Joint Ventures Paper 16 March 1999

Short of a merger: the competitive effects of horizontal joint ventures. Stephen P. King Professor of Economics Department of Economics The University of Melbourne Parkville Vic. 3052 E-mail: [email protected]

Horizontal joint venture agreements are an important tool of corporate organisation. These agreements, however, may have competitive implications. This paper considers the potential anti-competitive effects that can flow from horizontal joint ventures. We consider how different features of a joint venture agreement may either exacerbate or moderate these effects and present a set of guidelines that may be used to evaluate the competitive implications of joint ventures.

1

Introduction Arrangements between market participants often have implications for competition. Mergers, for example, that “would have the effect, or be likely to have the effect, of substantially lessening competition in a market” are illegal under s50 of the Trade Practices Act 1974. Section 47(6) of the Act makes third line forcing a per se offence – one firm cannot offer to supply goods or services on the condition that the buyer “will acquire goods or services of a particular kind or description directly or indirectly from another person”. Similarly, section 48 makes resale price maintenance a per se offence. The economic benefits of these arrangements are often ambiguous. While a merger may substantially lesson competition, it may also create efficiency gains for the merged firm. From an economic perspective, these gains need to be weighed against the competitive effects in analysing a merger.1 A tying or resale price maintenance contract may provide net economic benefits. For example, if there is a customer service component associated with retail sales, resale price maintenance can improve welfare by focusing competition on service rather than price.2 This paper considers the competitive effects of horizontal joint ventures.3 The term ‘joint venture’ covers a wide range of arrangements and is only poorly defined under the law. A joint venture may be incorporated but “unincorporated joint ventures are much more common in Australia and tend to involve a range of contractual provisions which may give rise to competition concerns”.4 It involves less integration of the relevant firms than a merger, and may provide conflicting pro- and anti-competitive effects. Our aim is to consider the competitive implications of joint venture

1 2 3

4

Williamson, O. (1968) “Economies as an antitrust defense: the welfare tradeoffs”, American Economics Review, 58, 18-36, presents a simple framework for analysing this trade-off. See Tirole, J. (1988) The theory of Industrial Organization, MIT Press, Cambridge, MA. The competitive implications of joint ventures are currently being analysed, for example, in the United States where the Federal Trade Commission established a joint venture project in 1996. Tonking, I. (1998) ‘Joint venture arrangements and long term contracts: animal farm revisited?’ Commentary presented to the 1998 Trade Practices Workshop, Melbourne Business School. 2

arrangements and provide insight into how these arrangements might be analysed by competition authorities.5 We will consider joint ventures involving agreements between horizontal competitors. In other words, the firms entering the venture compete or potentially compete against each other in some market. When two potential competitors enter into a joint venture arrangement, the degree of competition between the firms may be affected. A joint venture falls short of a horizontal merger between two competitors. But how do we judge the potential anti-competitive and pro-competitive effects of a joint venture?

Joint Ventures What is a Joint Venture? There is no standard economic definition of a joint venture. Rather the term is used to describe a range of activities between firms that fall short of a complete merger. For example, Williamson defines a joint venture as “two or more bodies co-operating for some common purpose. Beyond that it is really a matter of precise description rather than definition, for ‘joint venture’ is really a convenient label for a variety of activities, which may be undertaken in a variety of ways, with a variety of legal consequences”. 6 Similarly Harrison (1975 p117) notes that the “term ‘joint venture’ covers a variety of forms of co-operation. The term is vague enough to include all situations in which two or more persons or companies join forces to achieve some common goal”.7 Because joint ventures are only loosely defined, the term encompasses a variety of modes of co-operation. Pitofsky argues that the term is used to cover agreements 5

6 7

There are clearly many other important legal issues surrounding joint ventures. For example, what fiduciary duty is created by a joint venture agreement? What is the relationship between the creation of joint ventures and the taxation laws? These questions, while interesting, are beyond the scope of this paper. Williamson, D. (1977) “Trade practices law - its implications for mining and petroleum joint ventures”, Australian Mining and Petroleum Law Journal, 1, 59-108 at p.85. Harrison, F. (1975) “Joint ventures and the Trade Practices Act 1974: the American approach and its applicability to Australia”, Australian Business Law Review, 3, 117-131 at p.117. Similarly, Pitofsky, R. (1985) “A framework for antitrust analysis of joint ventures”, Antitrust

3

ranging from ‘naked’ cartels that involve no integration of resources or creation of new productive capacity, to mergers. In-between, and captured by the term joint venture are a broad range of co-operative business undertakings “ranging from purely contractual collaboration (eg., horizontal information exchanges, joint advertising arrangements) through the joint ownership or participation in use of existing assets (eg., sports leagues, patent cross-licenses), to the joint creation of a new enterprise to conduct manufacturing, distribution, or research and development activities”.8 In order to narrow the range of activities under consideration, some authors restrict attention to joint ventures that involve the explicit creation of a new entity. For example, a joint ventures involves “the creation of a new business organization which is owned by two or more enterprises”.9 Under this definition, a production swap or ‘tolling agreement’ would not be considered a joint venture. Even if a joint venture is defined as involving the creation of a new entity, the type of entity formed can still distinguish different joint ventures. For example, Backman distinguishes between joint ventures where the new organisation is temporary and those where it is permanent.10 There are also differences according to each party’s rights to use the products or services of the new organisation. For example, Tao and Wu distinguish between different forms of research and development (R&D) joint venture on the basis of the parent companies’ rights to use any joint venture innovations.11 Legal definitions also distinguish different types of joint venture. For example, Bensaid, Encaoua and Winckler note that EC competition law distinguishes between ‘co-operative’ and ‘concentrative’ joint ventures.12 “A joint venture is regarded as concentrative if two criteria are met: first the joint venture must perform on a lasting basis all the functions of an autonomous entity ...; second, there must be no

