. Joint Venture
. Joint Venture
Abstract and Keywords
This chapter introduces and develops a set of principles for analysing the development of collaborative activity. These principles are useful in helping to develop an efficiency rationale for the evolution of collaborative activities. The chapter focuses on the form of cooperative activity characterized by joint ventures. It begins by examining the nature of collaboration that sets the agenda for what an analysis of joint-venture activity must achieve. Market arrangements, hierarchical arrangements, and joint-venture arrangements to coordinate complementary assets are also examined. The chapter furthermore examines how the evolution of sectors may create the essential preconditions for the development of joint ventures. This analysis is then compared with actual patterns of joint venture and other collaborative activity.
The proliferation of cooperative activity between firms in recent years has been widely noted and commented on. First, however, another puzzle. If cooperation is such a good thing, as so many writers in this field appear to argue, why did firms generally wait until relatively recently before pursuing such activities so enthusiastically? We shall be concerned with this fundamental question in this chapter, as well as a number of related issues.
There have been a number of the reasons for cooperative or collaborative activity cited in the managerial literature. For the particular case of joint venture (a jointly owned subsidiary between two or more firms), Harrigan (1988a) lists 28 different motivations which can generally be reduced to benefits from sharing resources (e.g. finance, plant, distribution, technology, information, management practices) or demand side considerations (e.g. reducing competition). Similarly, Contractor and Lorange (1988a, pp. 9–10) argue that joint ventures, licensing, and other types of cooperative agreements can achieve at least seven or more objectives: (a) risk reduction; (b) economies of scale and/or rationalization; (c) technology exchanges; (d) co-opting or blocking competition; (e) overcoming government mandated trade or investment barriers; (f) facilitating initial expansion of inexperienced firms; and (g) complementary contributions of partners. Further, Hergert and Morris (1988) identify some recurring themes behind international collaborative activity as (a) capital requirements; (b) excess capacity; (c) modular design facilitating splitting of responsibilities; (d) economies of scale; (e) emergence of global products; (f) risk reduction; (g) foreign market entry.1 On a rather different tack, Ring and Van de Ven (1992) explore the implications of risk and reliance on trust for the formation of cooperative relationships.
One problem with studies of cooperative activity between firms (including joint venture) is that if there is one area which has been particularly marked by the problem of limited comparators, it is this one. All too often, analysis of reasons for conducting a specific type of activity implicitly or explicitly relate it to a do-nothing alternative. Indeed, it is often difficult to discern significant differences between checklists produced by different analysts to account for mergers, joint ventures, strategic alliances, and contractual alternatives such as licensing. There should be no surprise that checklists for various strategic options often display a strong family resemblance with key phrases such as ‘market access’, ‘R&D economies’, ‘reduction in competition’ recurring from list to list. Rather, there should be (p.178) surprise if the respective options did not display strong similarities. The reason for this is that strategic options are not typically different ways of doing different things, as the drawing up of checklists often implies. Instead, they tend to be different ways (modes) of doing the same thing (direction).
In this chapter we shall introduce and develop a set of principles for analysing the development of collaborative activity. It will be suggested that these principles are useful in helping to develop an efficiency rationale for the evolution of collaborative activities and that they are also consistent with a broad span of empirical evidence on these issues. It is argued that cooperative activity may impact at business, corporate, and group level. The key to understanding ‘the complex and inter-locking clusters, groups and alliances which represent co-operation fully and formally developed’ (Richardson, 1972, p. 887), lies in recognizing the implications of the interplay between multiple levels of analysis. Our primary focus will be on the form of cooperative activity characterized by joint ventures, though some of the analysis will generalize to other forms of cooperative agreements.
The basic ideas here have their genesis in an earlier paper by Kay et al. (1987) and Kay (1992b) and is also consistent with points developed separately by Hennart (1988a) and Buckley and Casson (1988). Kay et al (1987) pointed out that merger between two large diversified firms to exploit a limited venture opportunity could be akin to using a sledgehammer to crack a nut. In a similar vein, Hennart comments, ‘Besides the obvious case when governments restrict mergers and acquisitions, joint venture will be preferred when the assets that yield the desired services are a small and inseparable part of the total assets held by both potential partners or when a merger or a total acquisition would significantly increase management costs’ (Hennart, 1991, p. 99). Buckley and Casson (1988, p. 41) suggest that one reason why partners may enter into a joint venture rather than merge may be managerial diseconomies arising from the scale and diversity of the resulting enterprise. Kogut (1988, p. 176) suggests that it is reasonable to assume that acquisition may be ruled out as an option if the technology being transferred is only a small part of the total value of the firm, and Lewis (1990, p. 16) argues similarly that acquisitions are logical only if the value of the resources sought is a large part of the purchase.
This captures the essence of the case for joint venture, and this chapter develops a framework in which points such as these are set in context. It shows that analysis is likely to be flawed when based on the assumption that joint ventures can be low cost over the domain of the joint venture itself, or based on the assumption that joint ventures can generate gains that otherwise cannot be achieved by other options. It is argued that an efficiency rationale for joint venture must show that it can obtain the same gains as other options but at lower cost subject to governmental restraints such as in anti-trust or market access areas. Such a rationale must simultaneously reconcile this with a recognition that organizing ventures through split ownership (p.179) is itself typically unambiguously more costly and less efficient than single ownership options. From this is developed the analysis of collaborative activity that builds into an analysis of alliances in the next chapter. A number of elements in the chapter may be seen as following from, or consistent with, Richardson’s 1972 paper, including his ‘triple distinction’ (p. 896) between firm, cooperative and market modes of resource coordination; the role of complementary activities in collaboration; the role of future decision in stimulating cooperative arrangements; and the picture of the firm as a bundle of capabilities represented by appropriate knowledge, experience, and skills.
In section 9.1 below, we make some initial points regarding the nature of collaboration that sets the agenda for what an analysis of joint venture activity must achieve. Market arrangements, hierarchical arrangements, and joint venture arrangements to coordinate complementary assets are looked at in sections 9.2, 9.3, and 9.4 respectively. section 9.5 examines how the evolution of sectors may create the essential preconditions for the development of joint ventures. Section 9.6 compares the analysis of this section with actual patterns of joint venture and other collaborative activity.
9.1. Collaborative Activity
The role of complementary activities or assets has been recognized and analysed as a major stimulus to collaborative activity (Richardson, 1972; Teece, 1986a). Richardson and Teece argue that to bring complex projects to fruition, a variety of skills or resources may be required, not all of which may be possessed to the requisite degree or quality by one firm. Collaborative activity between firms providing different but complementary assets is one way that such resources can be combined to produce efficiency oriented economic activity. The significance of complementary assets in collaborative activity is clear, with numerous examples provided by Teece and other writers to support this point. However, there are points following from the arguments of the previous section that are required to set it in context.
To start with, complementary assets refers only to a sub-category of advantages from resource interaction. Complementary assets depend on heterogeneity of firms, but both market power and scale economies from resource sharing depend on some degree of homogeneity of interest or resources across firms. For example, Dunning (1993, pp. 256–7) surveys studies on cross-border cartels, and concludes that this form of collaborative activity is most likely to be successful where products are homogeneous and the ownership specific advantages of firms are similar. Also, on the supply side, many joint ventures in the extractive sector have traditionally been characterized by the desires of firms to share capital or spread risk, while the development of R&D clubs or consortia has been generally stimulated (p.180) by the collective motive of replicating the exploitation of R&D economies that otherwise could only be achieved by a prohibitively large and diversified firm. The key question here is ‘are differences between partners essential for this form of collaborative activity to exist?’ If the answer is no, then the objectives of the collaboration are more likely to be found in market power or scale reasons, rather than complementary assets. One important issue that relates to these points of clarification is that it helps confirm that the different types of competitive advantage that may be generated by collaboration is a simple one-to-one mapping of the types of competitive advantage that may be generated by merger. Collaboration can fashion market power, economies of scale and economies of scope from the same basic ingredients as merger activity. It can exploit synergistic as well as complementary relations.
