joint venture contracts

An Economic Analysis of Joint Venture Contracts1 Gilles H. Chemla2 Michel A. Habib3 Alexander P. Ljungqvist4 January 15, 2001 would like to thank Paolo Fulghieri, Sumantra Goshal, Pim Piers, Emily Sims, Diego Rodriguez, John Stopford, and Ernst-Ludwig von Thadden for helpful discussions; and Alexei Jiltsov for excellent research assistance. Part of this paper was written while Habib was visiting Ecole Superieure des A®aires de Beyrouth, whose hospitality is gratefully acknowledged. We are responsible for all errors. 2 University of British Columbia. E-mail: chemla@¯nance.commerce.ubc.ca. 3 London Business School, visiting Ecole des Hautes Etudes Commerciales, Universite de Lausanne. E-mail: mhabib@london.edu. 4 New York University. E-mail: aljungqv@stern.nyu.edu. 1 We Abstract We provide an explanation for various clauses found in joint venture contracts, such as put and call options, drag-along and tag-along rights, preemption rights, and initial public o®ering clauses. We view such clauses as serving to preclude i) transfers from the venture to one or both partners and ii) renegotiation that would distort the partners' investments. We provide an explanation for the Russian roulette clause sometimes used in put options. Felix Austria. Nube alii bellant [Happy Austria. You get married while others make war]. 1 Introduction Recent years have seen a profusion of joint ventures and other cooperative arrangements amongst ¯rms, with hardly a week gone by without some new joint venture, participation agreement, or strategic alliance having been announced.1 For many scholars and commentators, joint ventures, participations, and strategic alliances allow ¯rms with complementary resources to combine these resources more cheaply than would be the case through a merger or an acquisition, yet more solidly than would be the case through non-equity based arrangements such as licensing or cross-marketing agreements for example.2;3 Our purpose in this paper is to provide a justi¯cation for various clauses A joint venture consists in the creation of a new entity in which two or more parents hold equity stakes. A participation consists in having one ¯rm acquire an equity stake in another, but not the reverse. A strategic alliance with cross-shareholdings consists in having two or more ¯rms exchange equity stakes. 2 See Bhattacharyya and Lafontaine (1995), Darrough and Stoughton (1989), Hennart (1988), Kogut (1988), McConnell and Nantell (1985), Mohanram and Nanda (1998), and Rey and Tirole (1998) for an analysis of joint ventures; Aghion and Tirole (1994), Allen and Phillips (2000), and Lerner and Tsai (1999) for an analysis of participations; Chan, Kensinger, Keown, and Martin (1997) for an analysis of non-equity-based arrangements; and the books by Doz and Hamel (1998), Gomes-Casseres (1999), Lorange and Roos (1992), and Yoshino and Rangan (1995) for a general discussion of the various forms of strategic alliances. 3 Gulati (1995), Hennart (1988), Pisano (1989), and Oxley (1997) analyze the choice between equity- and non{equity-based arrangements. Balakrishnan and Koza (1993), Hennart (1988), Hennart and Reddy (1997), and Reuer and Koza (2000) analyze the choice between joint ventures and acquisitions. 1 1 that are found in joint venture contracts, speci¯cally:4 ² Put options, which give one partner in a joint venture the right to sell his stake to the other partner.5 The exercise price of an option may have been ¯xed at the time of writing the option, it may be determined according to some valuation formula that was agreed upon at such time or by some impartial external expert, or it may be arrived at through the use of the `Russian roulette' clause. Such a clause enables the ¯rst partner to name a price and allows the second partner to buy out or sell o® the venture at that price.6 ² Call options, which di®er from put options in that they give one partner the right to buy the other partner's stake rather than sell his stake to the other partner. The same methods that are used to determine the exercise price of a put option can be used to determine that of a call option. ² Drag-along rights. In case a partner sells his stake in the joint venture to an outside investor, these oblige the other partner to sell his stake to the same outside investor, at the same price and on the same terms as the ¯rst partner. ² Tag-along rights. In case a partner sells his stake in the joint venture to an outside investor, these allow the other partner to oblige the outside References on the legal aspects of joint ventures are Campbell and Reuer (2000) and Linklaters, Paines, and Nightingale (1990). 5 The majority of joint ventures have two partners. 6 The Russian roulette clause is also known as the `dynamite' or `shortgun' clause. It is reminiscent of the `cake-cutting mechanism' (Crawford and Heller, 1979; McAfee, 1992). 