Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
THE EFFECT OF BARGAINING POWER ON CONTRACT DESIGN Author(s): Albert Choi and George Triantis Source: Virginia Law Review, Vol. 98, No. 8 (December 2012), pp. 1665-1743 Published by: Virginia Law Review Stable URL: http://www.jstor.org/stable/23333519 . Accessed: 12/02/2014 14:25 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp . JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. . Virginia Law Review is collaborating with JSTOR to digitize, preserve and extend access to Virginia Law Review. http://www.jstor.org This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions VIRGINIA Volume 98 LAW REVIEW December 2012 Number 8 ARTICLES THE EFFECT OF BARGAINING POWER ON CONTRACT Albert Choi and George DESIGN Triantis* Over the past forty years, an irrelevance proposition has been preva lent in law-and-economics scholarship: bargaining power should affect and not terms a contract. In contrast, practition only price nonprice of ers and commentators in industry regularly invoke bargaining power to static and dynamic variance in nonprice contract terms. This Ar explain ticle unpacks and analyzes the assumptions weak— of the strong—and this irrelevance to of bargaining power proposition bridge the gap between theory and the real world. In the first half of the Article, we identify and discuss a variety of explanations for the effect of bargaining versions on contract design. These include the effects of shifts in market supply and demand and the effect of negotiating price first and nonprice terms later. In the second half of the Article, we present an in-depth ex power amination the impact of bargain of one set of explanations, concerning and information asymmetry on nonprice terms, when the val ing power ue and cost of nonprice * Albert tricia Kowal C. BeVier Professor Belfer terms vary across contracting parties. In the Research Professor of Law, University of Virginia, and James and Pa of Law, Stanford University, The authors thank Isaac respectively. and Neal Sangal (Stanford Law 2014) for their valuable research 2011) Law (Harvard assistance. They are also grateiul for helpful Andrew Ronald Richard Gilson, Daughety, comments from Ian Ayres, Richard Craswell, Michael Hynes, Kessler, Avery Katz, Daniel Mitchell Alan Schwartz, Robert Scott, Klausner, Bentley McLeod, Jay Mitchell, Polinsky, and Kathy Spier as well as participants at law school at Berkeley, Columbia, workshops and Virginia, and at the 2012 meetings of the American Harvard, Stanford, Texas, USC, Law and Economics Association and the Theoretical Law and Economics Conference at Washington Law, in St. Louis. Triantis thanks University Law School and Business at Harvard for financial support. the John M. Olin Economics, 1665 This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions Center for 1666 Virginia Law Review [Vol. 98:1665 extreme cases gaining surplus in which one or the other party enjoys overwhelming bar the that to a share power, efforts of party capture larger of the the efficiency of the by screening or signaling may compromise nonprice terms. when bargaining We show that this incentive disappears or is mitigated is more shared between the power evenly parties: for when the example, when a monopolist faces the threat of competition, can renegotiate, or when they engage in bilateral parties bargaining with more even bargaining a power. As a whole, the Article provides theoretical basis for interpreting the intuition among market partici that the impact of bargaining power extends beyond price terms. we Before concluding, briefly suggest implications for legal policy, par ticularly the contract law doctrine of unconscionability. pants Introduction I. What II. How A. B. C. D. 1666 Is Bargaining Bargaining Power? Power 1674 Affects Contract Design The Bargaining Power Irrelevance Propositions Power Can Alter the Optimal Nonprice Bargaining Power Can Lead to Inefficient Nonprice Bargaining Bargaining Power in Two-Staged 1677 1678 Terms Terms.... (Price-First) Under Monopoly and Perfect Competition 1696 of Uneven Bargaining Power Under Asymmetric Information 1701 A. The Irrelevance B. Private V. Effect 1686 1690 Negotiations iii. Product Quality IV. Effect 1680 Under Complete Proposition on Buyer Type Information 1. Dominant Seller 2. Dominant Buyer Bargaining of More Even Information 1704 1705 1706 1709 Power 1712 A. Threat of Competition 1713 B. Renegotiation Bilateral Negotiation 1719 C. VI. Bargaining Power, Discrimination, 1723 and Legal Policy 1732 Conclusion Appendix: 1728 A General Model This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1734 The Effectof Bargaining Power 2012] 1667 Introduction WHEN power two parties enter into a contract, their relative bargaining of bar affects the terms of their deal. While the allocation gaining power clearly determines price, it is an open question whether and how it also affects nonprice terms (what we are alternatively refer It is common for practitioners and industry ring to as "contract design"). one-sided seemingly nonprice terms to unequal bargaining power and to explain changes in nonprice terms over time as a result of shifts in such power. Consider the following examples of such observations: observers to attribute in a sales of warranties and limitations on damages (1) Disclaimers contract are due to the power of the monopolist. (2) Broad termination rights are included in a merger or acquisition agreement when the acquirer has the power to dictate the terms of the