8 9 10 11 12

Law Journal, 54, 893-914 at p.893, “[a] joint venture could involve any business enterprise in which two or more persons collaborate to achieve some commercial goal”. Op. Cit note 6 at p.894. Bernstein, L. (1965) “Comments on ‘Joint ventures in light of recent antitrust developments: joint ventures in the chemical industry’, Antitrust Bulletin, 10, 25-29 at p.27. Backman, J. (1965) “Joint ventures in light of recent antitrust developments: joint ventures in the chemical industry”, Antitrust Bulletin, 10, 7-24. Tao, Z. and Wu, C. (1997) “On the organization of cooperative research and development: theory and evidence”, International Journal of Industrial Organization, 15, 573-596. Bensaid, B, Encaoua, D, and Winckler, A. (1994) “Competition, cooperation and mergers: economic and policy issues”, European Economic Review, 38, 637-650. 4

coordination of competitive behavior between the parent companies or between parent company and joint venture. ... The Commission ... considers a joint venture to be concentrative where only one parent withdraws from the relevant market provided the other parent effectively plays the leading role in determining the industrial behavior of the joint venture. ... Any joint venture which does not satisfy the two proceeding requirements is considered as cooperative”.13 These two types of joint venture are treated differently by European competition laws.14 15 In Australia, joint ventures are defined under s.4J(a) of the Trade Practices Act 1974 as “an activity in trade or commerce (i) carried out jointly by two or more persons, whether or not in partnership; or (ii) carried on by a body corporate formed by two or more persons for the purpose of enabling those persons to carry on that activity jointly by means of their joint control, or by means of their ownership of shares in the capital, of that body corporate”. As with other definitions of a joint venture, this section of the Act encompasses a huge class of arrangements. In the absence of a clear definition of a joint venture, the economic literature must be approached with some caution. Different researchers will not only consider different aspects of firm behaviour with a joint venture, they will also analyse a variety of different institutions under the joint venture rubric. Similar care is necessary in the absence of a clear legal definition. Pitofsky argues that the lack of a rigorous legal definition of a joint venture in the US has resulted in uncertainty about their antitrust treatment and “uncertainties in enforcement policy have almost certainly blocked, delayed, or raised the cost of legitimate undertakings”.16

13 14

15

16

Ibid, p.640. See also Burnside, A. and Mackenzie Stuart, J. (1995) “Joint venture analysis: the latest chapter”, European Competition Law Review, 2, 138-146. Bensaid, et al model the differences between the two types of venture. In a concentrative venture, participants remain in competition in the final product but have incentives to ‘free ride’ on investments in the joint venture. In a cooperative joint venture, the free riding is internalised but there is a greater reduction in competition as the partners do not separately compete. Shenefield, J. (1985) “Export joint ventures”, Antitrust Law Journal, 54, 1039-1050 at p.1041) notes that export joint ventures often raise distinct legal issues. For the US, “[s]ince export joint ventures by definition involve trade or commerce ... with foreign nations, neither the Sherman Act nor Section 5(a) of the Federal Trade Commission Act will apply unless the challenged conduct has a ‘direct, substantial and reasonably foreseeable” anticompetitive effect on a protected interest, i.e., either domestic United States commerce, United States import commerce, or export commerce of a person engaged in such commerce in the United States”. Op. Cit. Note 6 at p.893. 5

The range and legal implications of joint ventures Joint ventures occur at a variety of production stages and differ in their coverage of the production chain. Williamson notes the range of potential production levels that can be covered by a mining or petroleum joint ventures, including “exploration, construction, extraction, processing, transport or marketing. Participants may wish to co-operate all the way along the chain, or they may choose to go separate ways at some point along the chain”.17 Similar distinctions can be made in other industries where joint ventures may deal with R&D, production and/or marketing. The effect of a joint venture will depend on both its location in the production chain and its coverage of the production chain. In this paper we focus on joint ventures between horizontally related parent companies. Many joint ventures fall into this category. For example, Pfeffer and Nowak look at 163 US joint ventures from 1960 – 71.18 They note that 56% involved parent companies in the same industry (horizontal) and only 14% involved vertically related parents (where one parent was in an industry where more than 10% of sales or purchases were with the industry of the other parent). In addition, 36% of the joint subsidiaries were in the same industry as both parents and in 19% of cases the joint subsidiary was vertically related to the two parent firms. Common vertical relationships include a joint-sales or marketing agency. In Australia, sections 45 and 45A of the Trade Practices Act 1974 deal with the potential anticompetitive effects of joint ventures. Joint ventures are treated differently to other arrangements between firms. Section 45A(2) of the Act for example moderates the per se illegality of price fixing that holds for most other contracts, arrangements or understandings between firms under s45A(1). Joint ventures might also attract attention under section 50 of the Act if they involve acquisitions that would result in a substantial lessening of competition. The Australian Competition and Consumer Commission (and before it, the Trade Practices Commission) can authorise joint venture arrangements under s.90 of the 17 18

Op. Cit. Note 5 at p.85. Pfeffer, J and Nowak, P. (1976) “Patterns of joint venture activity: implications for antitrust policy”, The Antitrust Bulletin, 21, 315-339. 6

Act. Examples include Re Electric Lamp Manufacturers (Australia) Pty Ltd; Pasminco Ltd, Australian Mining and Smelting Ltd; and Re SAGASCO Resources Ltd.19 Similarly, under United States law, joint venture arrangements might be considered a restraint of trade that violates section 1 of the Sherman Act. In addition, to the degree that a joint venture arrangement mimics a horizontal merger, it may violate section 7 of the Clayton Act. The United States v Penn-Olin Chemical Co et. al. provides an important example.20 Unlike Australia, the U.S. authorities cannot authorise a joint venture that would break these antitrust laws. The role of specific terms in a joint venture agreement. The potential economic benefits and costs of a joint venture will critically depend on the structure of the joint venture and the terms and conditions included in the joint venture agreement. Williamson notes that a joint venture agreement will include terms relating to at least some of: “the exercise of control over the venture by the joint venturers; the transfer of assets from the venturers to the venture, the regulation of asset usage by the venture, or the acquisition of assets from external sources; the provision of finance to the venture and the sharing of its costs or profits, if any; terms upon which new venturers may be admitted or upon which existing venturers may withdraw; resource exploration or product research and development; and the exploitation of venture production, including the terms and conditions upon which, and the prices at which, venture products are to be sold either by the venture or by the venturers”.21 The relationship between specific clauses in a joint venture agreement and the potential anti-competitive effects of the joint venture may be complex and requires careful analysis. An example is provided by a common form of joint venture that involves the construction of a facility which is jointly used by the participating firms (eg. a gas pipeline joint venture). The joint venture partners then compete in a downstream market. The participating firms may be allocated capacity in the joint 19 20 21