4 2 investor to buy the second partner's stake at the same price and on the same terms as he is buying the ¯rst partner's stake. ² Pre-emption rights, which oblige a partner wishing to sell his stake in the joint venture to o®er his shares to the other partner. ² Initial public o®ering (IPO) clauses, whereby the two partners agree in advance the circumstances in which they will take the joint venture public. The starting point of our analysis is Bhattacharyya and Lafontaine's (1995) argument that joint ventures between two parties arise as a response to a problem of double moral hazard. When each of the two parties must make a non-contractible investment towards the success of a project, an equity stake in the value of the project induces each party to make at least part of that investment. To the investment problem analyzed by Bhattacharyya and Lafontaine (1995), we add a transfer problem: we assume that each party can transfer part of the value of the joint venture to himself, at the cost of decreasing the value of the joint venture. Such a problem has long been a central concern of students of joint ventures and, in the extreme, has prompted some to argue that joint ventures `give away the future.'7 We further di®er from Bhattacharyya and Lafontaine (1995) in assuming that the continuation of the joint venture is but one of many outcomes: once the investments have been made, a joint venture can go public in an initial public o®ering; it can be sold to a trade buyer in a trade sale; one party can 7 See Reich and Mankin (1986). 3 buy out the other; and, of course, the joint venture form may be retained. The investments that are made are intended to increase the value of the venture when taken public or sold to a trade buyer or when one party buys out the other. We argue that the preceding clauses to joint venture contracts can be viewed as serving one or both of two roles: a clause may be intended to preclude a partner or an outside investor from engaging in a transfer; or it may be intended to preclude renegotiation. This is because renegotiation would alter the two parties' claims on the value of the venture away from the claims chosen at the outset, which would have been chosen to induce the parties to make the investments that maximize the value of the venture, subject to the constraints imposed by the problem of double moral hazard. Renegotiation would distort these investments and is therefore to be precluded. To illustrate, consider a put option. By giving a partner the right to sell his stake to the other partner at a price that re°ects the value of the venture absent any transfers, the put option serves to deny the second partner any incentive to engage in value destroying transfers. Now consider an IPO clause. By obliging both partners to take part in the IPO of the joint venture and denying either partner the right to veto the IPO, the clause precludes any renegotiation between the two partners, thereby precluding any attempt at changing the claims on the venture from those adopted at the outset. This avoids distorting the partners' incentives for making the non-contractible investments.8 Note that denying the right to veto the IPO to both partners, but allowing either partner to retain his stake in the now public company would not su±ce, for the ability of the remaining partner to engage in value destroying transfers would decrease the price 8 4 The remaining clauses can be explained similarly. Call options make each partner the unique residual claimant to the investment that enables him to buy out the other partner, thereby inducing each partner to make the ¯rst best such investment. Drag-along rights play a role similar to that played by an IPO clause, but they apply to a trade sale rather than an IPO. Tagalong rights serve to ensure that the bene¯ts a trade buyer may derive from the purchase of a controlling stake in a joint venture are shared by the two partners in the joint venture in proportion to their stakes, instead of being received only by the majority partner. Finally, pre-emption rights serve to preclude a sale to an outside investor that is motivated primarily by the ability of the outside investor to transfer more value from the joint venture than is the selling partner, when the use of a put option is precluded by the fact that the remaining partner is indispensable to the success of the venture. The paper proceeds as follows. Section 2 presents the basic model. Section 3 presents the results absent any uncertainty about the value of the venture in the various outcomes. The results pertain to put and call options (Sections 3.1 and 3.2, respectively), drag-along and tag-along rights (Sections 3.3 and 3.4, respectively), pre-emption rights (Section 3.5), and IPO clauses (Section 3.6). Section 4 extends the analysis to the case of uncertainty. Section 5 concludes. received for the venture at the IPO. 5 2 The model Consider two ¯rms a and b who choose to form a joint venture. Following the making of the requisite investments by the two ¯rms, now partners in the joint venture, the joint venture can be taken public in an IPO, it can be sold to a trade buyer in a trade sale, it can be acquired in its entirety by partner a or by partner b, or it can remain a joint venture. The values of the venture corresponding to these various outcomes are denoted Vipo (ia ; ib ), Vts (ia ; ib ), Va (ja ), Vb (jb ), and Vjv , respectively. They have probabilities pipo , pts , pa , pb , and pjv . The probability of a particular outcome or state represents the probability that the corresponding value is the highest amongst all possible values. We initially assume that Vipo (:; :), Vts (:; :), Va (:), Vb (:), and Vjv are non-random. The investments ia, ib , ja , and jb can be viewed as serving to increase the value of the venture when taken public, when sold to a trade buyer, or when one partner buys out the other over and above its value if