agreement to the target company. (3) The purchase order forms of a large corporation, facing many po tential suppliers, insist that all litigation will be held in the courts of the purchaser's The objective state. of this Article is to begin a systematic analysis of how the agreement to such apparently determine bargaining power might one-sided terms. An important normative question is whether the efforts of the stronger party to appropriate a larger share of the surplus through the size of the surplus. Would a more equal these terms compromise sharing of bargaining power be more likely to lead to efficient (surplus contract provisions? maximizing) In legal scholarship, the issues of one-sided contract terms bear on the antitrust regulation under the doctrine tained of monopolies, as well as the policing of contracts of unconscionability. main Early legal scholarship that monopolists often used contracts of adhesion that contained one-sided terms.1 Law-and-economics bargaining 1 power scholars argued in response that affects price, but not other terms.2 The basic argument Contracts of Adhesion—Some About Freedom of Contract, Kessler, Thoughts L. Rev. 629, 632 (1943). This concern was reflected in the common law. See, v. Bloomfield Motors, 161 A.2d 69, 78-79 (N.J. I960). e.g., Henningsen ~ terms can be confusing The distinction between because price and nonprice nonprice 43 Friedrich Colum. terms allocate buyer's surplus surplus as well. Conditional and increase the seller's on price, a limited In addition, profit. warranty many will both decrease nonprice This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions terms, such the as Virginia Law Review 1668 [Vol. 98:1665 can be found in an early work by now-Judge Richard Posner, who ar would offer product quality monopolist gued that a profit-maximizing that efficiently meets buyer preferences, that is, improving quality until the incremental cost of farther improvement outweighs the incremental to the buyer.3 Thus, a monopolist producer of cars should find it in its self-interest to offer any warranty for which the buyer is willing to value than the cost to the producer, just as if it were a seller in a competitive market. The difference between a monopoly and a perfectly market, then, should be the market price, not the warranty competitive pay more terms offered. The argument that a monopolist extracts rather than continues to be the conventional quality price and collateral warranty "price" in law-and-economics.4 scholars the leading its rent through wisdom among clauses, that the promisor like price terms in the sense that they stipulate a on the occurrence of an event (or a to the promisee often look has to "pay" condition). 3 21 Stan. L. Rev. 548, 584-85 and Its Regulation, Richard A. Posner, Natural Monopoly of Law § 4.9, at 116 (7th ed. 2007). see also Richard A. Posner, Economic Analysis (1969); 4 of this conventional wisdom as irrelevance In Section II.A, we identify two versions that bargaining First, the strong-form version stands for the proposition power propositions. only affects bargaining inefficient statements tract Law in the weak-form terms. Second, and has no effect on nonprice version, terms, but the parties are no more likely to agree to may affect nonprice Various rather than equal, terms under unequal, bargaining power. nonprice are found in Robert E. Scott & Jody S. Kraus, Con of the conventional wisdom price power and Theory 58-60 104 G. Baird, The Boilerplate Puzzle, Douglas L. Priest, A Theory of the Consumer Product (4th ed. 2007); L. Rev. 933, 934, 938 90 Yale L.J. 1297, Mich. Warranty, stantive Unconscionability, (2006); 1320-21 George Alan Schwartz, A Reexamination ofNonsub (1981); . . three [weak] L. Rev. 1053, 1072-74 ("Given. (1977) of whether the the same level of product quality regardless 63 Va. a firm will produce assumptions, or a perfect competitor."); firm is a monopolist 113 Law, Theory and the Limits of Contract Schwartz & Robert E. Scott, Contract 541, 552-54 (2003) ("Bargaining of the surplus, which is determined in the division by the price the surplus, which the price the contract terms so as to maximize power instead is exercised term. Parties jointly choose [sic] Alan Yale L.J. & Louis L. Wilde, Product Quality and Alan Schwartz unequally."); that where 258 (1985) 52 Rev. Econ. Stud. 251, 251-52, Information, (arguing are imperfectly informed about product prices and quality levels offered by the sellers (that is, positive search costs), and where there are low fixed costs to provid may then divide Imperfect consumers various ing quality, a profit-maximizing competitiveprice). seller will offer at least the optimal quality, but at a supra that bargaining recognize power might affect nonprice terms in and Ian Ayres & Robert Gertner, Strategic Contractual Inefficiency Richard Craswell, Choice of Legal Rules, 101 Yale L.J. 729, 733 (1992); the Optimal Prop 60 U. Chi. L. and Related Rules in Unconscionability Doctrines, erty Rules and Liability and the Economic Rev. 1, 39^0 Jason Scott Johnston, Strategic Bargaining Theory (1993); A small some number circumstances. of Contract Default of articles See Rules, 100 Yale L.J. 615, 621-22, 625 (1990); see also This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions infra note 55. The Effectof Bargaining Power 2012] 1669 literature on the relationship between bargaining power and terms, such as warranties, is thin but mixed.5 Recently, in a Empirical nonprice Professor study of terms in end user licensing