(1982) ATPR 50-033, (1988) ATPR 50-082 and (1992) ATPR 50-118 respectively. 378 US 158 (1964). Op. Cit. Note 5 at p.87 7

venture facility in a number of ways. For example, capacity shares may be fixed according to each firm’s share in the joint venture. “In a ‘cost-center’ cotenancy, production is overseen by an independent management company. Output and pricing decisions are made independently by the co-owners who essentially place orders for output with the management company”.22 In this case, dedicated capacity for each owner is determined by that owner’s share, but additional competitive rules may be used to increase interfirm rivalry. Gale considers the pro-competitive effects of a use-it-or-lose-it rule in a cotenancy joint venture.23 This rule means that each party to the joint venture may utilise any unused portion of the other parties’ capacity by paying the variable costs of that capacity use. The rule has been used in the US.24 Gale shows that the rule can improve facility use and competition. In particular, each firm has an incentive to use its capacity, knowing that otherwise it may lose capacity to a downstream competitor. Non-competitive benefits of joint ventures Why would firms form a joint venture? Kitch considers three types of potential noncompetitive benefits for joint venture parties.25 These are benefits due to externalities, economies of scale and transactional efficiencies. Externalities occur, for example, in advertising if one firm can ‘free ride’ on product (rather than brand) advertising expenditures of another firm. A joint venture can internalise this externality. Similarly, in the absence of a joint venture, firms may engage in either excessive or insufficient R&D. Excessive R&D may arise when there is a ‘winner takes all’ patent system in place while the absence of strong patent or copyright protection may lead to free riding and insufficient research. Again a joint venture can internalise these effects, potentially benefiting both the firms and society. Firms can use a joint venture to reap otherwise unobtainable economies of scale. Shapiro and Willig (1990) break these economies of scale into two subclasses; (1) 22 23 24

Gale, I. (1994) “Price competition in noncooperative joint ventures”, International Journal of Industrial Organization, 12, 53-69 at p.53. Ibid See ibid at p.54 for details.

8

synergies arising when venturers share complimentary skills or assets; and (2) the attainment of economies of scale and scope in production, procurement or logistics.26 The former, for example, may apply to a gas pipeline that can be shared by a number of closely located gas producers. The latter is an argument for cross production joint ventures. If production technologies are similar across firms but their final product demands are negatively correlated, then a joint venture to share production facilities may more efficient than each firm owning their own facilities. A similar argument applies to firms with plants located in different geographic locations for a product with high transport costs. It may not be profitable for each firm either to build a new factory in the other firm’s region or to transport significant quantities of final product between regions. If existing plants have excess capacity, a production joint venture may then allow both firms to compete in both locations at minimal cost. Of course, the firms themselves would only want to enter such a joint venture if it raised, rather than lowered total profits.27 Transactional efficiencies may arise from the allocation of risk. Backman (1965, p8) notes that “[r]isks may be undertaken in a joint venture that a single enterprise would be unwilling to assume”.28 For mining or petroleum, Williamson notes that joint ventures may be structured to help guarantee a market for product or to provide sufficient security to raise large capital borrowing.29 Co-generation joint ventures may improve the allocation of risk between parties. The joint venture may be formed by a firm with expertise in building gas-fired electricity generators and a firm that requires substantial, but variable quantities of electricity for its production. The joint venture is mutually beneficial by securing a guaranteed electricity supply at known price for the producer while providing a large customer for the supplier. There may also be economies of scope, for example, if the related firm can also use waste steam from the

25 26

27 28 29

Kitch, E. (1985) “The antitrust economics of joint ventures”, Antitrust Law Journal, 53, 957994. Shapiro, C. and Willig, R. (1990) “On the antitrust treatment of production joint ventures”, Journal of Economic Perspectives, 4, 113-130. Backman, op. cit. note 9 at p7 also considers a variety of ways in which a joint venture can improve economies of scale. A joint venture can be used to “pool technological knowledge, production know-how, selling ability, and/or financial resources”. The refinery sharing arrangements between petroleum companies in Australia may provide this type of benefit. Op. Cit. note 9 at p.8. Op. Cit. note 5. 9

generation facility. These types of vertical relationships are closely related to long term supply contracts, which are considered in the third section of this paper. Anti-competitive costs of a joint venture. Pitofsky notes that four anti-competitive effects have dominated US cases: “(1) reduction of potential competition; (2) tendency of parents to stifle what would otherwise be independent future growth of the joint venture or one of the parents; (3) “spill-over” effects with respect to competition among the parents of a joint venture; and (4) the existence of collateral restrictive agreements”.30 The first two of these effects may be summarised as effects relating to independent entry. Independent Entry. In the absence of the joint venture the parent companies may influence competition through their potential to enter in the future. This will affect current competition in the relevant product market to the extent that incumbent firms follow a different strategy because of the potential for entry. Further, if entry actually did occur in the future, then the parent company would obviously directly affect product market competition through its own participation in that market. To analyse the effects of a joint venture on competition, the U.S. courts have “typically ask[ed] what the parents would have done but for the joint venture”.31 To analyse what would have occurred without the joint venture, Williamson notes three potential counterfactuals: “(i) none of the joint venturers would have entered upon the scheme by itself; (ii) one of the joint venturers would have entered upon the scheme by itself, while others remained as potential competitors; or (iii) each of the joint venturers would have entered upon the same scheme as competitors”.32

