Termination clauses in partnerships

Abstract

We show that when designing a partnership agreement partner firms may prefer not to specify how to allocate the commonly owned assets should there be an early termination of the contract. By not including such a clause, firms induce litigation before a Court with positive probability. Firms create this ex-post inefficiency in order to increase the levels of non-contractible investments, i.e. increase the ex-ante efficiency. The absence of an asset allocation clause works as a “discipline device” that mitigates the hold-up problem within the partnership. In our set-up, no other contract but that without an asset allocation clause can credibly create an ex-post inefficiency.

Introduction

Strategic alliances, in the form of joint ventures (JVs) or looser modes of cooperation, are an increasingly common solution in response to the need to reduce start-up costs, share risks, enter new markets or develop new technologies. According to Dyer et al. (2001) the top 500 global businesses have an average of 60 major strategic alliances each. During the 1990s, the number of alliances grew at an annual rate of over 25% in the leading industrial nations and about 20% of the revenue of the largest US and European corporations comes from partnerships (see Contractor and Lorange, 2002, Harbison et al., 2000).

Even though the potential advantages of partnering are well known, the track record for joint ventures is not a glowing one. Instability is a commonly recognized problem affecting strategic alliances and the average life span of a JV is as little as four years (seven years for other studies) with a failure rate ranging between 50% and 70%.¹ Because of these prospects, partners should be aware of the difficulties they may encounter when managing an alliance and of the possibility of its early termination, when setting up a new relation. According to some commentators, partners should approach JVs as Hollywood marriages; they should plan their termination strategy from the very beginning by specifying in the initial agreement “what happens to assets, customers and existing contracts in the (likely) event of a break-up”.² Indeed, as is well documented in business literature, a non-amicable termination of an alliance may result in very long negotiations, large expenses and bitter legal battles.³

Surprisingly, JV participants devote relatively little attention to predicting what happens in case of termination of the alliance. A PricewaterhouseCoopers (2000) survey shows that less than half of the firms entering an alliance have a formal exit strategy. Similarly, several authors have observed that, of the many aspects of alliance management, planning its termination ranks among the most ignored by partners.⁴ Obviously, there are probably various reasons for such a lack of attention. Just as a pre-nuptial agreement, discussing a termination clause when forming the alliance might sour the deal; it might reveal a lack of trust among partners. Also difficulties in working out all the possible contingencies that might occur and designing what parties should do in these cases may justify the absence of a termination clause in a JV contract. A possible alternative explanation for such an absence can be envisaged in the case of Concert. When negotiating the terms of their joint venture (called Concert), British Telecommunications and AT&T explicitly decided not to include a termination clause. By not determining the rules for separation, partners wanted to demonstrate their commitment into the relationship.⁵ The model we present develops formally this idea.

We consider two firms setting up a joint venture to pursue a project.⁶ The project can succeed or fail with probabilities which depend on the investment levels chosen by partners. In case the project fails, firms terminate the partnership and decide upon the allocation of the assets belonging to the JV. If the JV-contract is silent about asset allocation, partners bargain just after terminating the partnership about the assets ownership and the related payments. If they are unable to reach an agreement then the case comes up before a Court, which takes the final decision about assets allocation. Litigation is costly due to the related legal expenses. We show that litigating with positive probability is an equilibrium strategy for partners.⁷ They could avoid litigation before the Court by simply including an asset allocation clause in their JV-contract. However, they benefit from not including this clause, under reasonable conditions. By not including it, firms worsen their own prospects in the event of failure of the project: not only do they not succeed in pursuing it but they also waste resources litigating. But this induces them to increase their investments in order to lower the probability of failure, thus mitigating the hold-up problem when investments are non-contractible.

Litigation before the Court occurs with positive probability because of asymmetric information. Following the argument put forward by several authors,⁸ we assume that partners are asymmetrically informed about the private value of assets. Namely, if the partnership fails, we assume that assets are (more) valuable for one firm which knows its exact private value, while the other knows only that its own valuation is lower. The attempt of the former firm to appropriate a larger part of the assets value during the bargaining stage induces the latter to reject an amicable settlement with positive probability so that firms resort to a costly outside option, the Court, in order to take a decision.