agreements ("EULAs"), found little evidence to support the hypothesis that Marotta-Wurgler market concentration causes terms to be more seller-friendly than they be in competitive markets.6 The study suggests, therefore, that market power alone should not be sufficient to trigger the scrutiny of a court under the doctrine of unconscionability.7 In contrast, Professors would Ben-Shahar and White examined auto-manufacturer supply contracts in nonprice terms, such as warranty and termina tion provisions, that seemed correlated with bargaining power.8 While factor in consumer reading and search costs may be a confounding and found variations standard-form in Marotta-Wurgler's contracts, such as the EULAs study, and White's it is sample pertinent to our analysis because business-to-business contracts where this factor is less likely to Ben-Shahar contains be an issue. Ben-Shahar In light of the conventional wisdom in law-and-economics, and White remarked that "given the enormous stakes, we that economic power would be used to dictate low prices, not expected selfish boilerplate."9 that the variation and the potential They speculated were due to internal inefficiency agency conflicts within the parties, but conceded ance that they did not "offer a satisfactory explanation for the vari across the different [original manufacturer equipment of terms 5 Practice Regarding Warranties in the See, e.g., George G. Bogert & Eli E. Fink, Business of Goods, 25 111. L. Rev. 400, 413-15 (1930) (stating that seller's bargaining power may lead to inadequate for buyers); Antonio Cabrales et al., Hidden Infor warranty protection Sale mation, 155-56 An Experiment, 14 Experimental Econ. 133, 135, results in principal-agent that experimental bargaining showing Florencia efficiency improves when multiple agents compete); Marotta-Wurgler, and the Quality of Standard Form Contracts: The Case of Software License Bargaining (2011) contracting Competition Power, and Efficiency: (describing 5 J. Empirical Stud. 447, 448-51 Legal (2008); George L. Priest, A Theory of Product Warranty, 90 Yale L.J. 1297, 1320—21 (1981) (finding no relationship between and warranty coverage); William C. Whitford, Law and the industry concentration Consumer Transaction: A Case Study of the Automobile Warranty, 1968 Wis. L. Rev. 1006, Agreements, the Consumer 1095-96. 6 Professor does find the expected between Marotta-Wurgler positive relationship price and market share or industry concentration. Marotta-Würger, supra note 5, at 451. 7 Id. at 475. 8 Omri Ben-Shahar & James J. White, Boilerplate and Economic Power in Auto Manufac 104 Mich. L. Rev. 953, 959, 971 (2006) for example, that orig turing Contracts, (observing, inal equipment manufacturers exert their power to extract broad warranties, discretion over little if any cor quantity, and rights to terminate without cause, while giving their suppliers responding right to cancel). Id. at 964. This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1670 Virginia Law Review contracts, ("OEM")] inefficient."10 or for the conjecture The puzzle that Ben-Shahar that dominate their industries [Vol. 98:1665 that some of these terms are and White raise is echoed repeatedly by in the business and legal press, who in practitioners and commentators voke bargaining power to explain both static and dynamic observations of contracting patterns. One static observation is that business entities tend to adopt different contractual alloca on whether they are buyers or sellers. risks depending When they are buyers or licensees, they demand extensive warranties and indemnification from their counterparts; when they are promises tions of similar or licensors, they disclaim, limit remedies, and demand indemni A dynamic observation is that contract from their customers. in loan and debt agreements, or material ad terms such as covenants sellers fication verse change ("MAC") clauses in mergers and acquisitions agreements, fluctuate over time between "buyer-friendly" and "seller-friendly" ver sions as market conditions change.11 when and why bargaining or market addressing and wor of nonprice terms remains unexplored The conventional law-and-economics theory offers The set of questions is a determinant power thy of investigation. a starting point for this analysis by offering two irrelevance propositions: (1) bargaining power affects the price but not the nonprice terms of a of bargaining contract (the strong-form version); and (2) the allocation may lead to different nonprice terms but does not change the like that the parties will agree to efficient terms (the weak-form ver sion).12 Each of these propositions depends on a set of implicit assump costs. We and zero transactions tions, including perfect information power lihood relax these assumptions in this Article to bridge theory and practice, by a theoretical framework to understand when and how bargain providing ing power determines contract design. To focus on bargaining power, we in their assume that each party can read and understand all provisions contract and set aside important effects that stem from lack of sophisti cation or bounded rationality.13 10 11 12 Id. at 982. See infra note 41; See Schwartz II.C. 13 & Section Scott, II.B; supra see also infra note 68; Section note 4, at 554; Section II.A; II.D. see also infra Sections II.B & set of issues with nonprice terms in contract, a common confounding dealing from the fact that weaker parties, particularly individual often do not read consumers, to or understand certain