30

31 32

Op. Cit note 6 at p.896. See also Harrison op. cit. note 6. He distinguishes between the first two effects as the “[e]limination of the present competitive force of possible future entry” and the “[e]limination of possible future entry thereby diminishing competition in the future” (p121-22). The former deals with the current effect of the elimination of a potential competitor. The latter asks whether the joint venture partners may have separately entered in the future but that this is now precluded by the joint venture. See Pitofsky op. cit. note 6 at p.897. For an example, see, for example, the Penn-Olin decision. Op. Cit. note 5 at p.87. 10

If, in the absence of the joint venture, both parent firms would have entered or are already in the relevant product market and the joint venture agreement implicitly or explicitly prevents interparent or parent-joint venture competition, then this would appear to be strong evidence that the joint venture is likely to have anticompetitive effects. Pitofsky claims that “[t]he defenders of joint ventures involving parents that would have competed but for the arrangement have a considerable burden to overcome”.33 Even if both joint venturers are either in, or would enter, the relevant market, this does not mean that the joint venture will be anti-competitive. For example, if there are significant cost savings from the joint venture then competition with the joint venture may lead to lower prices for consumers than competition without the joint venture. Even if prices are higher under the joint venture, due to its effect on reducing competition, the consumer loss due to the price increase may be more than offset by the savings due to increased productive efficiency.34 Alternatively, “[w]here neither parent was in the market or a likely entrant into the market served by the joint venture, it is hard to see why those joint ventures should be treated other than legal per se”.35 In these circumstances, the joint venture promotes entry and is likely to be pro-competitive rather than anti-competitive. “The more difficult joint venture case involves the situation, treated by the supreme court in Penn-Olin, where one parent is in the market or is a highly likely entrant and the other parent is not likely in the near future to enter on its own”.36 In this situation, the U.S. approach has involved evaluating the extent to which the non-entering parent was a potential entrant and effected competitive conduct in the relevant market. If this parent is now precluded from entering by the joint venture then there is a loss of potential competition.37 Pitofsky considers four additional factors that need to be shown before it can be claimed that the joint venture has reduced potential competition - the market is not 33 34 35 36 37

Op. Cit. note 6 at p.897. The same issue arises in horizontal merger analysis. Pitofsky, op. cit. note 6 at p.898. See also Backman op. cit. note 9 at p.16. Pitofsky, ibid at p.898. Pitofsky ibid at p.899 notes that parents often either explicitly or implicitly agree not to compete with the joint venture. See also Kitch op. cit. note 24. 11

competitive, other competitors in the market would have viewed the ‘in the wings’ competitor as a potential entrant, the non-entering parent was one of the most likely potential entrants, and the ‘in the wings’ entrant had a substantial pro-competitive effect on the market.38 In many circumstances these additional constraints will not be satisfied. The Electric Lamp Manufacturers joint venture in Australia provides a useful example. The joint venture involved four firms that otherwise might have been independent competitors in the Australian market. The joint venture also accounted for 85 per cent of the incandescent lamps and 70 per cent of the fluorescent lamps consumed in Australia in 1980. However, imports accounted for the remainder of Australian consumption and the Commission found that import competition provided an effective foil for the potential anti-competitive effects of the joint venture. Competitive Spillovers. Baker notes that the participants in a joint venture may “be tempted to cooperate in matters outside the joint venture. For example, local clearing house meetings once served as an occasion for the member banks to agree upon the prices they were going to charge their customers”. 39 More generally, joint venture agreements may lead to a variety of informal arrangements that limit competition. For example, the exchange of information between joint venture participants, including price information, may aid tacit collusion. Consider, for example, a joint venture where two firms ‘swap’ production between plants. This arrangement may limit competition by changing the competitive relationship between the firms. If one firm pays above marginal cost for production by a competitor, then that competitor has an interest in both its own sales and the sales of the other firm. This will tend to temper competition. If a firm acts aggressively in the market, raising its own sales at the expense of its rival, then this may result in a direct increase in profits. At the same time, the profits from producing the competitor’s product will fall. The ‘tolling’ agreement will reduce the incentives for a firm to act aggressively. The existence and extent of this effect will depend on a number of factors - is production being ‘swapped’ at prices that exceed marginal costs, are the 38 39

Ibid at p.898-9. Baker, D. (1974) “Antitrust as a positive force in relation to financial joint ventures”, paper delivered at the United States League of Savings Associations’ 82nd. annual convention, San Francisco, November 13, 1974. 12

amounts of exchanged production fixed, and what production will be affected by a marginal reduction in one firm’s sales? In addition, cross production makes it difficult for one firm to ‘surprise’ its competitor with new product developments and market strategies, as that competitor is involved in the production process.40 Pfeffer and Nowak argue that spillover effects will be more likely in horizontal joint ventures and will be most prevalent and most anticompetitive at intermediate levels of concentration. At low concentration joint ventures are ineffective as collusive devices. At high concentration they are unnecessary. They have the biggest effect at intermediate concentration levels.41 Collateral restrictive agreements. There have been a number of US cases involving collateral restrictive agreements. In the Yamaha-Brunswick case, the joint venture agreement restricted independent potential entry by the parents. More generally, joint venture agreements may reduce competition without any offsetting benefits by including collusive clauses that do not directly relate to the benefits of the joint venture. For example, a joint production agreement between two oil companies may also involve a joint marketing agreement, even when separate marketing would be equally efficient. Where the main aim of the additional agreement is to mute competition, the relevant clauses should be removed from the joint venture agreement. Can joint ventures increase competition? If a joint venture enables firms that would otherwise have been unable to enter a market to do so, then the joint venture will increase rather than decrease competition. Williamson notes that a joint venture may increase competition by “the lowering of entry barriers, ... the achievement of economies of scale, ... addition to productive