it remains a joint venture. Our interest is therefore in the incremental value these investments make possible. This is because, as we shall argue below, we consider those clauses not motivated by the desire to preclude transfers as motivated primarily by the changing structure of the venture. The investments ia and ib are made by the partners for the purpose of increasing the value of the venture to the public markets or to a trade buyer. For example, ia and ib may increase the value of the synergies available to a trade buyer or they may make the venture more `transparent' to in- 6 vestors, thereby increasing its valuation in public markets.9 The ¯rst best investments are precluded by the fact that neither party can be the unique residual claimant to the value created by the investment he has made.10 The investments ja and jb are made by the partners to enable one partner to buy out the other, at a price Pa paid by partner a to partner b or a price Pb paid by b to a. For example, if the joint venture is motivated by the complementarity of skills of the two partners, with one partner providing technical knowhow and the other marketing skills, ja and jb represent the investments made by the technical partner in acquiring marketing skills and by the marketing partner in acquiring technical knowhow, respectively. The acquisition of one partner's stake by the other can be a source of value because it removes the problem of double moral hazard for any further investment that might be made. All investments have unit marginal cost. Following the realization of the state, any party that has a stake in the venture can transfer value from the venture to itself. We let xa , xb , and xts denote the amounts transferred by partners a and b and the trade buyer, respectively, and ba (xa ), bb (xb ), and bts (xts ) denote the bene¯ts of such transfers.11 For example, in case both a and b were to engage in transferring value from the joint venture in the state where the joint venture form is to For a discussion of the link between transparency and valuation, see Krishnaswami and Subramaniam (1999). 10 This requires the investments made by the two parties to be simultaneous. When investments are sequential, Noeldeke and Schmidt (1998) have shown that simple options can be used to make each party in succession the unique residual claimant. 11 The bene¯t functions ba (:), bb (:), and bts (:) are increasing and concave in their arguments. They have value zero at zero. 9 7 be maintained, the value of the joint venture would be: Vjv (xa ; xb ) + ba (xa ) + bb (xb ) The values Vipo (ia ; ib ; xa ; xb ), Vts (ia ; ib ; xts ; xb ), Vts (ia ; ib ; xa ; xts ), Va (ja ; xa ), and Vb (jb ; xb ) are similarly de¯ned.12 Our notation implies that Vjv ´ Vjv (0; 0), Vipo (ia ; ib ) ´ Vipo (ia; ib ; 0; 0), Vts (ia ; ib ) ´ Vts (ia ; ib ; 0; 0), Va (ja ) ´ Va (ja; 0), and Vb (jb ) ´ Vb (jb ; 0). We assume all transfers are value destroying: ¯ ¯ ¯ ¯ @Vr (:; :; :; xt) @Vr (:; :; xs ; :) 0 0 < 0 and <0 + bs (xs )¯ + bt (xt )¯ ¯ ¯ @xs @xt xs =0 xt =0 where r 2 fjv; ipo; ts; a; bg, s 2 fa; tsg, and t 2 fb; tsg.13 No transfer therefore takes place after one partner has bought out the other, or after an IPO or a trade sale in which both partners sold their entire stakes. Notwithstanding the value destroyed by transfers, a partner may have an incentive to engage in a transfer when the cost of the transfer is borne in part by its partner in the venture. For example, if we denote partner a's stake in the joint venture by ° a, 0 < ° a < 1, and consider the state where the joint venture form is to be maintained, it is possible that: ¯ ¯ @Vjv (:; :; xa ; :) 0 °a + ba (xa)¯ >0 ¯ @xa xa =0 All values are decreasing and concave in the amounts transferred. Note that in the states where one partner buys out the other, only the remaining partner, now sole owner of the venture, can engage in a transfer. 13 Note that the function Vr has only two arguments in case r 2 fjv; a; bg. 12 8 In the absence of any transfers and renegotiation, ° a is chosen such as to: max pipo Vipo (ia ; ib ) + pts Vts (ia ; ib ) + pa Va (ja ) + pb Vb (jb ) + pjv Vjv °a (1) subject to: ´ i ´ ³ h ³ i i ° a pipo Vipo ba; ib + pts Vts ba ; ib + pjv Vjv +pa [Va (ja ) ¡ Pa ] + pb Pb ¡ ba i ia = arg max ba i ib = arg max bb i h ³ ´ ³ ´ i (1 ¡ ° a ) pipo Vipo ia ; bb + pts Vts ia ; bb + pjv Vjv i i +pa Pa + pb [Vb (jb ) ¡ Pb ] ¡ bb i h ³ ´ i ba ¡ Pa + pb Pb ¡ ba Va j j ja = arg maxpa and: ba j h ³ ´ i jb = arg maxpaPa + pb Vb bb ¡ Pb ¡ bb j j We now show how the various clauses we have presented can serve to preclude transfers and renegotiation and to induce the ¯rst best investments ¤ ¤ ja and jb , where: bb j ³ ´ ¤ j j ja = arg maxpa Va ba ¡ ba ba j (2) 9 and: ³ ´ ¤ j j jb = arg maxpb Vb bb ¡ bb bb j (3) 3 Results under certainty We analyze the various clauses in turn. 3.1 Put options We limit our analysis of put options to the state in which the joint venture form is to be maintained. This is because, as noted above, no transfer takes place after one partner has bought out the other, or after a trade sale or an IPO in which both partners sold their entire stakes.14 In such a state, a put option that grants partner a the right to sell his stake to partner b at a strike price ° a Vjv and one that grants b the right to sell his stake to a at a strike price (1 ¡ ° a ) Vjv can be shown to deter any attempt at transferring