nonprice terms, either because sophistication they lack the necessary it is not rational for them to spend the time and effort to read and under do so or because When stems This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions The Effectof Bargaining Power 2012] 1671 treats a price term as purely The strong-form irrelevance proposition distributional: it allocates the surplus and has no effect on size. Howev er, variations in price terms can change the optimal nonprice terms when there is a change in price. We give two instances of this phenomenon. First, a different price alters one (or both) party's rate of substitution be tween price and nonprice terms. For example, a business executive who is richly paid may be willing to sacrifice a greater amount in salary for an official title than one who is less well paid. Second, price determines the severity of the adverse-selection a higher or lower or moral-hazard we would For example, to have problems. expect a loan agreement and more extensive covenants to address these tighter problems when the market rate of interest is high than when it is low, all else equal.14 Turning to the weak-form proposition, we outline four ways in which the shift in the bargaining power might lead to a deviation from efficient in negotiations contract design. First, the pursuit of advantage exacer bates the inclination of negotiators to engage in value-claiming rather than value-creating the with strategies. Second, party greater bargaining power has better incentives to invest effort and resources in innovating and developing contractual opportunities to create value. Third, business often occurs in a two-stage process: the price and other im contracting portant terms are decided by the business principals, and the design "de tails" are delegated to their respective lawyers. The exercise of bargain ing power in the second stage, after price has been determined, may distort the agreement on nonprice terms. Fourth, in negotiations charac terized by information asymmetry, unequal bargaining power might en stand the terms. Such bounded can be very significant rationality and lack of sophistication factors leading to inefficient and one-sided terms. See, e.g., Xavier Gabaix & David Laibson, Shrouded and Information in Competitive Mar Attributes, Consumer Myopia, Suppression kets, 121 Q.J. Econ. 505, 509 (2006); Avery Katz, Your Terms or Mine? The Duty to Read the Fine Print in Contracts, 21 RAND J. Econ. 518, 519 (1990); Russell Bounded Korobkin, Standard Form Contracts, and Unconscionability, 70 U. Chi. L. Rev. 1203, 1206 Rationality, Alan Schwartz & Louis L. Wilde, Information in Markets for Contract (2003); Imperfect Terms: (1983); The of Warranties Examples Yeon-Koo Che & Albert H. Mass-Market Communicating Law and Econ. Research Contract Paper and Security Interests, Choi, Shrink-Wraps: Terms? 3-4 (Univ. Series, Paper No. http://ssrn.com/abstract=1384682. 4 Albert Choi & George Triantis, Market Conditions Debt Covenants and Collateral, 87 N.Y.U. L. Rev. 69 Who of Va. Va. L. Should Sch. Rev. Bear of Law 1389 1387, the Cost of John M. Olin 2009), available at and Contract Design: Variations in (forthcoming 2013), available at 2009-15, http://ssrn.com/abstract=2048621. This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1672 Virginia Law Review excessive courage signaling or screening [Vol. 98:1665 in the activity design of terms. nonprice Each of these effects warrants more detailed treatment and we analyze only one of them in this Article. We examine the impact of unequal bar valuations gaining power when markets have buyers with heterogeneous of alternative contract terms and these values are not observable by the on from the economics of industrial seller(s).15 Building insights organi that one-sided zation, we demonstrate theoretically bargaining power can lead to the inefficient use of nonprice terms to screen or signal (de the power lies with the seller or the buyer, respec that or otherwise dominant seller may a monopolist tively). screen by offering suboptimal contract quality to the buyer who values At less. the other when the buyer has all the bargaining extreme, quality similar the of to case power, perfect competition, the buyer who values on whether pending We show the nonprice term less will seek to avoid by agreeing sellers.16 to suboptimal nonprice other buyer types subsidizing to signal her type to the terms, in which sellers can To illustrate, consider a market for automobiles in which buyers have to either limited or extended warranties and agree different valuations careful or careless careful or careless. of the warranty provisions because they are either drivers. The seller cannot observe which drivers are that both groups of buyers would pay more it would cost the seller to provide, and than warranty therefore efficient. A monopolist would seek to warranty is Suppose for the extended the extended maximize and limited its share of the surplus in each sale by offering both extended warranties at prices that would induce the careful driver to purchase a limited warranty. Similar inefficient separation would occur in a perfectly competitive market. If sellers offered the extended warran the ty to all buyers at a single price, careful drivers would subsidize careless choose 15 drivers. The careful the limited warranty drivers, then, would have an incentive to so as to signal their type to the market and that are plausible across a significant domain of nonprice make several assumptions even if the underlying