40

41

This is similar to the argument presented by Reynolds, R. and Snapp, B. (1986) “The competitive effects of partial equity interests and joint ventures”, International Journal of Industrial Organization, 4, 141-153. They compare a joint venture with cross-ownership and note how the joint venture may provide an incentive for firms to be less aggressive. Op. Cit. note 17. In contrast, Pfeffer and Nowak argue that vertical joint ventures are more likely to be related to reducing uncertainty than anticompetitive reasons. “[T]he advantages [of a vertical joint venture] include ensuring a stable and predictable source of supply of important materials or components, or ensuring reliable outputs and markets for the product” (p332). 13

economic activity”.42 Barriers may include financing and risk sharing. Put simply, if a joint venture aids entry then it is likely to aid competition. The more complex case involves a joint venture between firms that are already participating in a market. Kwoka presents a simple model of a horizontal production joint venture that involves a new facility.43 The firms who form the joint venture gain by being able to commit to aggressive production and sales. When the new production facility commences operation and is managed independently from the joint venture parents, total production in the market can rise. Customers gain by lower prices, the joint venture partners gain by seizing a larger market share and higher total profits but other competitors lose. Overall, the joint venture is pro-competitive. The Kwoka model highlights the potential benefits for both joint venture partners and the public from a commitment to strict separation between a joint venture production facility and its parents’ operations. This separation must, however, be credible. Also a small spillover from the joint venture that increases industry wide cooperation and reduces overall competition can wipe out these gains and lead to a fall in output and higher prices. Comparing mergers and joint ventures. Joint ventures are distinguished from mergers and generally involve less commitment than a merger. “In a merger, two or more companies combine all of their assets to create a new entity. In a joint venture, two or more companies combine less than all of their assets to create a new entity”.44 Kitch (1985) considers that a joint venture lies on a continuum between a cartel or ‘naked price fix’ and a single firm, so it involves less restraint than a merger and more benefits than a naked price fix.45 As joint ventures involve less restraint than a merger, there is a consensus of opinion that they should be treated more leniently than a merger by antitrust authorities.46

42 43 44 45

Op. Cit. note 5 at p.89. Kwoka, J. (1992) “The output and profit effects of horizontal joint ventures”, Journal of Industrial Economics, 40, 325-339. Bernstein, op. cit. note 8 at p.25. Op. Cit. note 24. 14

Shapiro and Willig argue that “[t]o the extent that the production joint venture preserves the assets, the ability and the incentives of the parents to compete independently of one another and of the production joint venture, the venture will be recognized to pose less of a threat to competition than would a full merger of the parents”.47 Similarly, Pitofsky (1985, p897-8) notes that “joint venture parents do not permanently disappear from the scene but may continue competing in other areas of their business ... [and] unlike mergers where one or both parents usually lose their independent existence, joint venture parents remain independent and hence remedies may be imposed more easily and at a later date if the combination turns out to be anticompetitive”.48 Joint ventures may be viewed more favourably by antitrust authorities because (1) they allow for the continued existence of the parents and potentially continued competition between the parents and (2) in comparison with a merger, joint venture agreements are more easily modified or dismantled at a latter date. A reasonable rule of thumb is that if the joint venture parents would be allowed to merge, then there should be little concern about the parents forming a joint venture. “Any genuine joint venture should almost surely be allowed if the participants would be permitted to merge”.49 This discussion suggests that the ACCC’s merger guidelines should form the first stage of any joint venture analysis and that there should be consistency between merger and joint venture analysis. In particular, if the parent firms would be allowed to merge then the joint venture provides little concern for competition.

46

47 48 49

Care must of course be taken to compare the joint venture with a relevant ‘equivalent’ merger. For example, a joint venture between two multi-product firms involving all their products might raise more anticompetitive concern than the outright sale of one product division between the firms. But the joint venture will raise less concern than a complete merger between the two firms. Op. Cit. note 25, p.118. Op. Cit note 6 at p.897-8. Op. Cit. note 25, p.127, italics in original. Kitch op. cit. note 24 at footnote 10 states that, when comparing merger and joint venture market share cutoffs, the cutoffs should be higher for joint ventures except those relating to current marketing or sales, as the threat to competition is less. See also Brodley, J. (1990) “Antitrust law and innovation cooperation”, Journal of Economic Perspectives, 4, 97-112. 15

There are two minor caveats to this rule, Kitch notes that there may be overlapping joint ventures within an industry and that these would need to be considered when analysing either the effect of a merger or of a new joint venture.50 Nye (1992) points out that if a joint venture leads to less cost reduction than a merger then a merger may actually lead to less of an increase in price than a joint venture.51 In this second case, if a merger was more efficient from the parents’ perspective and a proposed merger would pass the scrutiny of the ACCC, then we would expect the firms to pursue merger rather than a joint venture. If they continue to pursue a joint venture, then it is likely that there is some additional factor that has been omitted from the Commission’s analysis. Joint ventures and the production chain. Two key issues arise when analysing joint ventures at different production stages. First, does the range and location of the joint venture down the production chain change the potential anti-competitive spillovers? Secondly, can a joint venture at a specific stage of production be used to limit product market competition? There is a broad belief that joint ventures have an increased probability of anticompetitive effects if they relate more directly to product market competition. In other words, the further the joint venture extends down the production chain, the more likely that it will aid tacit collusion and reduce competition. For example, Kitch claims that “[m]ost suspect would be joint ventures related to current marketing and pricing of the members’ product. Least suspect, in order, would be joint ventures related to research and development of future products, promotion of the product type, or a phase of the production process”.52 Similarly, “[t]he joint sales agency has been a classic cartel device, and there is ordinarily little economic justification for joint selling between firms that have market power”.53 Alternatively, a joint venture may enable effective monopolisation of one level in the vertical chain of production. The monopoly power achieved at this production stage 50 51 52 53