value on the part of either partner. The partner's expectation of having to pay fair value for the other partner's stake will deter the former partner from attempting any transfer, for it will negate any bene¯t from such transfer. To see this formally, consider an attempt by partner a to transfer an 14 We shall argue in Sections 3.3 and 3.6 that the partners will always sell their entire stakes in a trade sale or an IPO. 10 amount: xa = arg max° a Vjv (ba ; xb ) + ba (ba ) x x where: xa b Partner a would do so only if: xb = arg max (1 ¡ ° a) Vjv (xa ; xb ) + bb (bb ) b x xb b ° a Vjv (xa; xb ) + ba (xa ) > ° a Vjv (4) for he would otherwise prefer to exercise his put option. However, a's attempted transfer would prompt b to exercise his option to put his stake in the venture to a at a price (1 ¡ ° a ) Vjv , for equation (4) combines with the undesirability of transfers to imply:15 (1 ¡ ° a ) Vjv > (1 ¡ ° a ) Vjv (xa ; xb ) + bb (xb ) 15 The undesirability of transfers implies: Vjv > Vjv (xa ; xb ) + ba (xa ) + bb (xb ) 11 Partner b's exercise of his put option would make a's payo®:16 ° a Vjv (xa; 0) + ba (xa ) + (1 ¡ ° a ) Vjv (xa; 0) ¡ (1 ¡ ° a ) Vjv = Vjv (xa ; 0) + ba (xa ) ¡ (1 ¡ ° a ) Vjv From the undesirability of transfers, it is clear that partner a would set xa = 0. In other words, the expectation that b will exercise his put option in case of attempted transfer on the part of a serves to deter the latter partner from engaging in such transfer. The preceding reasoning holds equally for any attempted transfer on the part of partner b. 3.2 Call options We start our analysis of call options by considering the case where pa > 0 = pb : whilst there is a state in which partner a should buy out partner b, there is no state in which the converse is true. In this case, a call option that grants a the right to acquire b's stake at a strike price Pa = (1 ¡ ° a ) Vjv will induce is desirable. Partner a will be precluded from exercising the call option prior to a trade sale or an IPO by the tag-along provision and the IPO clause, respectively (see Sections 3.4 and 3.6). He will have no interest in exercising the call option in the state where the joint venture is to be maintained, for he would Note that partner b's exercise of his put option will preclude him from engaging in any transfer. 16 ¤ a to make the ¯rst best investment ja and to buy out b when such buy out 12 then buy at a price (1 ¡ ° a) Vjv a stake (1 ¡ ° a ) in a venture that will have value Va (ja ) < Vjv following his acquisition.17 He will make the ¯rst best ¤ investment ja for his payo® will be (neglecting the states in which he does not exercise the call option): pa [° a Va (ja ) + (1 ¡ ° a ) Va (ja ) ¡ (1 ¡ ° a ) Vjv ] ¡ ja = pa [Va (ja ) ¡ (1 ¡ ° a ) Vjv ] ¡ ja ¤ This has solution ja from the comparison of the preceding equation with equation (2). Now consider the case where both pa and pb are strictly positive. Denote sa the state in which partner a should buy out partner b; sb that in which the converse is true; Va (:; s) the value in state s of the venture when bought out by a; and Vb (:; s) the value in state s of the venture when bought out by b; s = sa ; sb . We assume that Vr (:; s) > Vjv for r = a; b and s = sa ; sb . The buy out of one partner by the other partner is preferable to the continuation of the venture in states sa and sb irrespective of the identity of the buying partner. As we shall see below, this ensures that neither partner exercises his put option in states sa and sb . ¤ a and a call option of strike price Pb = ° a Va (ja ; sb ) to partner b will achieve ¤ Granting a call option of strike price Pa = (1 ¡ ° a ) Vb (jb ; sa ) to partner the desired e±ciency, for a will wish to exercise his option in state sa as ¤ Va (ja; sa ) > Vb (jb ; sa ), whereas b will wish to exercise his option in state sb The state in which the joint venture form is to be maintained is one in which the value of the joint venture Vjv is larger than the values of all alternatives. 17 13 ¤ as Vb (jb ; sb ) > Va (ja ; sb ). Partner a will then have payo®:18 ¤ pa [° a Va (ja; sa ) + (1 ¡ ° a ) Va (ja ; sa ) ¡ (1 ¡ ° a) Vb (jb ; sa )] ¤ +pb ° a Va (ja ; sb ) ¡ ja ¤ ¤ = pa [Va (ja ; sa ) ¡ (1 ¡ ° a ) Vb (jb ; sa )] + pb ° a Va (ja ; sb ) ¡ ja ¤ This has solution ja from the comparison of the preceding equation with equation (2). The same can be shown of partner b. Nonetheless, a problem will arise from the desire of both partners to ¤ ¤ exercise their options in state sa if Vb (jb ; sa ) > Va (ja ; sb ) or to do so in state ¤ ¤ sb if Va (ja ; sb ) > Vb (jb ; sa ). In the former case, partner b who has made the ¤ e±cient investment jb will wish to exercise his call option in state sa as his payo® from doing so is greater than his payo® from selling his stake to a at ¤ Pa = (1 ¡ ° a ) Vb (jb ; sa ). Formally: ¤ ¤ ¤ (1 ¡ ° a ) Vb (jb ; sa) + ° aVb (jb ; sa ) ¡ ° a Va (ja ; sb ) ¤ ¤ = Vb (jb ; sa ) ¡ ° a Va (ja ; sb ) ¤ ¤ > Vb (jb ; sa ) ¡ ° a Vb (jb ; sa ) ¤ = (1 ¡ ° a ) Vb (jb ; sa) = Pa ¤ ¤ where the inequality is true from Vb (jb ; sa ) > Va (ja ; sb ). In the latter case, ¤ partner a who has made the e±cient investment ja will wish to exercise his call option in state sb as his payo® from doing so is greater than his payo® 18 ¤ Note that partner b will not exercise his put option, as Vb (jb ; sa ) > Vjv . 