nonassignable, including: (1) contract rights are generally affects not only the value of idiosyncratic (2) a buyer's preference product can be resold; attributes terms to the buyer, but also the cost to the seller; and (3) unlike physical nonprice to another and also to of a product, contract terms are easier to vary from one consumer We terms, modify, even after sale. 16 See Philippe Aghion & Benjamin Hermalin, Legal Enhance 6 J.L. Econ. & Org. 381, 381-82, Efficiency, problem of inefficient signaling). Restrictions 400-01 on Private (1990) Contracts (describing This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions Can a similar 1673 The Effectof Bargaining Power 2012] receive that exceeds a price reduction the value of the foregone warran ty.17 The problem of extreme allocations of bargaining power in either di in from the fact that one party has the power to dic rection stems, part, tate the terms of trade. The party with this power is willing, in many cir to sacrifice some of the aggregate cumstances, surplus in order to of the a more "even" sharing capture a larger share surplus. Therefore, in contract design. The of the surplus may address these inefficiencies balance bargaining power in this manner, After however, all, competitive markets essentially al locate all the power to the buyers without leaving any surplus for the sellers. We suggest three ways of representing more "even" sharing of bargaining power. In the first, some competition is introduced by allow that would circumstances are not obvious. seller (an entrant) to compete with an existing seller (an in In the second, the power of commitment is reduced by letting cumbent). the contracting parties renegotiate the original contract. In the third, the ing another seller's power to dictate the terms is curtailed by allowing the buyer to the terms with some delay. In each of these cases, we demon strate the key result that the conditions mitigate or eliminate the ineffi dictate ciencies of screening and signaling through nonprice terms.18 The Article is organized as follows. Part I investigates various mean for ings unequal bargaining power. Part II states the weak and strong forms of the bargaining and then out propositions, power irrelevance lines various paths by which bargaining power might determine contract III reviews related literature in industrial organization con Part design. cerning how monopoly and perfect competition can distort the quality of goods provided. Part IV provides a more detailed analysis of the effect of unequal 17 bargaining power on contract design when parties are asym This is familiar in markets for health insurance and can be applied more argument to contracting patterns in other industries. When both the healthy and sick people are the healthy will be subsidizing the sick, and they insurance, pooled together for a common will have an incentive to drop out of the common less than policy by, for instance, choosing broadly full insurance Equilibrium at a much lower in Competitive 90 Q.J. Econ. See, e.g., Joseph Stiglitz & Michael Rothschild, premium. Insurance Markets: An Essay on the Economics of Imperfect see also Albert H. Choi & Kathryn E. Spier, 629, 638 (1976); Information, Should Consumers be Permitted and Selection to Waive Products Product Safety, Private Con Liability? at available 25, 2011) (unpublished manuscript, http://ssrn.com/abstract=l 680932) (making similar argument with respect to product safety). The Appendix de presents a more general model in which a social planner (mechanism tracts, Adverse signer) can choose, and the seller. without 18 knowing (July buyer type, what types of contract to offer to the buyer This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1674 Virginia Law [Vol. 98:1665 Review metrically informed as to the value and cost of a given nonprice term. Part V shows that more even bargaining power can attenuate inefficient screening and signaling through nonprice terms. Despite the inefficien cies caused that legal power, we suggest briefly in Part VI to be able to mitigate these problems with a freedom of contract. We conclude by unequal bargaining institutions are unlikely the parties' by constraining comment as to possible directions I. What for future research. Power? Is Bargaining Although bargaining power is often cited as a critical determinant of contractual terms, neither the meaning of power nor the path of its influ ence is very clear.19 People differ in the meaning they attach to the ex The slipperiness of the term is due at least partly to the fact pression. that "bargaining power" frequently boils down to a tautology: one party had bargaining power when the resulting agreement is more favorable to use of the ex that party than its counterpart.20 In light of the ubiquitous and then to its we to its and hope meaning vagueness, clarify pression affect contract can of the in which some bargaining power explore ways design. Consider isolate 19 See a deal struck between the meaning David A. Lax of bargaining & James K. and Competitive Gain Cooperation is "notoriously Richard slippery"); a buyer and seller of a good. We can power by setting aside contract design Sebenius, 249 (1986) A. Posner, The Manager as Negotiator: for Bargaining power (noting that the concept of bargaining of Law 101-04 Economic (3d ed. Analysis the concept of unequal bargaining power is whether "the general 1986) (raising question Duncan fruitful, or even meaningful"); Kennedy, References Contract and Tort Law, with Special gaining