Op. Cit. note 24 at footnote 10. See Nye, W. (1992) “Can a joint venture lessen competition more than a merger?”, Economic Letters, 40, 487-489. Op. Cit. note 24. Brodley, op. cit. note 47 at p.101. 16

can then be used to raise final product prices and achieve monopoly profits. “[E]ven if the parents’ production and marketing decisions remained independent, the venture could transfer all its output to all the parents at a sufficiently high price to ensure that the parents would independently charge elevated prices to their customers, and the venture would be the locus for the earning of monopoly profits at the expense of the public interest”.54 The ability to seize monopoly profits from one stage of production depends on the availability of substitute inputs and on the ability to price above marginal cost. If there are few or no alternative inputs then the input will be necessary for production and the joint venture will control an essential facility. In fact, the essential facility doctrine in the US is often traced back to the terminal railroad case that involved the conduct of a railway joint venture. When dealing with joint ventures involving key facilities, Pitofsky takes the view that “[i]f participation in the joint venture confers a ‘significant competitive advantage’ and the venture itself holds a substantial market position, the law may require that competitors be allowed to participate in the joint venture or obtain the advantages of membership on reasonable and non-discriminatory terms. A significant competitive advantage does not mean the membership in the joint arrangement is ‘indispensable,’ but only that lack of opportunity to participate would seriously diminish the company’s ability to compete”.55 Potential remedies when a joint venture controls a key facility involve use-it-or-lose-it clauses as discussed above; a requirement to allow new participation, as suggested by Pitofsky; and rules for third party access. For example, a joint venture may be viewed more favourably if it has made an access undertaking under s.44ZZA of the Trade Practices Act. For a joint venture at one stage of production to be used as a collusive device by parent companies, the joint venture must be able to price to its parents above marginal cost. “To avoid the antitrust concern ..., a consortium may be structured to transfer its 54 55

Shapiro and Willig, op. cit. note 25 at p.115. Op. Cit. note 6 at p.902. Opening access to a joint venture, or any other facility, is an area of considerable controversy. It is easy to create undesirable incentives if inappropriate access rules are used. See for example, King, S. and Maddock, R. (1996) Unlocking the infrastructure: the reform of public utilities in Australia, Allen and Unwin, Sydney. 17

output to members at marginal cost”.56 There may be problems enforcing marginal cost pricing within a joint venture. However, transfer pricing is a key element that needs to be analysed when considering the potential anti-competitive effects, say, of a production joint venture. Example – research and development (R&D) joint ventures. Much of the economic literature has focussed on R&D joint ventures. These joint ventures offer obvious benefits, in terms of internalising research spillovers, muting socially undesirable patent races, and allowing synergies between research skills. Ordover and Baumol list the benefits as (i) spreading the risk of investment in knowledge production, (ii) reducing wasteful duplication, and (iii) facilitating information dissemination while not undermining its appropriability.57 They also note potential costs - reduction in total R&D investment, a reduction in the number of research approaches, and a possible reduction in downstream competition. R&D joint ventures might also raise competitive concerns. In this section, we focus on R&D joint ventures to illustrate some of the points raised in the discussion above. Consider the trade off between potential cost savings and anticompetitive spillovers. Yi compares independent and cooperative R&D.58 If one firm independently engages in cost reducing research and/or development then it creates two externalities for other firms. First, some of the savings may spill over to competitors, particularly if the processes cannot be protected by patent or copyright. These spillovers lower all firms’ costs and tend to raise profits. However, independent R&D also tends to intensify competition. By lowering its costs relative to its competitors, the firm engaging in R&D will act more aggressively in the market raising its profits at the expense of its rivals. Both of these externalities have positive benefits for society as a whole. While R&D spillovers are also a positive externality between the firms, the increase in competition is a negative inter-firm externality. By cooperating on R&D through a joint venture, firms internalise the interfirm externalities. If the positive benefits of cost savings are more important, firm and social interests are aligned and 56 57 58

Shapiro and Willig, op. cit. note 25 at p.117. Ordover, J. and Baumol, W. (1988) “Antitrust policy and high-technology industries”, Oxford Review of Economic Policy, 4, 13-34. Yi, S-S. (1996) “The welfare effects of cooperative R&D in oligopoly with spillovers”, Review of Industrial Organization, 11, 681-698. 18

the joint venture will lead to more R&D and increased social welfare. If the negative effects of increased competition dominate firm decisions, then there is a conflict between social and firm welfare and the joint venture will result in less R&D and less final product competition.59 An upstream research and development joint venture may feed into final market competition. But the analysis by Yi shows that the net public benefit is extremely sensitive to changes in costs and competition. Cost savings from an upstream joint venture are socially beneficial but these may be partially offset or outweighed by anticompetitive spillovers. The exact form of the joint venture may also have competitive implications. Tao and Wu distinguish between two types of R&D joint venture.60 With an equity research joint venture (RJV), each parent company owns shares in the RJV. “[T]he RJV owns the future innovations, and the participating firms need to pay royalty fees for the use of the innovations”. In contrast, with a non-equity development (COD) participating firms can freely use any innovations. Tao and Wu show that, when the participating firms compete in the final product market, then an RJV leads to strictly higher profits (and prices) after innovation than a COD. The reason for this was outlined in the previous section; the royalty fees that the participating firms need to pay to the RJV allow the firms to limit product market competition. The fees will be set to mute competition and to transfer and distribute ‘collusive’ profits through the RJV. In contrast, with a COD, post-innovation competition cannot be avoided due to the ability of each participating firm to freely use the innovation. This does not necessarily mean, however, that the COD is socially preferred compared to the RJV as the choice of joint venture will influence the rate of innovation. If an R&D joint venture is socially desirable, then its desirability may be reduced if it extends a considerable distance down the production chain, allowing increased anticompetitive spillovers. However, Jorde and Teece argue that R&D joint ventures can only be fully successful if they operate well down the production chain.61 For