14 ¤ from selling his stake to b at Pb = ° a Va (ja ; sb ). Formally: ¤ ¤ ¤ ° a Va (ja ; sb ) + (1 ¡ ° a) Va (ja ; sb ) ¡ (1 ¡ ° a ) Vb (jb ; sa ) ¤ ¤ = Va (ja ; sb ) ¡ (1 ¡ ° a ) Vb (jb ; sa ) ¤ ¤ > Va (ja ; sb ) ¡ (1 ¡ ° a ) Va (ja ; sb ) ¤ = ° a Va (ja ; sb ) = Pb ¤ ¤ where the inequality is true from Va (ja ; sb ) > Vb (jb ; sa). Note that only one ¤ ¤ of the two problems will arise, as the two inequalities Vb (jb ; sa ) > Va (ja ; sb ) ¤ ¤ and Va (ja ; sb ) > Vb (jb ; sa ) cannot simultaneously be true. The problem of simultaneous exercise of the two options can be avoided by giving one partner precedence over the other partner in exercising his option. ¤ ¤ Thus, in the case where Vb (jb ; sa) > Va (ja ; sb ), it is clear that precedence should be given to a; thereby denying b the exercise of his option in state sa . Partner a will not exercise his option in state sb , as his payo® from doing ¤ so is lower than his payo® from selling his stake to b at Pb = ° a Va (ja ; sb ). Formally: ¤ ¤ ¤ ° a Va (ja ; sb ) + (1 ¡ ° a) Va (ja ; sb ) ¡ (1 ¡ ° a ) Vb (jb ; sa ) ¤ ¤ = Va (ja ; sb ) ¡ (1 ¡ ° a ) Vb (jb ; sa ) ¤ ¤ < Va (ja ; sb ) ¡ (1 ¡ ° a ) Va (ja ; sb ) ¤ = ° a Va (ja ; sb ) = Pb 15 ¤ ¤ where the inequality is true from Vb (jb ; sa) > Va (ja ; sb ).19 In the case where ¤ ¤ Va (ja ; sb ) > Vb (jb ; sa ), precedence should be given to b, who can be shown not to exercise his option in state sa by the same reasoning as was used for a. We conclude this section by noting that, in accordance with our analysis, there is evidence in joint venture contracts of an `appropriate party' having precedence over the other party.20 3.3 Drag-along rights In case the venture should be sold to a trade buyer, it is clear that neither partner should be allowed to veto the sale. This is because a veto on the part of either partner would be nothing but a pretext for renegotiation, which would alter the partners' payo®s in that state from ° a Vts (ia ; ib ) and (1 ¡ ° a ) Vts (ia ; ib ) to: ° aVjv + ¯ a [Vts (ia ; ib ) ¡ Vjv ] and: (1 ¡ ° a ) Vjv + (1 ¡ ¯ a ) [Vts (ia ; ib ) ¡ Vjv ] Note that partner a may exercise his option in state b with the intention of negotiating the sale of the venture to partner b, but the irrevocable nature of many joint venture clauses makes such a sale unlikely (see Linklaters, Paines, and Nightingale, 1990, p. 187). 20 See Linklaters, Paines, and Nightingale (1990, p. 61): \... the appropriate party will have the right within a speci¯ed period of time to purchase all of the other's shares. Failure to exercise that right will entitle the other to purchase the former's shares." 19 16 where ¯ a denotes the bargaining power of partner a.21 Yet, it is the payo®s ° a Vts (ia ; ib ) and (1 ¡ ° a ) Vts (ia ; ib ) that maximize the value of the venture, subject to the constraints imposed by the problem of double moral hazard. Drag-along rights amount to more than the denial of the right of veto, as they mandate the sale of both partners' stakes to the trade buyer. Such an obligation may be made necessary by the desire to avoid the transfers that would occur if only one partner were to sell his stake to the trade buyer and ownership of the venture were shared between the trade buyer and the remaining partner. Of course, the desire for e±ciency would have the trade buyer buy out the remaining partner, but the resulting negotiation might again distort ex ante investments. Thus, if the remaining partner were partner b, who had bargaining power 1 ¡ ¯ ts when negotiating with the trade buyer, b's payo® would be: (1 ¡ ° a ) Vts (ia; ib ; xts ; xb ) + bb (xb ) + (1 ¡ ¯ ts ) [Vts (ia ; ib ) ¡ Vts (ia ; ib ; xts ; xb ) ¡ bts (xts ) ¡ bb (xb )] The preceding payo® is to be contrasted with the desired payo®: (1 ¡ ° a ) Vts (ia ; ib ) Note that the put options discussed in Section 3.1 would not deter the trade buyer and the remaining partner from engaging in value transfers, as the di®erence between the value of the venture once sold to the trade buyer 21 Note that the status quo consists in maintaining the joint venture form. 17 Vts (ia ; ib ) and the value of the venture implied by the strike prices of the put options Vjv would a®ord the parties some leeway to engage in such transfers. In particular, xts and xb would be such that: ° aVts (ia ; ib ; xts ; xb ) + bts (xts ) = ° a Vjv and: (1 ¡ ° a ) Vts (ia ; ib ; xts ; xb ) + bb (xb ) = (1 ¡ ° a) Vjv respectively. 3.4 Tag-along rights Consider the case where the value added by a trade sale consists exclusively of synergies received entirely by the trade buyer.22 Denote such synergies Sts (ia ; ib ) > 0, which are such that: Vts (ia; ib ) = Vjv + Sts (ia ; ib ) That the synergies are received entirely by the trade buyer implies that, should any partner in the venture fail or be unable to sell his stake to the trade buyer and therefore remain a partner in the venture now shared with the trade buyer, such partner would have a stake in a venture worth, to This assumption is made to simplify the exposition and is stronger than necessary. It is su±cient that part of the value added is not re°ected in the value of the venture after the trade sale. 22 18 him, Vjv < Vts (ia ; ib ). For example, if the remaining partner were partner b, his stake in the venture following the sale to the trade buyer would be (1 ¡ ° a ) Vjv < (1 ¡ ° a ) Vts (ia ; ib ). The lower payo® of a partner who fails to sells his stake implies that no partner will fail to sell his stake if able to do