Power, bargaining power See Thomas 41 Md. L. Rev. as "internally C. Schelling, 563, 623 (1982) Distributive and Paternalist Motives in Terms and Unequal Bar to Compulsory of (referring to the doctrine of inequality incoherent"). The Strategy of Conflict 22 (1960) ('"Bargaining power,' skill' suggest that the advantage goes to the powerful, the 'bargaining are defined to mean only that ne It does, of course, if those qualities to be are won by those who win. But, if the terms imply that it is an advantage gotiations more physical more intelligent or more skilled in debate, or to have more financial resources, strength, more military potency, or more ability to withstand losses, then the term does a dis strength,' 'bargaining strong, or the skillful. in bargaining are by no means universal situations; they advantages qualities of the Contract at Will, 51 U. Richard A. Epstein, In Defense a contrary value."); Chi. L. Rev. 947, 974 (1984) "[t]he ques (stating that, with respect to a bilateral monopoly tion of inequality of bargaining power can now be helpfully restated: which side will appro service. These often have between priate most of the surplus in any negotiations of bargaining be said to possess an inequality power than half the surplus"). them? when can therefore ... An employer more he is able to appropriate This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions The Effectof Bargaining Power 2012] 1675 that the nonprice terms have been fixed. That is, all rights the purchased by buyer are well defined and settled, and the only ques is The tion parties will agree to a price within a bargaining range price. that is defined by the parties' respective reservation prices. The negotia and assuming tion literature calls these boundaries Settlement Negotiated seller's BATNA the parties' Best Alternatives to a At the bottom of the range is the ("BATNA").21 or reservation price: the value of the seller's next best (for example, the seller might choose to sell it to another use of the good buyer or use it herself). At the top end of the range is the buyer's BATNA or reservation price: the value of the buyer's next best use of her funds (for example, the price at which she can buy the good from another seller or the foregone benefit if she walks away from the pur models presume that the point within this chase). Many game-theoretic range at which the price is agreed upon is determined by the relative pa tience and risk aversion of the parties, as they look at the prospect of continued bargaining and delayed agreement.22 In contrast, the negotia tion literature and practitioners think of bargaining power as adjusting of the bargaining range itself, in addition to placement the boundaries within the range. From this perspective, price is a function of the seller's and buyer's of the two reservation prices (each party's own respective perceptions and that of her counterpart).23 The perceived bounds for the bargaining range, and the price ultimately chosen within this range, are determined or endogenous to the nego by a mix of factors that might be exogenous tiations. We divide these factors into five categories: (1) demand and supply conditions, (2) marketconcentration,(3) private information,(4) patience and risk aversion, and (5) negotiating skills and strategy. The first category of exogenous factors consists of the demand and supply conditions in the relevant market. When there is a significant in crease in the demand ket-clearing 21 See, for the product or reduction in the supply, the mar will tend to increase and sellers are often said to have price e.g., Roger Fisher 97-105 (2d ed. 1991). 22 et al., Getting to Yes: Negotiating Agreement Without Giving In In complete, Ru models, such as those by Professors symmetric information bargaining and Stahl, the party who is more patient—that is, who has a lower discount rate— binstein Ariel Ru gets a larger share of the surplus. See Ingolf Stahl, Bargaining Theory 121 (1972); in a Bargaining 50 Econometrica In the binstein, Perfect Equilibrium Model, 97, 108 (1982). is partly reflected through the dis two-period bargaining games that we present, the patience count factor S. 23 Fisher et al., supra note 21, at 102 ("The better your BATNA, the greater your power."). This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1676 Virginia Law [Vol. 98:1665 Review bargaining power. An example we discuss further below is the tightening of credit during and following the 2007 financial crisis: indus substan try participants noted that when the supply of credit decreased increased tially, lenders enjoyed greater "bargaining power" over their borrow ers.24 A second A factors is market concentration. category of exogenous market is often referred to as its monopolist's power bargaining power. A buyer's no-agreement alternative is limited by the fact that there are no other sellers in the market and his reservation price is corresponding the same good from a competing ly higher than if he could purchase seller. Typically, market concentration on the seller side increases price and concentration on the buyer side decreases it. third category of exogenous ad factors contains informational the other par vantages that one party may enjoy by knowing more about about A with infor or information itself. party private ty by concealing A can be thought of as having a type of monopoly stemming from this information.25 In the analysis later in this Article, we isolate the pri vate information relating to one's own reservation price