Thus, consistent with the points made in the last section, the important issue is not what collaboration can do. Collaboration can do more than just exploit complementary assets, but even the extended list is a simple replication of sources of competitive advantage that may be exploited by alternative modes such as merger or internal expansion. Figure 9.1 builds on these points to help provide a focus for analysis of collaborative activity. The circles in Fig. 9.1 represent IBUs, and as before, each unit, considered as an independent unit, is assumed to be identical to all others in terms of profitability, growth, and scale. This simplifying assumption allows us to suppress product-market considerations here and concentrate instead on the implications of business unit linkages along the lines opened up in Chapter 4. We shall also focus here on efficiency oriented economies from sharing and leave issues of competition on one side.
The top left of Fig. 9.1 illustrates the reference case of stand-alone or abstention modes, with three single unit firms engaged in head to head competition (C). No linkages in the form of economies from shared resources exist between the firms. Arranged along the top are two alternative options. In (S), one unit exploits complementary assets provided by the two other units. One link is market related and could represent, for example, shared assets in marketing, distribution, advertising, and/or reputation. The other link is technological and could represent, for example, shared assets in plant, equipment, work force, and/or R&D. Given that all business units are presumed equal in value terms when operated as independent units, any market or technological synergies or economies of scope exploited in arrangement (S) means that it is efficiency and value enhancing compared to (C). However, the same argument holds for (E). Here the respective units share identical or similar resources. If it is a collaborative arrangement, it could be a farmers’ cooperative, oil companies co-developing an oil field, or electronic companies participating in an R&D club. As with (S), the value of the combination when a resource link is coordinated exceeds the value of the three-unit competitive scenario in (C). Berg et al. (1982, pp. 21–2) point out that joint ventures to share scale, finance, and risk are a common feature of the mining and metals industries.
(p.181) Both complementary and substitutable asset cases may be exploited by a variety of options and institutional arrangements. However, in Fig. 9.1 we concentrate on the case of complementary assets for illustrative purposes. The dotted circles or ellipses represent corporate boundaries. Figure 9.1 illustrates three possible options to exploit the (S) case of complementary assets, S1, S2, and S3. The collaborative option in this case (S2) is a joint venture with overlapping corporate boundaries thrown around the middle business unit indicating co-ownership and co-direction of this unit by the two firms associated with the other respective units.
A myopic checklist approach to the collaborative option might compare arrangement (C) with combination (S2) and conclude that a rationale for (S2) would be that it beats (C) on value enhancing grounds. However, not only are there varieties of collaborative options but there are also alternatives to collaboration in the form of co-ownership. For example, combination (S1) represents a full-scale merger with corporate boundaries thrown round all three linked business units, while S3 represents an extreme case of market organization in which all linkages are exploited by market contracts. As Richardson (1972) points out ‘co-ordination (of complementary (p.182) activities) can be effected in three ways; by direction, by co-operation, or through market transactions’ (p. 890). If merger, joint venture, and systems of market contracts are all available to exploit the gains from complementary assets in case (S), choice of mode can focus on the comparative advantage of the respective options. This leads to some interesting conclusions on the comparative merits of the respective modes displayed in Fig. 9.1.
Figure 9.1 displays some likely hierarchical arrangements for the respective options with dotted ellipses again indicating the boundaries of firms. The contractual arrangements for exploitation of complementary assets do not involve any additional hierarchical arrangements over those that exist already within business units. The merger option here is presumed to create a multidivisional or M-form corporation (Williamson, 1975) built around the three units responding to a single HQ. The joint venture option leads to with joint ownership of one division as indicated by the overlapping ellipses. The hierarchical arrangement for the joint venture may look rather strange, indeed awkward. It is not usual to find a two-division M-form corporation, and, in fact, small, relatively specialized firms as in this example might be more likely to continue to operate functionally organized unitary form or U-form structure (Williamson, 1975). The strange representation is a consequence of two related issues; it is true that joint ventures typically tend to be coordinated between divisionalized business units operating with an M-form corporation, but, as pointed out above, usually at least one (and often both) firms are large diversified systems2 in which the domain of specific collaboration of the type described in Fig. 9.1 above is limited to a fairly localized region of the overall firm. This means that Fig. 9.1 provides a partial or incomplete picture of corporate systems involved in collaboration.
As we shall argue below, it is the fact that representation of collaborating firms as in Fig. 9.1 tends to be incomplete that holds the key to understanding the evolution of joint ventures. However, consistent with most conventional treatments of this subject, we shall first look at the efficiency and cost implications of joint ventures by considering the administrative and contractual issues associated with units directly involved in coordinating complementary assets. Since merger and joint venture both involve the creation of hierarchy to coordinate complementary assets, we shall start by considering why a hierarchical option may be preferred to a contractual solution, and then assess the relative efficiency implications of joint venture versus the merger alternative. Both parent companies now operate two divisions, their original IBU and the new (joint shared) venture.
9.2. Market Arrangements to Coordinate Complementary Assets
The issue of whether a contractually based solution will tend to be preferred to a hierarchical (merger or joint venture) solution is likely to depend on (p.183) the nature of the assets being transferred. Contractually based coordination tends to come into play when the services provided by the assets being coordinated tend to be fairly standardized and well specified. Complexity of assets is not likely to be an issue as long as the other conditions are satisfied. Expensive, complex and technologically indivisible assets such as computers, satellites, and space shuttles can all be simultaneously exploited by multiple users using variants on standard market contracts. Where contractual solutions tend to encounter problems is more commonly in situations involving ambiguity, uncertainty, and novelty, as in the case of technological innovation and market entry.
Thus, contractually based coordination of complementary assets is most likely to be adopted where we observe well established and stable technologies conducive to standard, off the shelf solutions such as licensing, rental, leasing, and franchising (as we saw in Chapter 3). Some researchers have fused asset complexity with contractual uncertainty, but in fact the concepts are quite distinct. Arnold Schwarzenegger was able to hire Harrier jets from the Pentagon for US$1600 an hour, but the Post-it concept for adhesive notes was invented, developed, produced, and marketed by personnel working in-house with the 3M corporate hierarchy.3 Few would dispute that Post-it technology is intrinsically simpler than jet fighter technology, yet, like most advanced high technology corporations, 3M relies mostly on internal R&D for its innovation search rather than contracting out. The hierarchical imperative ‘find new products’ is fundamentally much more problematic than the contract ‘lease existing technology’. The critical point being made here is that when elements that are the subject of the contract are innovative, then these transaction costs are likely to be especially marked and indeed the contractual option may breakdown completely and fail to coordinate the potential gains from complementary assets. As we argued earlier, contractual arrangements work best when there are few decisions left to be made concerning the cross coordination of complementary assets between business units that cannot be resolved by reference to the terms and conditions of the respective contracts themselves.
The tendency for coordination costs to fall as uncertainty and incompleteness of specification are removed from contractual situations may help to explain why a contract may cope well with stable, established technologies. However, it does not explain why contractual arrangements may cope better than hierarchical arrangements such as merger and joint venture in such circumstances. As discussed above, anything a contractual arrangement can do, in principle a hierarchical arrangement can do also. Only if contractual arrangements involved lower coordination costs than hierarchical alternatives would we then have an efficiency rationale for coordination by contract in such cases.