Review of the relevant literature: There are different strands of economic literature that are related to our paper. A relatively recent series of studies stemming from the paper by Cramton et al. (1987) focuses on partnership dissolution. There are two main issues tackled: (i) Under what conditions is there efficient partnership dissolution (i.e. dissolve it when it is efficient to do so and assign the assets to the partner that evaluates them the most)?⁹ (ii) What are the relative merits of commonly used dissolution clauses such as the so-called Texas-shootout?¹⁰ Our paper departs from this literature quite substantially. We consider the relationship between investment and termination decisions, while the literature on partnership dissolution focuses exclusively on the latter decision.¹¹ Said differently, we study the effect of different termination procedures more on the ex-ante efficiency and less on the ex-post efficiency. Ex-post inefficiency (in our paper, litigation before the Court) generated by the absence of a termination clause might be beneficial in order to improve ex-ante efficiency (in our paper, to induce larger investments). The idea that there is a trade-off between ex-ante and ex-post efficiency is similar to the one presented in quite different contexts by Bordignon and Brusco (2001) and Gerardi and Yariv (2008). In the former paper, the authors show that the lack of exit rules in federal constitutions can be a commitment device; high costs of secessions (secessions are possible only by “independence wars”) increase the stability of the federation, and therefore the ex-ante benefits of joining it. Gerardi and Yariv (2008) present a mechanism design model with a principal and a committee of agents/experts who collect some (privately costly) information regarding a pay-off relevant variable. The design problem rests on the determination of the size of the committee and on a rule for aggregating agents’ information in order to come up with a decision. The authors show that an imperfect aggregation of the available information (an ex-post inefficiency) is useful in order to stimulate more agents to collect information in the first place (increased ex-ante efficiency).

The problem of how economic institutions and contracts should be designed in order to provide correct investment incentives to parties has been extensively studied by the hold-up literature. In a typical hold-up model, two parties involved in a transaction are required to make some relationship-specific investment and contract incompleteness forces them to rely on ex-post bargaining. The crucial difference between our paper and most of this literature relates to the informational structure at the bargaining stage; the vast majority of these papers assumes that bargaining takes place in a context of symmetric information and therefore yields to outcomes that are ex-post efficient. We show that, in a context where other solutions of the hold-up problem are not effective,¹² the presence of asymmetric information increases the investment incentives of firms. The study of hold-up problems with asymmetric information at the bargaining stage is relatively recent. Lau (2008) analyzes the role of asymmetric information in a standard buyer–seller relationship where only the former party makes trade-specific investments. She shows that the joint profits of the two parties increase with the probability that buyer and seller are asymmetrically informed about the investment chosen by the former party. An increase in the probability that parties are asymmetrically informed enhances the rents of the buyer thus increasing her/his investment incentives. This incentive effect dominates the negative effect due to failure to trade because of asymmetry of information. In a similar setting, Gul (2001) shows that if the investment of the buyer is unobservable and the time between offers is small, then the hold-up problem disappears. In these papers, asymmetric information lessens the hold-up since it decreases the seller's ability to expropriate the efficient investment made ex-ante by the buyer, at the bargaining stage. In our paper, which deals with two-sided hold-up problems, asymmetric information at the bargaining stage and the possibility of resorting to an outside option induce an ex-post inefficiency which turns out to be ex-ante beneficial.

Finally, our paper is also related to the stream of literature which takes into consideration strategic reasons for contract incompleteness. Non-contigent contracts as a signaling/screening device are analyzed in Aghion and Bolton (1987), Diamond (1993), Hermalin (2002), and Spier (1992). Bernheim and Whinston (1998) show that contracts that contain some “gaps” may help in establishing the appropriate incentives for parties. In a context where certain actions are observable by parties but not verifiable by Courts, incomplete contracts that expand the set of discretionary choices/strategies may be used in order to induce parties to coordinate on Pareto superior equilibria.

Outline of the paper: In Section 2, we describe the set-up of the model and we develop the benchmark. In Section 3, we derive the main results of the paper focussing on a comparison between two specific types of contracts: the Texas-shootout and one without an asset allocation (the one we wish to analyze). In particular we derive under which condition the latter is preferable to the former. Section 4 is devoted to showing that the results of Section 3 hold also when considering more general settings and alternative contracts. Finally, in the concluding Section 5 we relate the results shown in the paper to some of the stylized evidence on strategic alliances. All the proofs that are not essential for an understanding of the main arguments of the paper are presented in the appendix.