so. However, in the case where the trade buyer can receive the desired synergies simply by buying a majority stake in the venture, the minority partner may be unable to sell his stake as the trade buyer and the majority partner choose to divide the value added between them rather than share it with the minority partner. The expectation of such behavior would distort the minority partner's incentive to invest, as he would not pro¯t from the bene¯ts of his investment in the case of a trade sale. Tag-along rights, which essentially oblige the trade buyer to buy both the majority and the minority partners' stakes on the same terms, serve to reassure the minority partner that he will pro¯t from a trade sale in proportion to his stake in the venture and therefore ensure that no distortion in investment will take place. 3.5 Pre-emption rights Consider the case where one partner, say partner a, is indispensable to the success of the venture.23 Partner a will therefore be prohibited from withdrawing from the venture and will be denied the put options and tag-along rights discussed in Sections 3.1 and 3.4.24 This implies that pipo = pts = pb = 0. See Linklaters, Paines, and Nightingale (1990, p. 59): \If one party possesses critical skills and facilities, the withdrawal of that party in the early years of the joint venture would e®ectively mean dissolution and an absolute prohibition may be appropriate." 24 23 19 In the case where the joint venture form is to be maintained, the restrictions on a will expose him to transfers initiated by b or by a trade buyer who, perhaps being in a position to make the same contribution as b to the venture yet having a greater incentive than b to transfer value from the venture, b0ts (x) > b0b (x) 8x, may acquire b's stake in a trade sale exclusively for the purpose of transferring value from the venture. Pre-emption rights, which oblige b to o®er his stake to a at a price (1 ¡ ° a ) Vjv in case b should wish to sell his stake in the venture, can be viewed as serving to preclude any such trade sale. Pre-emption rights thereby avoid the threatened destruction of value.25 Note that a's pre-emption rights need not expose b to expropriation by a, for b retains the put options discussed in Section 3.1. 3.6 Initial public o®ering clauses The issues that arise when the venture is to be taken public are very similar to those that arise when it is to be sold to a trade buyer in a trade sale: should one or both partners have the right to veto the IPO or be allowed to retain a controlling stake in the now public venture, the resulting renegotiation would distort the partners' investment.26 Both partners must therefore be denied the right to veto the IPO and both must be made to sell their stakes in the IPO. IPO clauses serve to achieve the desired results, by having the partners agree in advance the circumstances in which the venture will be taken public, Pre-emption rights do not, however, preclude the transfers initiated by b. Whilst such transfers can be precluded by having a buy out b, such a solution will not be applicable to the state in which the joint venture form is to be maintained. 26 The formal proof is identical to that in Section 3.3, with Vts replaced by Vipo . 25 20 with both partners selling their stakes.27 4 Results under uncertainty e e e e e We now consider the case where Vipo (:; :), Vts (:; :), Va (:), Vb (:), and Vjv are and V r its expectation conditional on the realization of the relevant state. ° a remains the solution to (1), with Vr now replaced by V r . We note that drag-along and tag-along rights and initial public o®ering e e clauses are not a®ected by the randomness of the values Vts and Vipo . Dragvalue of the venture in that they require the partners to specify in advance the conditions in which they will take the joint venture public, but such reference can be avoided by extending drag-along rights from trade sales to IPOs. Put and call options and pre-emption rights involve a direct reference to the value of the venture. The manner in which they are structured must therefore change with the introduction of uncertainty. Consider partner b's put option. It is clear that a ¯xed strike price such as (1 ¡ ° a ) V jv for example will induce b to exercise his put option when Vjv < V jv even if a has attempted no transfer. Such opportunism can be avoided by setting the strike price of the option equal to the value of b's stake This in e®ect assumes that the state in which the venture is to be taken public is contractible. 27 e random variables. We denote Vr , r 2 fipo; ts; a; b; jvg, the realization of Vr along and tag-along rights involve no direct reference to the value of the venture. Initial public o®ering clauses involve an indirect reference to the 21 in the venture, (1 ¡ ° a ) Vjv , for b will then exercise his put option only when a attempts a transfer. This requires the valuation of the venture, which can be based on some valuation formula or performed by some impartial external expert. The valuation need not result in the exact value Vjv , but it must be unbiased: the strike price of a partner's put option must equal the value of the partner's stake in the venture in expectation. One or the other partner would otherwise have an incentive to exercise his put option. The valuation must be carried out before the attempted transfer is expected to take place, for it would otherwise