In fact, it is not difficult to develop such a rationale, since it is no more than the standard defence and justification for the market system. If (p.184) hierarchical alternatives are chosen to coordinate complementary assets, then the range of assets to be drawn from is limited by the boundaries of the hierarchies involved, whether single firm or joint venture. By way of contrast, contract opens up the possibility of matching assets to their highest valued uses as we saw in Chapter 3, not just to the constrained uses to which they may be put within a specified hierarchy. Also, as Williamson points out (1985, Chapter 6), hierarchies may have intrinsic bureaucratic and incentive disadvantages compared to contract, including a possible tendency to more easily ‘forgive’ poor performance that would not be tolerated in a market environment. This is the realm of privatization, contracting out, outsourcing, and downsizing arguments that has been well trodden by management theorists and politicians alike since the 1980s. Consistent with such arguments, if contracts can be well specified then we would expect to find no justification for the evolution of hierarchy to coordinate the exploitation of complementary assets, as we saw in Chapter 3. Commodities such as base or precious metals tend to be standardized and well specified and so their exploitation tends to be coordinated by systems of market contracts.
9.3. Hierarchical Arrangements to Coordinate Complementary Assets
The costs and disadvantages inherent in hierarchical arrangements do not disappear as more innovative and incompletely specified elements are introduced into asset coordination between business units. Indeed, control problems and administrative costs are likely to increase as senior management find themselves less experienced and less familiar with the tasks being carried out by lower level personnel, and consequently less able to police and coordinate activities. The only argument in favour of hierarchy is that the effect of introducing incomplete specifications into market contract alternatives is typically much worse.
As we observed in Chapter 3, hierarchy begins to exert absolute advantage over market coordination alternatives where significant decisions still remain to be made in the future concerning the coordination of assets between business units. This is likely to be the situation where the relevant activity involves innovative elements, and where specifications of particular assets are poorly developed, or as yet absent. The coordination of complementary assets cannot then be adequately regulated by a pre-set contractual recipe. Where contractual arrangements rely heavily on the important resource relocation terms and conditions being decided in the present, hierarchy allows for such situations involving incompletely specified assets by creating decision-making capacity to make strategic future decisions. Such capacity can help to both generate and clarify asset specifications, and coordinate the exploitation of assets across business units in the future once their (p.185) characteristics are more fully known. Interestingly, both market contracts and hierarchy rely on repetition to facilitate coordination, though in very different ways. Market contracts frequently rely on standardization and homogeneity of goods to specify closely what is to be provided by particular goods and services, while hierarchy often uses continuity and repetitious employment of individual managers to make overall assessments of their performance. It may be possible for individual managers to conceal poor performance or opportunism on a one-off basis, but more difficult to do so on a continuing basis.
Thus, the evolution of single firm or joint venture hierarchy as in Fig. 9.1 is a recognition that significant strategic decisions remain to be made concerning cross unit asset coordination. This follows from the argument of Chapter 3 that the need to defer future strategic commitments in cases where assets are incompletely specified is the fundamental reason for the evolution of hierarchy. Unless provision is made to renegotiate contracts, market solutions tend to compress strategic decision-making into present commitments to future action. However, it is also worth noting that hierarchy facilitates the protection of property rights by giving the asset holder more direct control over their use and development. This has been argued as being particularly important in the case of intellectual property rights where legal safeguards to protect licensing agreements (e.g. patent law) may prove difficult or too expensive to invoke. Doing it all yourself and/or keeping developments secret can be an attractive option where technology could leak to other firms.
Such arguments may be plausible in the context of the single firm hierarchical solution in Fig. 9.1, but it leaves us with a puzzle. If hierarchy is a means of retaining control, why do firms create joint venture hierarchies where the method of coordination involves the dual system of hierarchical control illustrated in Fig. 9.1? In the next section we shall explore more fully the difficulties associated with the joint venture option.
9.4. Joint Venture to Coordinate Complementary Assets
In Fig. 9.1, there are two hierarchical alternatives available to enable the coordination of complementary assets—the single firm solution (S1) and the joint venture solution (S2). Again, it bears repeating that both forms can achieve the same gains from facilitating the sharing of resources across business units. Accordingly, an efficiency explanation for the evolution and choice of joint ventures must focus on its associated costs relative to the single firm option.
These relative costs may be analysed with the help of the hierarchical arrangements illustrated in Fig. 9.1 for the respective merger (S1) and joint venture (S2) modes. Both the merger and joint venture cases here involve (p.186) the operation of three business units controlled and coordinated by higher level HQ decision-making. However, there are three significant differences between joint venture operation and merger, each of which may have a bearing on relative costs: (a) dual control by two HQs in joint venture (the middle ‘V’ portions of the joint venture hierarchy in Fig. 9.1); (b) contractual aspects of joint venture (the double headed arrow in the joint venture hierarchy in Fig. 9.1); (c) overlapping ownership in joint ventures (the area of overlap between the dotted ellipses spanning the joint venture hierarchy in Fig. 9.1). We shall discuss each in turn; none of the points are contentious, since the elements of each are already well recognized in the literature as confirmed by Hennart’s (1991) arguments concerning possible cheating, divergent goals and control problems in joint venture.
1. Dual control Being the servant of two masters is not a problem if the masters share the same objectives, the same strategies regarding the attainment of these objectives, and share common views regarding the nature and timing of actions to pursue these strategies. Unfortunately, such happy and ideal coincidence of interests is unlikely to be observed in practice, and the joint venture literature provides much evidence of the existence and consequences of differing objectives and strategies of partners.4 If procedures for conflict resolution do not exist or do not work adequately, costs are imposed on the joint venture operation. However, even if procedures for conflict resolution do work well, they will inevitably impose costs on the joint venture by involving the commitment of managerial time and resources. Consequently, dual control will tend to add costs to joint venture operation compared to the merger or single firm option (Madhok, 1995, p. 126). Kogut (1988) reviews a number of studies in which parental conflict led to instability in the joint venture itself. Also, after interviewing managers involved in thirty-five joint ventures in North America and Western Europe, Killing concludes that because of dual control, ‘if one asks any manager, North American or European, which is more difficult to manage, a wholly owned subsidiary or a joint venture, the chances are high that the joint venture will get the nod’ (Killing, 1983, p. 8).
2. Contractual aspects These exist in joint venture where no corresponding regime of contracts exist in single firm operation. There may be substantial transaction costs following from managerial and legal resources required to search for a partner, negotiate the contract and police the agreement. Berg et al (1982, p. 42) report that one steel joint venture with multiple partners required documents nineteen inches thick to spell out the rights and obligations of all the parents. Since such costs follow from the splitting of ownership rights that are otherwise consolidated in single firm operation, these must also be counted as an additional cost of joint venture relative to the merger option illustrated in Fig. 9.1.
3. Overlapping ownership The contractual aspects above are likely to involve direct resource costs and associated efficiency losses. However, there (p.187) is a further contractual issue that may or may not involve efficiency, but is likely to be perceived as an uncompensated cost by at least one partner and accordingly discourage or impede joint venture formation. This is the issue of appropriability problems regarding the use of intangible assets such as intellectual property and managerial know-how. The problem here is that partners may exploit such intangible assets in areas not covered by the joint venture contract without compensating their partners, and might even use the know-how obtained through the joint venture to become a direct rival to its existing partner in future strategic moves (Richardson, 1972). From an economic perspective, such diffusion may be efficient and desirable, since knowledge tends to have public good characteristics that should generally encourage its dissemination to any users for whom it has value.5 However, such considerations should also encourage parents to restrict or slow transfers of sensitive technological information to the child, which should in turn impair the performance of the joint venture relative to that of wholly owned alternatives.
These three issues—dual control, contractual aspects, and overlapping ownership (appropriability problems)—are interrelated and partially interchangeable in normal conditions. For example, joint venture partners may commit a great deal of resources to establishing and policing the contract in attempts to mitigate possible dual control and appropriability problems. Correspondingly, firms may skimp on contract costs and trade them off for possible dual control and appropriability problems.