59 60 61

See also Martin, S. (1994) “Private and social incentives to form R&D joint ventures”, Review of Industrial Organization, 9, 157-171. Op. Cit. note 10 at p.574. Jorde, T. and Teece, D. (1990) “Innovation and cooperation: implications for competition and antitrust “, Journal of Economic Perspectives, 4, 75-96. 19

example, an R&D joint venture may only fully realise potential cost savings if it extends into marketing, so as to allow feedback from customers and sales staff to the research staff. Jorde and Teece emphasise that R&D is not a linear process but is a ‘simultaneous’ system with other production activities. Jorde and Teece also argue that significant downstream integration and/or cooperation will be needed if patent protection is poor. Similarly, if technology is costly to transfer, possibly because of the technical nature of an innovation, the upstream R&D process will need to be integrated with the downstream production and sales. “A public policy stance that treats only early stage activity as potentially requiring cooperation is misguided and will thwart both early and later stage activities. Most firms will not have much incentive to engage in early stage, joint development if later stage, stand alone commercialization appears too expensive to accomplish profitably”.62 The issues raised by Jorde and Teece make competition analysis complex. If allowing further downstream cooperation increases both the social benefits and the costs of an upstream joint venture, where do authorities draw the line? In the case of R&D, it may be desirable to treat ‘new’ products differently to new versions of existing products.63 At the same time, Brodley presents examples of patent cross-licensing and pooling being used to sustain anticompetitive practices such as price fixing.64 In addressing these concerns, Ordover and Baumol argue that because research joint ventures reduce duplication and ‘free-riding’, improve technological diffusion and dissemination (particularly where licensing would be problematic) and involve less coordination than merger, antitrust policy should be “highly permissive”.65 They place some caveats on this. If the joint venture will substantially affect future production of knowledge, by controlling “a large share of the assets that are critical to the production, dissemination and adoption of new technology”, then this will make new entry to the final product market more difficult.66 The joint venture may reduce downstream competition and they note that “the less vertically integrated downstream

62 63 64 65 66

Ibid. p.92. See Brodley, op. cit. note 47. Ibid. Op. Cit. note 55 at p.29. Ibid at p.30. 20

is the venture, the less likely it is to be used to facilitate collusion in the product market”.67 However, as already noted, if there is strong downstream competition between participants, then this competition may wipe out the gains from a research joint venture, so in some cases restrictions on ex post competition between joint venture participants may be desirable. Ordover and Baumol also express concern about the competitive effects of excluding competitors from the joint venture. Joint venture stability In many circumstances, neither the ACCC nor the courts will be able to accurately judge the potential anti-competitive effects of a joint venture. While joint ventures that have blatantly anti-competitive intentions may be obvious, in most cases it will be difficult to accurately weigh costs and benefits. To minimise the possibility of approving joint ventures that offer little real benefit in terms of reducing costs, the Commission may want ensure that participants have the ability to terminate the joint venture agreement if claimed cost savings do not eventuate. The Commission may view joint ventures more favourably if they are less stable to the competitive pressures that are created by an anti-competitive arrangement. Kogut considers the stability of joint ventures.68 He finds that joint ventures are more stable if there are other ‘ties’ between the participants. In particular the existence of other joint ventures and/or licensing arrangements between the parties (horizontal ties) improves the stability of the joint venture. Vertical ties (supply agreements) have no effect. “Embedding the venture in other long-term economic relationships among the partners has a stabilizing influence on cooperation” as each party has more to lose by behaving opportunistically in the joint venture, as it could effect other agreements.69 Industry stability may influence the success of a joint venture. “[C]ompetitive conflicts among the partners are engendered by the growth of external opportunities”, such as market growth and changes in relative firm size.70

67 68 69 70

Ibid at p.30. Kogut, B. (1989) “The stability of joint ventures: reciprocity and competitive rivalry”, The Journal of Industrial Economics, 38, 183-198. Ibid at p.193. Ibid at p.195. 21

Joint ventures involving R&D tend to be relatively stable. Also, in industries of medium concentration, there may be more incentive to form joint ventures to mitigate competition, but also, these joint ventures may be more fragile due to the potential for opportunistic behaviour in these industries. Kogut’s analysis suggests that the competition authorities should be less concerned about the anti-competitive effects of joint ventures in the same type of industries where tacit collusion is unlikely to be sustainable. For example, if rapid new entry is feasible; where the joint venture partners each have only a relatively small market share, raising the incentive for opportunistic behaviour even if the total share of the parties is high; where firms can quickly gain market share through contracts with key customers; and where there are few formal or informal ties between firms other than the joint venture. Approaches to joint venture analysis. Given the range of difficulties raised by joint venture analysis, it is not surprising that most commentators have offered only vague and poorly defined rules for competition authorities to judge joint ventures. Williamson suggests four principal considerations to be used to judge anticompetitive potential: “(i) whether the joint venture possesses or threatens to posses substantial market power; (ii) whether the joint venturers, individually or collectively posses such power; (iii) whether the joint venturers are competitors; (iv) whether the product of the joint venture is closely related - horizontally or vertically - to those produced by the joint venturers”.71 With the exception of the last of these items, these considerations are identical to the issues raised by merger analysis, and simply highlight that such analysis should be the first stage in judging a joint venture. Jorde and Teece address the relationship between mergers and joint ventures when considering changes to U.S. antitrust laws and suggest (1) a ‘safe harbor’ if the relevant firms have less than 25% of the market and (2) “integration by contract or alliance should be treated no less favorably than full mergers”.72