result in the value Vjv (xa ; xb ) < Vjv which would fail to deter the transfer. An alternative to having the venture valued by some valuation formula or some impartial external expert is to have the valuation done by partner b himself. The obvious scope for abuse this provides b with can be circumvented through the use of the Russian roulette clause: by giving a the right to sell out to b at the price put forward by b, the Russian roulette clause ensures that b has no incentive to put forward a value greater than Vjv . He has no incentive to put forward a value smaller than Vjv , for the value he puts forward is the strike price of his put option.28 The advantage of the Russian roulette clause over the valuation formula or the external expert is that it avoids the greater noise these two methods are likely to introduce in the valuation. As is required of these two methods, the Russian roulette clause requires the valuation to be carried out before the attempted transfer is expected to take 28 The combined e®ect of the put option and the Russian roulette clause is to make partner b long a put option and short a call option. The two options share the same strike price, which is set by b himself. 22 place.29 Turning to call options, we note that the very advantage of a valuation formula, an external expert, or a Russian roulette clause in the case of put options is a disadvantage in that of call options. Setting the strike price of a partner's call option equal to the value of the other partner's stake in the venture makes the price depend on the investment made by the second partner. This implies that the second partner is no longer the unique residual claimant to the investment he has made. It thereby distorts the second partner's investment. To see this, consider the case where the Russian roulette clause is used and partner a has precedence over partner b in exercising his call option. Partner a will set strike prices (1 ¡ ° a) Vb (jb ; sa ) and (1 ¡ ° a ) Vb (jb ; sb ) in states sa and sb , respectively, thereby appropriating to himself the entire value that results from the acquisition of b's stake in state sa and the sale of his stake to b in state sb . His payo® will be (neglecting the states in which he does not exercise the call option): pa [° a Va (ja ; sa ) + (1 ¡ ° a ) Va (ja ; sa ) ¡ (1 ¡ ° a ) Vb (jb ; sa )] +pb ° a Vb (jb ; sb ) ¡ ja = pa [Va (jb ; sa ) ¡ (1 ¡ ° a ) Vb (jb ; sa )] + pb ° aVb (jb ; sb ) ¡ ja An expectations damage clause (see for example Che and Chung, 1999), which would compensate partner b for the transfer e®ected by partner a, would achieve the desired result. Perhaps surprisingly, such clauses appear to be rarely used in joint venture contracts. 29 23 whereas partner b's payo® will be: pa (1 ¡ ° a ) Vb (jb ; sa ) +pb [° a Vb (jb ; sb ) + (1 ¡ ° a ) Vb (jb ; sb ) ¡ ° a Vb (jb ; sb )] ¡ jb = pa (1 ¡ ° a ) Vb (jb ; sa ) + pb (1 ¡ ° a ) Vb (jb ; sb ) ¡ jb The comparison of the preceding equations with equations (2) and (3) ¤ ¤ ¤ reveals that ja = ja and jb 6= jb . Similarly, we can show that ja 6= ja ¤ and jb = jb in the case where partner b has precedence over partner a. Interestingly, the Russian roulette clause appears to be used primarily for put options (see Linklaters, Paine, and Nightingale, 1990, pp. 186-189). The ¯rst best can nonetheless be achieved by setting:30 h i e ¤ Pa = (1 ¡ ° a) Max Vb (jb ; sa) h i e ¤ Pb = ° a Max Va (ja ; sb ) and: and granting a precedence over b when:31 h Note that the maxima exist, as Vb (:; sa ) < Va (:; sa ) by de¯nition of sa as the state in which the venture should be acquired by partner a; and Va (:; sb ) < Vb (:; sb ) by de¯nition of sb as the state in which the venture should be acquired by partner b. 31 Note that the two inequalities cannot simultaneously be true. 30 i h i ¤ eb (jb ; sa ) > Max Va (ja ; sb ) e ¤ M ax V 24 and b precedence over a when: h i h i e ¤ e ¤ M ax Va (ja ; sb ) > Max Vb (jb ; sa ) ¤ ¤ The achievement of the ¯rst best investments ja and jb can be shown in a manner similar to that in Section 3.2. Finally, we note that the valuation formula, external expert, or Russian roulette clause used for put options can also be used for pre-emption rights. 5 Conclusion We conclude this paper by noting that many of the clauses found in joint venture contracts can also be found in venture capital contracts, with either the same name (call options, drag-along rights) or slightly di®erent names (liquidity rights in place of put options, buy-sell agreements in place of the Russian roulette clause, co-sale rights in place of tag-along rights, ¯rst-refusal rights in place of pre-emption rights).32 The similarity between joint venture and venture capital contracts should not be surprising, for a start-up can, in many ways, be viewed as a joint venture between an entrepreneur and a venture capitalist. The former typically provides human capital and technical expertise, the latter commercial and ¯nancial knowhow. A venture capital investment can also plausibly be characterized by a double-sided moral hazard. A start-up is similar to a joint venture in that it can be sold to a trade buyer or taken public, but di®ers For information on venture capital contracts, see for example The Growth Company Guide (www.growco.com). 