However, what we are left with is a remarkably clear and unambiguous conclusion. The implications of dual control, contractual aspects, and appropriability problems all point to joint venture being a less efficient method for coordinating complementary assets than is single firm operation as in the merger case in Fig. 9.1. The existence of one set of problems would have been sufficient to allow this conclusion to be drawn; the existence of all three merely serves strongly to reinforce the conclusion. If joint venture as described in Fig. 9.1 could offer some category of efficiency gains relative to single firm alternatives, we would have a basis for comparing the relative merits of joint venture versus unified hierarchy in different circumstances, and predicting conditions which would encourage the evolution of joint venture in preference to single firm solutions. However, the consistency with which joint venture appears to be an inefficient solution compared to single firm operation closes off such options. Although Fig. 9.1 represents a standard type of joint venture arrangement, it seems to signal that there is no obvious efficiency rationale for the evolution of joint venture activity. On this basis, single firm operation would typically be preferred to joint venture coordination of complementary assets. This is consistent with much of the organizational and managerial literature. For example, from interviews of executives involved in joint ventures in the chemicals industry, Berg et al found that ‘joint ventures were often viewed as a last resort, dominated by (p.188) other forms of intercorporate activity such as mergers’ (Berg et al, 1982, p. 39).
The discussion here also suggests that it may be extremely misleading to think of joint venture as a mixed or hybrid mode which occupies an intermediate position between hierarchy and contract on some hypothetical spectrum, drawing elements from both polar modes. However, this has become the standard transaction cost interpretation of cooperative activity in general, and joint venture activity in particular (Zaheer and Venkatraman, 1995). For example, Williamson (1985, pp. 83–4) identifies a dimension arrayed in terms of the degree to which parties to deals maintain autonomy, with discrete transactions located at one extreme, centralized hierarchical transactions at the other, and hybrid transactions (including joint ventures) located somewhere in between. He also argues that
roughly speaking, transaction cost economics predicts that there will be a shift out of markets (which have the stronger incentive intensity properties) into hierarchies (which feature adaptability) as the condition of asset specificity builds up. Hybrid modes of organization (joint ventures, franchising, regulation, various form of long term contracting), are interpreted as governance structures for which an incentive intensity/adaptability compromise has been reached (Williamson, 1990, p. 68).
The idea of cooperative agreements such as joint ventures being a ‘hybrid’ or a ‘compromise’ appears to involve some idea of there being a trade-off in terms of market/hierarchy characteristics in this perspective. Also, Lorange and Roos (1993, p. 3) begin their discussion of strategic alliances by identifying joint venture activity as lying midway on a spectrum of degree of vertical integration between the polar extremes of free market activity at one end, and hierarchy (and mergers and acquisitions) at the other.
The concept of hybrid implies a selective fusion of elements from different sources, as in the case of plant hybrids, with the implication that hybrids might trade-off some features of one mode in exchange for aspects of another. However, the idea that joint ventures occupy some intermediate point on some organizational form spectrum is misplaced. An analogy might be drawn with someone carrying either a 1kg bag of sugar or a 3kg bag of flour. If he or she carries both, the burden is 4kg, not 2kg; the weight is augmented, not averaged. Similarly, joint venture carries the burden of both hierarchical and contractual arrangements, and indeed the respective market or hierarchy costs may be greater than in the corresponding pure form; for example, it is difficult to see how the complicated dual control hierarchy of joint venture could be more efficient than the simpler conventional structures associated with the wholly owned alternative. We are dealing with a phenomenon that lies outside conventional market and hierarchy modes, not some simple fusion of them.
Of course, just because we cannot find an efficiency rationale for joint venture around the circumstances represented in Fig. 9.1 does not mean (p.189) that it is impossible to find such a rationale. We may have missed something, or could have constructed a straw man to justify our conclusions. However, whilst our way of constructing the analysis may be novel, the elements of our argument are not. The arguments concerning dual control, contractual, and appropriability problems are well documented in the literature. If there are serious objections to our arguments, it is more likely to come from outside the consensus of research findings in this area. Alternatively, it could be argued that efficiency rationales may not be necessary or appropriate to explain the evolution of joint ventures. For example, some theorists have argued that power rather than efficiency objectives may influence the development of the corporation.6 However, it is difficult to see how such arguments may be easily applied to cases of joint venture activity being preferred to single firm operation, since it performs badly on power as well as efficiency criteria. Control is diffused and often scrambled in joint venture arrangements compared to single firm options.
Yet it is not difficult to find examples of joint venture activity where it is adopted, even where it is appears less efficient than the single firm option. The clearest examples are cases where Third World governments require joint venture with a local firm as a prerequisite for market access. Here the inward investor faces Hobson’s choice if it wishes access; the choice is restricted to (C) versus (S2), with (S1) taken out of consideration. Additionally, inefficiency is not necessarily a guarantee of joint venture failure. If collaboration results in a degree of monopoly control and competitive advantage in the domain over which the joint venture operates, the joint venture may survive comfortably and indefinitely, even though its costs of operation may be higher than if a single firm solution had been adopted.
Thus, our provisional conclusions above regarding joint venture inefficiency have to be qualified by recognition that such inefficient solutions may be sustainable in certain situations. The problem is that such a qualification does not take us very far in helping to develop a coherent and logical explanation of joint venture activity, and it certainly does not help to account for the proliferation of joint venture agreements between progressive high technology companies in developed countries in recent years.
In the next section we shall look at the relationship between diversification and collaboration options and suggest that this holds an important key to analysing the evolution of joint venture activity in the firm.
9.5. The Evolution of Collaboration in the Diversified Firm
A fundamental problem with conventional approaches to collaborative activity between firms is that they tend to take the agreement as their basic unit of analysis, and then they analyse possible motives firms may have for pursuing individual agreements. Just as analysis of product-markets in (p.190) isolation may obscure patterns of diversification, so analysis of collaborative agreements in isolation may obscure patterns of collaboration and fundamental reasons for their existence. Very few current analyses of collaborative activity could be said to provide a satisfactory explanation as to why they evolve in particular contexts and at particular times. In most studies, the various objectives ascribed to collaborative activity generally do not differ from those that have been traditionally attached to merger activity, and it is typically not clear why one mode should be preferred over another.
In this section we shall extend the analysis of previous chapters into a simple mental experiment to explore conditions likely to be associated with the emergence of collaborative activity amongst firms. We shall start with a population of sixty-two IBUs constituted as separate firms, and assume that each is pursuing growth opportunities. As in Chapter 7, such opportunities may or may not involve sharing market or technological links. However, we shall not worry about imposing strict coherence conditions on our firms, not because coherence is not deemed to be important in practice, but because it is an unnecessary restriction in this case. The firms will still be recognizably related-constrained, related-linked, and conglomerate, and it is these attributes which will drive the development of the system in our example, irrespective of whether or not the firm displays the strict coherence and strong patterns associated with the analysis of Chapter 7. We shall focus throughout on decisions from the point of view of a particular firm, or group of firms (indicated with a background mat as in earlier analysis).
If we were to draw parallels with evolutionary biology, Fig. 9.2 is an economic equivalent of the primordial ooze. The firms in Fig. 9.2 are simple specialized IBUs operating in a fragmented, unstructured environment. There is nothing to differentiate any firm in Fig. 9.2 from any other in terms of scale or profitability. Now, suppose the firms begin to explore growth (p.191) opportunities. In our experiment, this will involve either merger or joint venture, with merger involving the combination of existing firms and joint venture involving the setting up of a new IBU jointly owned by two firms. For simplicity we shall ignore the possibility that the firms may expand using wholly owned internal expansion and/or multinational options. To begin with we shall ignore possible implications of environmental turbulence and conglomerate strategies, but we shall deal with these issues later. So, what strategies and modes are likely to emerge in such a process?
As our analysis in Chapter 4 suggests, the first preference for growth purposes is likely to be specialization strategies in which both market and technological links are exploited. Thus, in Fig. 9.3 firms have sought and merged with IBU twins to pursue this strategy. As Chapter 4 also suggests, problems of coordinating the linkages involved in such strategies through market transactions are likely to encourage the adoption of the internalization (merger) option. In Fig. 9.3, forty-eight firms have merged to form twenty-four new firms formed around IBU pairs to exploit double linkages associated with specialization. Fourteen IBUs have been unable to find suitable partners to merge with at this stage.