and treat it dis mation tinctly from bargaining power. Thus, in Parts IV and V, the buyer has private information as to how much he values the good being sold, while the other aspects of bargaining power shift between the buyer and the seller. such as pa We identify a fourth category containing characteristics tience and risk aversion that may determine where the agreed price will fall within a given bargaining range.26 Bold parties, for example, may do better than timid players, and the patient negotiator typically enjoys higher returns than her impatient opponent. Patience may be, in turn, a function of other factors, such as the solvency and liquidity constraints of the party, or its ability to diversify the risk of an unfavorable bargain ing outcome. tactics that can In a fifth category, we put the various negotiating reservation of either the actual or party, so as to price change perceived induce a favorable shift in the bargaining range.27 For example, a party 24 text. and accompanying of a valuable re et al., supra note 5, at 135 ("[T]hey are the sole 'owners' source—information about their type."). 26 text. See supra note 22 and accompanying 27 of bargaining ex determinants the exogenous Rather than analyze power, negotiation its counterpart's perts focus on the means by which a party can increase its own and decrease 25 See infra notes 41^3 See Cabrales This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1677 The Effectof Bargaining Power 2012] if she can improve her alternative to reaching an to third parties that increase the cost of will be more successful or make commitments agreement in the negotiations. Or, a party might take steps to granting concessions her opponent's outside worsen (or appear to worsen) opportunities, or We credible threats otherwise.28 might put in this category the through hold-up issue in contract theory: the tactic of inducing the investments that can later in other party to make relationship-specific loss on that from flict a significant non-agreement.29 Strategic ne party gotiators also exploit the cognitive biases and errors of opponents, par well-known ticularly the tendency of individuals and be overconfident in their abilities. to anchor, escalate commitment, In some cases, bargaining through results. These skills are the subject of one or more agents might improve and we do not attempt to summarize many books on negotiation here. them In any given transaction, one or more of these factors may be in play. Which ones are present may determine the path by which unequal bar gaining power affects contract design. This Article is meant to set a framework ing power for further investigation terms. of the interaction II. How Bargaining Power Affects Contract in bargaining In this Part, we explore how changes influence terms. We start by articulating may nonprice weak-form versions of the bargaining power irrelevance irrelevance between bargain and nonprice is one that flows proposition in the process. advantage of itself has often proved logically Design power balance the strong- and An proposition. from a set of restric Lax & Sebenius, supra note 19, at 257 ("Analyzing to be a sterile exercise. However, directly focusing of the bargaining set and the ways that such changes in and 'power' on factors that can change perceptions influence out comes seems more fruitful for both theory and practice."). 28 The Competi K. Dixit & Barry J. Nalebuff, See, e.g., Avinash Thinking Strategically: tive Edge in Business, Life 291 (1991) matters is his outside Politics, and Everyday ("[W]hat to that of his rival. He will do better in the bargaining even if he makes a both parties' so long as that of the outside opportunities, rival is damaged more severely."); G. Richard Shell, Bargaining for Advantage: Negotiation for Reasonable 103 (2d ed. 2006) Strategies People (noting that threatening your opponent opportunity commitment relative or a threat that lowers with losses "as astute negotiators have resulting from the failure to agree works because, for centuries and psychologists have repeatedly losses loom larger proven, potential in the human mind than do equivalent gains"). 29 Structure 26-27 Oliver E. See, e.g., Oliver Hart, Firms, Contracts, and Financial (1995); known Williamson, The Economic ing 52-54, 61-62 (1985). Institutions of Capitalism: Firms, Markets, Relational This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions Contract Virginia Law Review 1678 [Vol. 98:1665 tive assumptions that are suspected to be both unrealistic and binding. in law-and-economics Two famous irrelevance are Modi propositions and the Miller's of gliani ("MM") significance proposition concerning debt and equity in corporate finance30 and Coase's of legal entitlements.31 The assumptions proposition on which the bargaining power proposition is based are very strong. In the choice between about the allocation par they are similar to those of the MM and Coase propositions, the of information and no transaction ticularly assumptions symmetric costs. These are the assumptions that we begin to unpack in this Article. deed, then divide We our hypotheses of influence of bargaining power on terms into two First, contrary to the strong-form nonprice categories. a in version, change bargaining power, through its effect on price, may alter the optimal nonprice terms. Second, contrary to the weak-form ver sion, bargaining power may lead to inefficient terms as a result of the of the bargaining exercise power. We also briefly introduce the more complicated question of how bargaining power influences contract de when sign negotiations are conducted in stages and through agents. A. The Bargaining Some nopolist theorists have will competitive Power Propositions that the contract terms offered by a mo as those offered by a seller in a asserted be essentially market.32 Each Irrelevance the same provision of a contract creates value for at 30 Finance and the Franco Modigliani & Merton H. Miller, The Cost of Capital, Corporate The authors later presented 48 Am. Econ. Rev. 261, 262-64 (1958). Theory of Investment, an irrelevance the effect of dividend policy on firm value. Merton H. proposition concerning of Shares, 34 J. Dividend Miller & Franco Modigliani, Growth, and the Valuation Policy, Bus. 411,412-14(1961). 