs related to persuasion and learning costs with the providers of outsourced external capabilities. Bf* dynamic management bene"ts, which are positively correlated with the ability to exploit monopoly power, asymmetric knowledge, economies of scope, and cumulative tacit know-how when facing competence-enhancing innovations.

6


Cf

static management costs, which are related to the costs of monitoring large bureaucracy and high sunken R&D costs. Cf* dynamic management costs relate to the costs of persuading, negotiating and teaching within the "rm when a new capability has to be generated. It also includes the inability to cope with radical uncertainty. Now, we are ready to use these implicit equations (1) and (2) above when explaining the rise of inter-"rm partnerships in light of our approachrelated explanatory categorization introduced in Table 1. The cost types Cm and Cf belong no doubt to the domain of the Coasean}Williamsonian basic model with some Teecean extras (i.e. cell C#some elements of A), Cm* and Cf* belong to the domain of until now relatively ill-developed dynamic transaction cost economics (i.e. cell D), whereas, the dynamic bene"t functions Bm* and Bf* represent our own contribution i.e. they cover a combination of cells A and B.

4. Paradigmatic explanations for emergence of partnerships In this section we shall put together our explanatory categories (A, B, C, D) and our explanation functions (1 and 2) and introduce in terms of them some interesting paradigmatic explanations for the emergence of inter-"rm partnership relationships. First we will take cell C, which combines the exchange-based view with static temporal isolation, i.e. represents the most paradigmatic Coasean} Williamsonian transaction cost case. In this very aggregative framework, which is actually built for the analysis of the basic choice between the markets and hierarchies, the rise of di!erent governance options can be interpreted as follows. Vertical integration is the best option, when (i) uncertainty, the danger of opportunism and complexity are high, (ii) asset speci"city is high and there are only few providers of complementary capabilities, and (iii) trust between partners is lacking. The market option is preferred when (i) the degrees of uncertainty and complexity are minor and the danger of opportunistic behavior is small,

(ii) there are many potential partners available, and (iii) transactions do not need any speci"c investments. In this static TCE-oriented interpretation the partnership solutions can best be regarded as strange hybrids between well-de"ned market and hierarchies. Intermediate governance structures are most preferable when there are determinants, which simultaneously speak both for insourcing (e.g. uncertainty, danger of opportunism, high asset speci"city) and for outsourcing (the need for highpowered incentives). A typical precondition for the emergence of networks is also the pursuit of economies of scale and scope at the same time, even if this dynamic extension can be imported only as an ad hoc relaxation into the static framework. These two important concepts actually belong to the domain of dynamic transaction/management cost, i.e. to cell D. In addition, trust and reciprocity among partners are badly needed to impede opportunism. The next step is to go to cell A, where Teecean transaction cost elements can be taken into account as well. Main additional elements here are the nature of innovation (systemic vs. autonomous), the role of complementary assets and the tightness of the appropriability regime. We can conclude that the partnership agreements will be looser (tighter), i.e. closer to market (vertical integration) solutions, if (i) innovation launched is autonomous (systemic) by nature and requires small (large) speci"c investment, (ii) appropriability of new knowledge is tight (weak), thus implying no (great) danger of free rider imitators, and (iii) the markets of complementary capabilities are e$cient (ine$cient), thus fostering the pressures towards hold-up problems. Fig. 2 illustrates the basic governance choice problem in terms of a `governance costa indicator or pendulum. The higher the management costs and the lower the transaction costs are, the more preferable is the market option and vice versa. Fig. 2 also shows how the boundaries of the "rm change subject to governance costs. Although all the results derived above grasp some important issues as to the rise of inter-"rm partnership solutions, they also leave some crucial elements out of the explanans (i.e. the set of explainers) and, consequently, out of the explanandum (i.e. the set of explained items) as well. Most of these omitted explainers can best be dealt with in


7

Fig. 2. Dynamic transaction and management cost-based choice illustration.

terms of static and dynamic governance bene"ts and their underlying mechanisms, i.e. by taking a step towards the union of cells A&B.

5. On partnership-related bene5ts In this section we shall analyze how transaction and management bene"ts may give rise to inter-"rm partnerships. The necessary (but not suf"cient) reason for a partnership is that the aggregate level of quasi-rents derived from the joint production is higher than would be the sum of independent investments done by the parties alone. Let us start with a set-up where we have only two parties, say an incumbent buyer (B) and a small seller (S). Aggregate quasi-rents (QR) can now be measured as QR"
partnership is that the expected net joint surplus is positive, i.e.