At some point firms will run up against barriers to further specialization. The most obvious barrier is simply running out of room for further expansion opportunities for a particular business, although the possibility of antitrust intervention may impose barriers to such expansion at an earlier stage. In this experiment we assume that such saturation takes place in each case after the merger of two IBUs. If the firms now seek further expansion opportunities they can exploit single links at most. In the figure, we have also added shaded background mats to help in the identification of certain strategies. In Fig. 9.4 eleven pairs of firms have combined to form 4-IBU related-constrained strategies. There are three firms on the middle two mats which have been formed around single links to form one 3-IBU and two 2-IBU related-constrained firms. Outside the mats, two single and two double IBU firms have been left behind in this expansion process. On the mats themselves, a series of firms are considering further expansion opportunities. The top two mats (p.192) both have three firms considering ways of coordinating potential marketing and technological linkages, whilst on both of the two bottom mats there are pairs of firms negotiating starting a new IBU (the white IBU in both cases) which would draw upon the marketing expertise of one company and the technological expertise of the other. Again, these processes of expansion are consistent with Ollinger’s (1994) study of the historical growth of firms in the US oil industry, with firms pursuing strongly linked (horizontal and domestic) growth opportunities to begin with, then into more weakly linked petrochemical technologies, energy, and international markets, and then finally into unrelated technologies and markets as opportunities for further related growth opportunities became scarcer.7 This would be consistent with a weak pattern of linkages in the sense described in Chapter 5, with the firm progressively moving into areas of weaker linkages as it expands over time.
Just like specialization (or double-linked) strategies, individual linkages are likely to encounter saturation limits eventually. It is reasonable to assume that the opportunities associated with individual marketing and technological competencies can only be stretched so far at any point in time. In Fig. 9.5 we assume these limits are reached after two, four, or six IBU combinations, depending on the particular market or the technology. The fourteen related-constrained firms from Fig. 9.4 have still been seeking (p.193) growth opportunities, and in Fig. 9.5 all but two of them managed to find merger possibilities. The six 8-IBU firms can still exploit linkages though combination, but linkage saturation limits have forced each of them into a related-linked strategy. Figure 9.5 shows that the firms on the top three mats from Fig. 9.4 have merged to coordinate potential linkages and have emerged as related-linked firms in each case. The firms on the bottom mat in Fig. 9.4 have also merged to internalize linkage opportunities, and now form the 9-IBU combination on the left of the bottom mat in Fig. 9.5.
The firms on the largest two mats in Fig. 9.5 are now considering starting new IBUs that could draw upon their marketing or technological expertise in the respective cases. On the bottom mat, a single IBU firm possesses a technological expertise that could be married with the marketing expertise of the two larger firms to produce a new IBU in the respective cases. Again, these new opportunities are indicated by white IBUs on the bottom mat. On the large mat at the top, the middle firm is similarly exploring new IBU opportunities that could draw upon complementary expertise to be provided by itself and the other two firms in the respective cases. The merger option is still the most obvious mode to coordinate these new opportunities, but now some problems are becoming apparent if this option is to be continued with. If the firms were to merge to exploit the emergent IBU opportunities in the respective cases, it would create a 20-IBU firm on the bottom mat, and a 26-IBU monster on the top mat. At the moment, no firm covers more than nine IBUs and 13 per cent of those in this sector. If the (p.194) amalgamations take place, the two new firms would cover two-thirds of the IBUs in this sector.
What is happening here is that merger opportunities at each stage carry with them the baggage of the past. It raises the possibility that the evolution of cooperative activity may be a path dependent phenomenon, and this is something we explore in more detail in the next chapter. The four new business opportunities on the two large mats draw upon market and technological links provided by separate firms, just as in the cases of the firms in Figs. 9.4 and 9.5 that combined to form the firms on the two smaller mats in Fig. 9.6. However, in these earlier cases the new opportunity represented a significant proportion of the economic activity of the combined firm (as much as one-third of the firm in the case of the 3-IBU combination on the smallest mat in Fig. 9.6.). By way of contrast, if two of the firms on the top mat merged to exploit one of the new opportunities, this major shift in corporate boundaries would be undertaken to coordinate the exploitation of an opportunity that would constitute only about 6 per cent of the activity of the new firm. Such a solution would be tantamount to taking a sledgehammer to crack a nut (Kay et al, 1987). More precisely, merger to exploit the venture possibility may have serious efficiency and regulatory implications for the separate or collective operation of business units in the combined firm that go well beyond the business strategy implications of coordinating complementary assets for the new business opportunity.
If the effect of merger was restricted to the IBUs directly involved in the linkage opportunities associated with the new business opportunity, these scale effects would pose absolutely no problem at all. However, merger may (p.195) well have knock-on effects on corporate performance in two major categories. The first and most obvious implications follow from the scale of the merged firm if two large diversified firms are involved. Against the gains from coordinating complementary assets with the new opportunity must be set any additional costs of bureaucracy in running a considerably enlarged enterprise. Unless these are zero, which may be taken to be a possible but unlikely situation, merger to exploit the new business opportunity will incur costs that are externalities from the perspective of the IBUs involved with the new opportunity (and consequently ignored in analyses like Fig. 9.1), but which may have significant and adverse efficiency implications seen from the perspective of corporate strategy. In such circumstances, the firms may pass up exploitation of the common venture opportunity, or consider alternative forms of hierarchical control such as joint venture. The second major category of problem is the more units the respective firms operate, the more likely it is that merger may raise competition or anti-trust issues. If any two units that had been formerly operated separately within the respective firms are now likely to jointly control a significant portion of a market post-merger, then merger itself may be discouraged or prohibited. A possible solution could be to divest one of the offending units, but this would entail sacrifice of the gains from complementary assets that the divested unit would otherwise have exploited with neighbouring units before merger took place. Such reorganization creates further opportunity costs of merger to set against the local gains from exploiting complementary assets with the new business opportunity. Thus, the firms in the large mat in Fig. 9.6. decide instead to opt for joint ventures to coordinate linkages (indicated by white dotted lines for marketing links and black dotted lines for technological links).
Thus, when the focus changes from the new venture to the new firm (and from business strategy to corporate strategy), it becomes apparent that single firm or merger solution to exploiting complementarities may generate significant additional costs or prohibitions that are ignored if attention is paid only to the direct costs and benefits associated with the new business opportunity. This is indicated in Fig. 9.6. with the four new business opportunities being coordinated through joint venture in the respective cases. Whilst joint venture is certainly likely to be more expensive than the merger alternative over the region of the relevant collaborating IBUs in the respective cases, it may well be a cheaper or more acceptable alternative at the level of the respective firms.
What about the single-IBU firm considering new venture opportunities with the two larger firms on the bottom mat in Fig. 9.5? Clearly merger to exploit both opportunities would create an extremely large 20 IBU firm and this could lead to the same problem of quantum jumps in the scale of firms discussed above. However, if the single-IBU firm were to merge with one of its potential partners (say the 8-IBU firm) on the bottom mat, this would (p.196) mean that one new venture was internally coordinated within the enlarged 10-IBU firm, whilst still leaving the opportunity to exploit the other venture opportunity with the 9-IBU firm through a collaborative agreement. This internalizes one of the two potential agreements within corporate boundaries with only a modest increase in the scale of one of the firms.