71 72

Op. Cit. note 5 at p.90. Op. Cit. note 59 at p.89-90. 22

To move beyond merger analysis, antitrust authorities need to consider additional aspects of the joint venture. Pitofsky argues that “given equal magnitudes of anticompetitive effect, legality or illegality of joint ventures should be influenced by (1) the extent of integration and efficiencies and (2) whether the restrictions on competition growing out of the joint venture are reasonably related to achieving those efficiencies”.73 While the extent of efficiencies will generally be difficult to measure, there should be a reasonable suspicion of joint ventures that appear to extend unnecessarily down the production chain. In addition, any extraneous agreements that are not directly related to the efficiencies of the joint venture are likely to be socially undesirable.74 The relationship between the joint venture parents and their relationship with the joint venture are of key importance. Pfefer and Nowak (1976, p323) believe that “joint ventures created by horizontally related firms are the most anticompetitive, with conglomerate joint ventures the least potentially harmful to the structure of competition”.75 Shapiro and Willig note that “welfare is higher and consumers are better off if the consortium’s market power is limited by independent competition from its members. Therefore, the antitrust authorities should and do require that any restrictions on independent conduct by members be necessary for the success of the venture. Antitrust authorities must be on the alert to detect ventures that offer the appearance but not the reality of independent action”.76 Overall, the more independence each parent and the joint venture have from each other, the fewer are the competitive concerns. The discussion above highlights the importance of different forms of joint venture. For example, does the form of the joint venture artificially reduce competitive 73 74

75 76

Op. Cit. note 6 at p.904. When considering whether or not a restriction in a joint venture agreement is reasonable, Pitofsky, op. cit. note 6 at p.911 states “[t]he point is not that joint ventures should be found illegal if the same or similar efficiencies can be achieved in a less restrictive manner. As several U.S. courts have noted, that formulation is too demanding since it would place joint venture organizers at the hazard that others might come along later and think of some method of achieving similar efficiencies in a manner that is somewhat less restrictive. But when efficiencies are offered to justify what would otherwise be anticompetitive restrictions, it is essential to discount those efficiencies where they could be achieved with a substantially lesser level or the complete absence of restraint”. A conglomerate joint venture involves firms that are in other lines of business and not vertically related. Op. Cit. note 25 at p.116. 23

incentives? Can the transfer prices set by the joint venture be used to limit downstream competition? When considering anti-competitive effects, could the joint venture be restructured to provide equivalent benefits with fewer competition concerns? Competition authorities may also want to consider the potential for procompetitive additions to a joint venture agreement, such as third-party access requirements, or non-discrimination rules to new participants. Conclusion The discussion on joint ventures presented in this paper suggests a number of questions that authorities, such as the ACCC, need to answer when evaluating joint ventures: (1) Would a merger between the joint venture partners be approved? If so, then there is little concern about the joint venture and it should be allowed to proceed unhindered. (2) What are the potential benefits claimed by the joint venture parties? In particular, are there likely to be improvements in productive efficiency or other benefits from the joint venture, and can these be substantiated? If there are no real savings likely to emerge from the joint venture then it is hard to view it as anything other than an anti-competitive arrangement. (3) Is the joint venture operating in a market with one or more of the parent companies? If the joint venture is essentially a new operation in a market where the parents have little or no presence, then it is likely to be pro-competitive rather than anti-competitive. (4) Are there additional informal or formal links between the joint venture partners? A joint venture creates more cause for concern if it is one of a pattern of horizontal arrangements or agreements between the parent companies. (5) What stages of the vertical production chain are effected by the joint venture? Does the joint venture extend further than necessary down the production chain? When evaluating the vertical extent of the joint venture, authorities need to keep in mind the importance of feedback between final customers and suppliers and

24

product development, particularly for research and development joint ventures. At the same time, as a rule of thumb, a joint venture that covers less of the vertical production chain and is confined to upstream activities is less likely to raise competition concerns than a joint venture that is significantly spread over the production chain or is focussed on downstream operations such as marketing. (6) Does the Joint venture control a critical stage of the production chain? The anticompetitive potential of a joint venture will depend on the importance of the product or service produced by the joint venture and downstream production. A joint venture that controls an essential facility creates more competitive concern than a joint venture agreement to produce a product with readily available substitutes. (7) How does the joint venture agreement restrict the behaviour of the parent companies? For example, can the parents continue to operate in competition with the joint venture? If a parent is not currently in the relevant industry, can it enter in the future? Can the parent companies make independent decisions over the use of their share of the joint venture output? In general, the more restrictions that the joint venture agreement places on independent action by the parent companies, the more likely it is that the agreement will reduce competition. (8) To what extent is the joint venture independent of the parents? If the joint venture is relatively independent of the parent companies then this will moderate competitive concerns even if both the joint venture and the parents are in the same market. Of course, competition authorities must be wary of arrangements that create the illusion of independence. (9) What is the structure of the relevant industry? While there is not a simple correspondence between ‘structure, conduct and performance’, anti-competitive joint ventures are more likely in industries with medium level concentrations. At the same time, competition authorities should view joint venture agreements more favourably if they are more vulnerable to anti-collusive economic forces. For example, if all the participants in the joint venture are small and can easily leave the arrangement then the joint venture is likely to be unstable unless realised cost savings really offset any anti-competitive effects.

25

(10)

Are there specific clauses in the agreement that might give rise to competitive

concerns? Joint venture agreements may involve a wide range of terms and conditions that will interact with competition and consumer welfare. These rules must be considered carefully and it may be desirable to change or augment some of the joint venture rules to improve competitive outcomes without undermining the economic benefits that can be achieved from the joint venture. (11)

Where relevant, what transfer prices or fees pass between the joint venture and

the parent companies and do these differ substantially from marginal cost? Transfer payments set above marginal cost can be used for anti-competitive purposes and serve no pro-competitive purpose. While it may be difficult for the authorities to verify the relationship between transfer prices and production costs, any prices that are significantly out of line with costs should raise concern. (12)

Are there additional conditions that can be included in the joint venture

arrangement to alleviate competitive concerns? A use-it-or-loose-it clause is one simple example. Are there any additional actions that the joint venture can take to improve competition? For example, providing an access undertaking might be appropriate when the joint venture will control an essential facility.

26