32 25 from a joint venture in that one partner rarely buys out the other and it is frequently liquidated. Uncertainty also appears to be greater in the case of start-ups, requiring dynamic adjustments in the partners' stakes (Gompers, 1995; Kaplan and StrÄmberg, 1999). o To illustrate the similarity, consider the case (discussed in Section 3.5) where one partner is indispensable to the success of the venture. We argued that such a partner should be denied put options and tag-along rights. In a venture capital setting, the entrepreneur can frequently be thought of as controlling some asset, say human capital, that is hard to codify and disembody. This in turn may make him indispensable to the venture, at least in the early phases of the venture. We would predict, therefore, that the entrepreneur should be denied put option and tag-along rights for as long as he is indispensable to the venture. Similarly, consider our argument in Section 3.5 that pre-emption rights can protect one partner against a sale of the other partner's stake in the venture to a trade buyer who transfers value from the venture. In a venture capital context, it seems plausible that a trade buyer could transfer value (perhaps by shutting the venture down). We would predict, therefore, that the entrepreneur should have pre-emption or veto rights with respect to trade sales. 6 References Aghion, P. and J. Tirole, 1994, The management of innovation, Quarterly Journal of Economics 109, 1184-1209 Allen, J. and G.M. Phillips, 2000, Corporate equity ownership, strategic 26 alliances, and product market relationships, Journal of Finance 55, 27912815. Balakrishnan, S. and M.P. Koza, 1993, Information asymmetry, adverse selection, and joint ventures: Theory and evidence, Journal of Economic Behavior and Organization 20, 99-117. Bhattacharyya, S. and F. Lafontaine, 1995, Double-sided moral hazard and the nature of share contracts, Rand Journal of Economics 26, 761-781. Campbell, E. and J. Reuer, 2000, Note on the legal negotiation of strategic alliance agreements, Insead Note 300-054-6. Chan, S.H., J.W. Kensinger, A.J. Keown, and J.D. Martin, 1997, Do strategic alliances create value? Journal of Financial Economics 46, 199221. Che, Y.-K. and T.-Y. Chung, 1999, Contract damages and cooperative investments, Rand Journal of Economics 30, 84-105. Crawford, V.P. and W.P. Heller, 1979, Fair division with indivisible commodities, Journal of Economic Theory 21, 10-27. Darrough, M.N. and N.M. Stoughton, 1989, A bargaining approach to pro¯t sharing in joint ventures, Journal of Business 62, 237-270. Doz, Y.L. and G. Hamel, 1998, Alliance Advantage: The Art of Creating Value through Partnering, Harvard Business School Press, Boston, MA. Gomes-Casseres, B., 1999, The Alliance Revolution, Harvard University Press, Cambridge, MA. Gompers, P., 1995, Optimal investment, monitoring, and the staging of venture capital, Journal of Finance 50, 1461-1490. Gulati, R., 1995, Does familiarity breed trust? The implications of re- 27 peated ties for contractual choice in alliances, Academy of Management Journal 38, 85-112. Hennart, J.-F., 1988, A transaction cost theory of equity joint ventures, Strategic Management Journal 9, 361-374. Hennart, J.-F. and S. Reddy, 1997, The choice between mergers/acquisitions and joint ventures: The case of Japanese investors in the United States, Strategic Management Journal 18, 1-12. Kaplan, S.N. and P. StrÄmberg, 1999, Financial contracting theory meets o the real world: An empirical analysis of venture capital contracts, working paper, University of Chicago. Kogut, B., 1988, Joint ventures: Theoretical and empirical perspectives, Strategic Management Journal 9, 319-322. Krishnaswami, S. and V. Subramaniam, 1999, Information asymmetry, valuation, and the corporate spin-o® decision, Journal of Financial Economics 53, 73-112. Lerner, J. and A. Tsai, 1999, Do equity ¯nancing cycles matter: Evidence from biotechnology alliances, working paper, Harvard Business School. Linklaters and Paine, with C. Nightingale, 1990, Joint Ventures, Longman, London. Lorange, P. and J. Roos, 1992, Strategic Alliances: Formation, Implementation, and Evolution, Basil Blackwell, Oxford. McAfee, R.P., 1992, Amicable divorce: Dissolving a partnership with simple mechanisms, Journal of Economic Theory 56, 266-293. McConnell, J.J. and T.J. Nantell, 1985, Corporate combinations and common stock returns: The case of joint ventures, Journal of Finance 40, 519- 28 536. Mohanram, P. and A. Nanda, 1998, When do joint ventures create value? working paper, Harvard Business School. Noeldeke, G. and K.M. Schmidt, 1998, Sequential investments and options to own, Rand Journal of Economics 29, 633-653. Oxley, J.E., 1997, Appropriability hazards and governance in strategic alliances: A transaction cost approach, Journal of Law, Economics, and Organization 13, 387-409. Pisano, G., 1989, Using equity participation to support exchange: Evidence from the biotechnology industry, Journal of Law, Economics, and Organization 5, 109-126. Reich, R.D. and E. Mankin, 1986, Joint ventures with Japan give away our future, Harvard Business Review, 78-86. Reuer, J.J. and M.P. Koza, 2000, Asymmetric information and joint venture performance: Theory and evidence of domestic and international joint ventures, Strategic Management Journal 21, 81-88. Rey, P. and J. Tirole, 1998, Divergence of objectives and the governance of joint ventures, working paper, IDEI. Yoshino, M.Y. and U.S. Rangan, 1995, Strategic Alliances, Harvard Business School Press, Boston, MA. 29