One reason why this solution may not be adopted is that assimilation of the single IBU firm by one of the larger firms might compromise the other potential agreement. The two larger firms on the bottom mat may be content to cooperate with the small independent firm, but they may also be rivals that would not wish to enter into collaborative agreements with each other. The large number of businesses in which they operate may bring them into conflict in more than one area. The firms may therefore maintain a cooperative umbrella over both venture possibilities as indicated by the dotted lines on the bottom mat in Fig. 9.6. Whilst this may be expensive, it may also be the only way to ensure that both opportunities are pursued. Merger with one could sacrifice cooperation with the other. As we shall see in the next chapter, this may be an important issue surrounding the continuing independence of small firms in the biotechnology sector.
These arguments clearly pave the way for a rationale for joint venture activity based on efficiency criteria, and this may be illustrated more clearly with Fig. 9.7 which shows the joint venture set up by the two top firms in the top mat in Fig. 9.6. The mat in Fig. 9.7 shows the region associated with the joint venture in this case, covering the donor IBUs from the respective firms as well as the jointly owned venture itself. The domain of the joint venture is likely to be an area of intense administrative and contractual activity as discussed earlier. Just as unusual activity in the sun’s corona in the form of a solar flare is associated with a rise in the temperature over the local hot spot, so intense managerial activity channelled into the joint venture will be associated with higher coordination costs compared to the single firm option. However, while there may be a flare up of heightened coordination costs across the joint venture hot spot, collaboration also helps prevent significant spillover of costs into other regions of the participating firms. The scale of the respective firms is only marginally greater than (p.197) before, (expanded by their respective shares in the new IBU), and any antitrust implications that would have followed from full blown merger are likely to have been defused. The joint venture option can therefore be preferred over the merger option.
Thus, joint ventures are indeed locally costly relative to single firm options as earlier analysis suggests, but it is the systems-wide costs of coordinating the respective alternatives that matters, not just the direct costs involved in this local hot spot. Just as it would be a mistake to infer the average temperature of the sun’s surface from that of the region of solar flares, so it would be a mistake to infer the overall costs of institutional arrangements by focusing on the local costs of specific ventures.
One further point that requires emphasis here is that the form of diversification is likely to be an important influence on the development of collaborative activity. The arguments of Kay et al (1987) and Hennart (1988a) are based on the idea that it is difficult to decompose or disengage the various business assets in a firm pursuing related diversification, whether related-constrained or related-linked. Hennart uses the expressive phrase ‘tangled assets’ to capture this notion. However, there is one diversification strategy in which untangling or disengaging assets is relatively straightforward—the conglomerate. If the bottom firm in Fig. 9.7 was a conglomerate, then we would have an arrangement as in the conglomerate joint venture of Fig. 9.8. Such an arrangement is unnecessarily complicated; the conglomerate could simply decouple its contributing business unit and sell it to the other firm for a price that allows both to share in the resulting enhanced value, with the top firm continuing its related-based strategy. This is the asset trading option in Fig. 9.8. The sale of the donor IBU by the conglomerate to the collaborating firm results in the latter fully internalizing linkages that would otherwise have been coordinated by joint venture. The three IBUs that would have been involved in the joint venture are indicated by the mat in this case.
Conglomerate decomposability would facilitate asset transfer and coordination that reflects their highest valued uses. A possible brake on this process is that, carried to its logical conclusion, it might tend to shrink conglomerates as in Fig. 9.8 to an unsaleable rump composed of elements of little strategic value or interest to other corporations. Such a fate may not be entirely palatable to conglomerate management whose objectives may include seeking to maintain the corporate system and managerial coalition. A managerial limit on asset trading could be reflected in the collaboration option being pursued even for the conglomerate case in some instances.
We can summarize our position so far as suggesting there are three conditions for the evolution of joint ventures. If analysis does not recognize each of them as necessary elements, then it is liable to be incomplete or misleading, and indeed one of the problems of limited comparators in this area is that the three conditions have not always been recognized. First, there (p.198) must be gains from complementary assets compared to the competition or abstention option. In Figure 9.1, this is represented by the (S) option versus the (C) option in the top row. This prerequisite is uncontroversial, the problem is that many of the checklist based approaches stop at this prerequisite. Secondly, the hierarchical option [(S1) and (S2) type options in Fig. 9.1] must be a more efficient way of coordinating complementary assets than are contractual options. Contract is more efficient when strategic decision-making concerning the nature and integration of complementary assets is effectively complete, as is the case when dealing with commodity-type markets. ‘Where buyer and seller accept no obligation with respect to their future conduct however loose and implicit the obligation might be, then co-operation does not take place and we can refer to a pure market transaction’ (Richardson, 1972, p. 886). As we have argued, hierarchy is more efficient where important strategic decisions concerning the nature and integration of complementary assets remain to be made in the future, as in the case of market entry and technological innovations. New markets and new technologies are both likely to necessitate creating and maintaining capacity for managerial decision-making capacity (and so hierarchy) with managers occupying space that would otherwise be occupied by contract in stable, established situations. It is this need for future decisions that creates hierarchy. Hierarchy buys time by providing a bridge to the future that would otherwise be constrained by premature contractual commitments.
Thirdly, joint venture must be a more efficient way of coordinating complementary assets than are single firm options, subject to regulatory restrictions being fulfilled. The problem here is that over the region of the joint (p.199) venture, any gains that can be made by joint venture can typically be made more cheaply by single firm alternatives; it is this localized costliness that must be compensated for if joint venture is to be justified. ‘Regulatory restrictions’ refers to constraints such as anti-trust policies or governmental restrictions on multinational access to local markets. This prerequisite means there may be two major categories of joint venture: (a) joint ventures that help deal with tangled asset problems in pursuing specific new business opportunities. This is more likely to be the case with large diversified firms than small specialized firms, as we have seen above; (b) joint ventures that are resorted to by firms if government restrictions preclude single firm operation or place too high a cost on such options. This prerequisite means that it is not necessary to identify the spurious gains that a joint venture can achieve and a single firm cannot. Also, it is entirely consistent with the joint venture domain itself remaining a coordination cost hot spot.
Some studies have attempted to justify joint venture on the basis that the venture itself may be low cost if trust and mutual forbearance exists between partners.8 But, even though goodwill may lower the coordination cost temperature of the joint venture hot spot, it is unlikely that such dual control systems could ever be reasonably expected to operate as simply and efficiently as conventional unified hierarchies. Similarly, efficiency gains from joint venture that have been identified in the economics literature include a sharing of knowledge, risk, and capital.9 However, as long as the concept of the joint venture coordination cost hot spot holds, these studies do not explain why joint venture is more efficient than single firm operation in these regards. It is only when the joint venture hot spot is set in the context of the wider corporate systems that efficiency rationales for joint venture evolution begin to emerge.
Before we look at actual patterns of joint venture activity, it should be noted that arguments made here concerning coordination of complementary assets tend to apply similarly to cases involving substitutable assets as in Fig. 9.1.
9.6. Empirical Patterns
We have a nested set of prerequisites for the evolution of joint ventures. Exploitation of shared assets must be more efficient than abstention or do-nothing alternatives, In turn, for the range of institutional arrangements that can exploit shared assets, hierarchy must be more efficient than contractual alternatives. Finally, for the range of hierarchical arrangements than can exploit shared assets, joint venture must be more efficient than single firm operation after allowing for regulatory restrictions. We have argued that systems most conducive to the evolution of joint ventures are likely to be those in which systems wide efficiency implications dominate (p.200) considerations that hold over the limited domain of the new venture opportunity itself. We have argued that this is most likely in circumstances involving large diversified firms. Joint venture tends to evolve where there is a need to implant hierarchy to enable future strategic decision-making in cases where single firm alternatives are inferior or prohibited.
This perspective allows us to make sense of various patterns of joint venture evolution.
1. Stages in practice
Joint ventures to facilitate access to Third World markets by host governments have been a feature of international business for many years. However, Dunning (1993, p. 257) notes that the proliferation of joint venture activity between technologically progressive and sophisticated partners is a relatively recent post-war development, while Harrigan (1987, p. 195) estimates that joint ventures (and other forms of cooperative agreements) have only become particularly widespread since 1980. This phenomenon is understood most easily by standing it on its head and asking why such patterns of collaboration had not been observed in earlier periods if they are indeed efficiency enhancing and desirable.
However, we need to note first that cooperative or collaborative agreements cover a wider set of modes than just joint venture. For our purposes, the important distinction is between cooperation agreements that; (a) make provision for future decision-making with strategic content between the parties, and install appropriate hierarchical controls with this in mind, and (b) effectively treat the agreement as contract with all important strategic decisions established to begin with, and little, if any, provision for future strategic decisions in relation to the area of cooperation. Joint venture is likely to fall into the first category, while a straightforward licensing agreement would be more likely to fall into the second category, though of course the degree of hierarchical provision for future-related decision-making may vary from case to case. A problem is that whilst we are interested in the extension of hierarchy as represented by cooperation agreements such as joint ventures, databases have an unnerving habit of aggregating the two types together in their analysis of general trends.
In this context, Hagedoorn and Schakenraad (1991, p. 6) make an extremely helpful distinction between cooperative agreements that are aimed at the ‘strategic, long term perspective of the product market combinations of the companies involved and cost-economizing agreements which we think are more associated with the control of either transaction costs or operating costs of companies’. These distinctions appear to closely parallel our type (a) and type (b) distinctions. Hagedoorn and Schakenraad (1991) further distinguish between two groups of modes, finding that the first (joint ventures, research corporations, joint R&D agreements, and equity investments) (p.201) are over 85 per cent strategically motivated, while only a small proportion of the second group of modes (technology exchange agreements, one-dimensional technology flows, and customer-supplier agreements) are strategically motivated. In some cases, Hagedoorn and Schakenraad (1991, pp. 7–8) suggest that this latter group of agreements may be analysable in transaction cost terms, following Williamson (1985).
It is the long-term strategically motivated agreements that may require provision for future decisions (and the creation of joint administrative or hierarchical arrangements) that we are interested in. A separate paper by Hagedoorn and Schakenraad (1990) provides information that may be helpful in assessing the relative importance of cooperative agreements that are likely to be oriented towards long-term strategic positioning, compared to those that are more likely to be concerned with short-term cost economizing. They studied the development of alliances and inter-firm agreements in three core technologies (new materials, information technology, and biotechnology) up to 1989 using the MERIT-CATI database which logged 4619 agreements in these areas (out of over 7000 cooperative agreements). Of the 4619 agreements in the three core technologies, group 1 type agreements (joint ventures, research corporations, joint R&D, and equity investments) accounted for 2739, or 59 per cent. The proportion of group 1 modes was well above 50 per cent in each of the three sectors. In all three sectors, over 90 per cent of agreements had been established in the 1980s. This is consistent with other studies of the incidence of cooperative agreements. Analysing the INSEAD database on international collaborative agreements reported in the Financial Times and Economist over the period 1975–86, Hergert and Morris (1988) found that over the period they grew from ‘almost zero to the point where new ventures are announced on a daily basis’ (p. 99)10 whilst in a wide ranging review of a number of empirical studies, Hagedoorn (1993a) concludes that cooperative agreements on a large scale first appeared in the 1970s (p. 134). This is also broadly consistent with results from the LAREA/CEREM database (Mytelka, 1991) which collected 1086 inter-firm agreements (1980–87) in biotechnology, information technology, civil aeronautics, and the auto industry, involving at least one European based collaborator. Mytelka (1991) shows that the average number of agreements per year rose from 67 in 1980–82, to 133 in 1983–5, to 243 in 1986–87.11
This allows us to make three points. First, it is possible to distinguish between cooperative agreements in terms of their likely emphasis on hierarchy versus market coordination, as Hagedoorn and Schakenraad (1990, 1991) have done. Some (like licensing) may be heavily characterized by market exchange characteristics and contractual arrangements. Others like joint ventures are likely to add provision for future strategic decisions (with implications for extension of hierarchy into the area of the agreement). Secondly, it appears that strategically-oriented agreements constitute a (p.202) significant proportion of cooperative agreements. As a first approximation, it would seem that about half of the cooperation agreements in the core technology sectors surveyed by Hagedoorn and Schakenraad (1990) are likely to invoke some degree of continuing strategic decision-making, while the other half are likely to be largely contractually oriented. Thirdly, whilst most measures of cooperative agreements between firms tend to fudge the distinction between those that are strategically oriented and those that are short-term agreements, the evidence is that both broad categories of cooperative agreements have achieved their present importance from a base close to zero in the 1970s.
These patterns are consistent with our expectations that it was the diversification strategies pursued by large corporations in mature economies in the first half of the post-war period (Rumelt, 1986; Channon, 1973; Dyas and Thanheiser, 1976) that provided the crucial set of triggers required for the evolution and proliferation of cooperative activity in the second half of this period. We would expect these trends to have stimulated the growth of all types of cooperative ventures, since large diversified firms considering the potential exploitation of linkages (whether strategically oriented or cost economizing) are likely now to be forced to turn to modes other than merger to do so. In particular, the coordination solutions provided by options such as joint venture were more easily provided by merger options in earlier decades. Before the Second World War, even large corporations tended to be small and specialized when compared to the larger and diversified corporations that evolved later. Where firms found areas of overlapping interest, as in the cases of complementary and substitutable assets in Fig. 9.1, merger was typically the obvious solution (up to the limits provided by regulatory constraints on expansionary zeal). Systems wide effects of the type associated with related-linked expansion were likely to be nonexistent or of minor importance. It was the post-war diversification boom that created the conditions that really stimulated the growth of joint venture activity.
Further, in a study of the evolution of cooperative agreements in the biotechnology industry, 1971–89, Barley et al (1992) concluded that as the creation of new subsidiaries of diversified firms and new freestanding firms began to decline in the early eighties, so the rate of formation of new cooperative agreements markedly increased. Again, this is consistent with the early phase of diversification triggering subsequent cooperative activity in this particular sector.
2. Size, diversification, and joint venture activity
Cross-sectional analysis tends to display a link between size of firm and propensity to form joint ventures (Boyle, 1968; Berg and Friedman, 1978).12 Since, ceteris paribus, large firms also tend to be more diversified, these (p.203) results are consistent with joint venture being stimulated by prior diversification activity. It would be useful if we could separate out the effects of diversification from size, and Berg and Hoekman (1988) provide extremely useful evidence in this respect. They analysed joint venture activity in the Netherlands as of 1981 and again confirmed that size of firm had a stimulating effect on joint venture activity (noting also that similar patterns appeared when comparable Swedish, US and German data were analysed). They then analysed propensity to undertake joint venture activity for the 150 largest Dutch corporations separated into those that (a) stayed close to a specialized base with less than 10 per cent of its turnover outside its core business, and (b) diversifying or vertically integrating firms (DIV/VI firms), with more than 10 per cent of turnover outside their core business. The corporations were ordered in groups of 25, ranked according to turnover. For the first group of 25 (i.e. the largest corporations), DIV/VI firms were only 48 per cent of this cohort but accounted for 89 per cent of the 209 joint ventures formed. For the second group of the next largest 25 firms, DIV/VI firms were only 25 per cent of the cohort but accounted for 75 per cent of the 105 joint ventures formed. DIV/VI firms were 24 per cent of the third cohort, but accounted for 73 per cent of the 74 joint ventures, they were 8 per cent of the fourth cohort but accounted for 40 per cent of the 51 joint ventures, they were 16 per cent of the fifth cohort but accounted for 76 per cent of the 50 joint ventures, they were 12 per cent of the last cohort of 25 (the smallest firms here when ranked by turnover), but accounted for 64 per cent of the 25 joint ventures. In this case the stratification of the corporations into groups of like-sized firms helps to suggest a clear association between diversification activity (or more generally, movement away from a core business) and subsequent joint venture activity on the part of the firm. The evidence is generally consistent with prior diversification triggering subsequent joint venture activity. Thus, the suggested relationships between joint venture, diversification, and merger appear to hold over time and at a point in time, consistent with the perspective developed here.
3. Merger/joint venture relationships
The relationship between merger and joint venture is more involved than is generally acknowledged. It is usual to think of them as substitute options, but merger can also serve as the crucial source of diversification that triggers subsequent joint venture activity. For example, Scherer and Ross (1990, p. 92) point out that the diversification of US corporations 1950–75 was carried out to a considerable degree through merger and not internal growth. Consequently, not only does merger stand as a nominal alternative to specific joint ventures, merger activity served as the precursor to later joint venture activity by creating the pre-condition of diversification that helps foster joint venture activity.
Williamson (1985, pp. 83–4) supposes transactions to be arrayed along a dimension representing the degree to which parties to a trade maintained autonomy, with conventional market contracting at one end, and centralized hierarchy at the other. Cooperative agreements, such as joint venture, would be located in between. Whereas before he had earlier regarded cooperative agreements as both infrequent and inherently unstable (1975, pp. 108–9), with a bimodal distribution of transactions clustering around the extremes of the market and hierarchy dimension, Williamson later modified this position in the face of empirical evidence to the contrary to acknowledge that the distribution was typically more uniform than he had earlier conceded (1985, pp. 83–4). He concludes ‘whatever the empirical realities, greater attention to transactions of the middle range13 will help to illuminate an understanding of complex economic organisation. If such transactions flee to the extremes, what are the reasons? If such transactions can be stabilized, what are the governance processes?’ (Williamson, 1985, p. 84).
However there is no mystery concerning the proliferation of joint ventures in the view developed here. There is no tendency for cooperative agreements such as joint ventures to flee to the extremes of pure market or hierarchy, nor a need to stabilize them, for the simple reason that the mistake is to conceive of them as representing some form of hybrid or intermediate state somewhere between pure market and pure hierarchy on some hypothesized dimension. As far as the individual venture opportunities themselves are concerned, the trigger for the adoption of the joint venture mode is not to be found in the characteristics of the ventures themselves, but in the context in which they are set. Unless the role of previous corporate diversification is identified, the joint venture will remain an enigma inviting spurious explanations based on the supposed merits and demerits of the mode considered in isolation. In fact, joint ventures lie outside traditional market and hierarchy modes and are not simple hybrids produced by selectively fusing elements from market and hierarchy extremes. As argued above, joint venture is the extension of hierarchy by a less direct route, and is triggered when the natural evolution of firms and sectors produces blocks to simpler hierarchical alternatives. This helps to explain the increased frequency and persistence of joint venture activity, despite these phenomena representing puzzles when viewed from the perspective of conventional transaction cost economics.14
5. Joint venture and product-life cycle considerations
We would expect to find hierarchical solutions such as multinational expansion and joint venture used at earlier stages in the product life cycle than contractual alternatives such as licensing, given that innovative stages tend (p.205) still to have significant strategic decisions to be made in the future. This is both consistent with an internalization extension of the stages of development approach (Young et al, 1989, p. 33), and with empirical work by Davidson and McFetridge (1985) who studied 1200 intra-firm and market technology transactions by US firms 1945–78. They concluded that the probability of internal intra-firm hierarchical transfer was greater for R&D intensive companies and for newer technologies and technologies with few previous transfers.
As with the analysis of diversification and multinationalism, the key to understanding the phenomena of joint ventures is to maintain the firm as the basic unit of analysis. Studies which have analysed the joint venture in isolation have tended to miss the point that the joint venture is an extremely expensive device through which to coordinate resource allocation, and that it is typically only resorted to as a device of last resort. Whilst this may include host governments’ requirements for a local partner, in recent years it has increasingly been associated with large diversified firms imposing hierarchy through joint venture solutions because a merger of the two firms is costly and impracticable. It is not the characteristics of the joint venture that typically poses problems in these respects; it is the variety of businesses and linkages that merger would bring together that is the problem. The resource-based perspective helps bring these issues centre-stage, and hopefully contributes to an analysis of a variety of phenomena: these include reasons why joint ventures have proliferated only relatively recently; relationships between firm size, diversification and joint venture; the role of previous merger activity in triggering later joint venture activity; joint venture stability and frequency; and the place of joint venture in the product life cycle.
(2) . Indeed Williamson points out that size, and especially diversity, tend to be prerequisites for the adoption of the M-form structure by individual corporations.
(3) . Schwarzenegger’s film company hired the aircraft from the Pentagon for the making of the 1994 film True Lies (The Sunday Times, London, 21 August 1994, p. 19). The development of the Post-it has become the stuff of legend in the (p.206) innovation literature, and it could be argued that much serendipity surrounded the eventual commercial impact of this innovation and that neither market nor hierarchical arrangements could have anticipated the initial development or subsequent phenomenal success of this product. However, the important point here is that serendipity, uncertainty, and surprise are all concepts which can be difficult or impossible to handle in a contractual context, but which may be more appropriately handled within a hierarchy. Mintzberg (1979, pp. 273–7) also points out that environmental complexity differs from stability or uncertainty and develops implications for bureaucratization and decentralization.
(4) . This is not to say that multiple objectives and conflict resolution techniques are not observed in a unified, single firm hierarchy (e.g. Cyert and March, 1963, analyse the implications of multiple objectives in managerial coalitions and discuss associated conflict resolution procedures). However, the point being made here is that dual control is likely to add a further layer of cost elements in this respect, compared to those associated with unified hierarchy.
(5) . The general argument here is that knowledge (such as R&D) tends to have high fixed costs of generation, but once generated it can frequently be provided to a new user at low additional cost per user. In this respect it can be similar to other public goods; it may cost a great deal to construct lighthouses to service just one user, but a second user can use the services provided at zero marginal cost. Since allocative efficiency obtains where price equals marginal cost, the allocatively efficient price is zero when marginal cost is also zero. Consequently, there is an efficiency argument for the free and unfettered dissemination of knowledge.
Obviously, such principles would dim or eliminate the private incentives to search for new technologies, since innovators would not be able to protect their intangible assets from appropriation by others. Consequently, systems such as patent law tend to reflect a balance between private incentives and public good arguments.
(7) . Ollinger sees the subsequent restructuring of firms and shedding of unrelated assets in this sector as shareholder interests (and profitability concerns) asserting themselves over growth-oriented managerial interests. If anything, this suggests that what may be a failed investment in terms of shareholder interest may be satisfactory from the point of view of managers—in the absence of shareholder intervention.
(8) . See Kay(1992b) pp. 205–6.
(10) . Though this trend may be overstated to the extent that this phenomenon only became one of general interest in the 1980s and so may have been underreported in these sources in the early part of this period. International collaborative activity was certainly not invented in the 1970s; for example Gomes-Casseres (1988) traces a broadly increasing trend in the relative incidence of joint ventures in new manufacturing subsidiaries formed by US MNEs 1900–75, with a sharp decrease in the relative incidence of joint venture formation in the 1960s. Gomes-Casseres attributes this latter behaviour to cyclical phenomena.
(12) . Boyle’s results were especially striking, finding with his US based sample that (p.207) joint ventures had parents from the top 100 of the Fortune 500 in 42 per cent of cases, whereas parents from the next 100 only occurred in 4 per cent of cases.
(13) . That is, cooperative agreements such as joint ventures.
(14) . It is important to distinguish questions of persistence of solutions at the corporate level from those at the business level. Cooperative agreements may become established as stable and continuing phenomena pursued by corporate management, whilst the lives of individual cooperative agreements are more likely to be tied to (and curtailed by) lower level product life cycle considerations. The difference lies in the level being looked at; as discussed in Chapter 3, higher level institutions may reveal a stability and permanence that is not visible if the focus is solely on the characteristics and behaviour of constituent elements.