31 R. H. Coase, The In some Econ. 1, 15-16 (1960). of the can be thought of as an extension setting, parties bargain over legal rights so as to elim proposition. if any, that is engendered Id. In the inate the inefficiency, by the initial legal entitlement. there are no default nonprice terms; rather, the bargaining power irrelevance proposition, sense, Coase the bargaining Problem of Social power irrelevance In a standard Coase Cost, 3 J.L. & proposition the surplus from the transaction. parties choose a set of nonprice terms that maximize 32 In an early article, Alan Schwartz based his analysis on three very weak assumptions demand for quality does not vary with that now seem especially (1) "consumer incomplete: the amount of physical industry use (2) "all firms within a competitive product consumed," of the level of industrywide technology regardless output," and (3) "the production is 'similar' to that of a competitive of a monopolist industry in the sense that the factor combinations at and competitive monopolist industry face the same cost-minimizing the same function supra note 4, at 1073. In a footnote, any level of output." Schwartz, to set aside that the monopolist does not price-discriminate assumes n.44, he also the possibility raised by id. at 1075 This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions The Effectof Bargaining Power 2012] 1679 least one party and that party may view the provision as part of the good being sold. For example, a warranty, a termination right, or a selected dispute resolution venue is "sold" by one party to the other. The or service states that, if the cost and demand curves are the proposition same for both monopolist and seller in perfect competition, each will of fer the same nonprice terms. The proposition would extend to bilateral between and seller in which the balance of power buyer bargaining irrelevance shifts between the two parties. The proposition is based on the observation that a monopolist that re fuses to sell the quality desired by its customers, when their willingness to pay exceeds his cost of providing that quality, is simply leaving mon theorists refute the ey on the table. On this basis, law-and-economics concern leads of more to unfair Scott provide traditional contract contract terms. scholars Professors Alan that bargaining Schwartz and power Robert a recent statement of this refutation: It is widely believed that parties exercise bargaining ing weaker contracting partners to take unfavorable that superficially appear one-sided are commonly power by requir terms.... Terms described as the product of "unequal bargaining power." But when bargaining power is determined prior to contract formation, as is common in business con texts, these views are incorrect. Bargaining power instead is exercised in the division of the surplus, which is determined by the price term. Parties jointly choose the contract terms so as to maximize the surplus, which the parties may then divide unequally.33 Our interest in this Article terms nonprice of is the effect of bargaining power on the We assume a successfully completed bargain. to understand the throughout that both parties have the sophistication terms of the contract. This assumption ensures that the contracts we are discussing clearly improve the welfare of both parties, compared to each literature suggests that par position. Experimental to fail to reach a welfare-improving if agreement they have significantly unequal, as opposed to roughly equal, bargaining party's no-agreement ties are more likely Professors Michael Mussa and Sherwin Theory 301, 301 (1978). 33 Schwartz monopolist & looks Rosen, Monopoly and Product Quality, 18 J. Econ. supra note 4, at 554; see also Baird, supra note 4, at 941 for efficient warranty terms. Using inefficient terms compromises to extract rents. She is much better off providing quality goods Scott, nopolist's ability cient terms and charging as much as she can from them."). This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions a ("Even the mo and effi 1680 Virginia Law Review power.34 The relevant power affects contract question [Vol. 98:1665 in this Article is whether bargaining of the bar design, gaining surplus. We return to the simple sale example introduced above and ask whether a shift in the relative bargaining power between the seller and buyer might alter the agreed-upon warranty. A warranty allo in addition to the distribution cates the risk of product malfunction, depending on features such as its In the sets incentives for the sell and duration. so, warranty scope doing er to raise the quality of the good and the buyer to take care in using it. It might also be a means by which the seller can signal the quality of the good. One could think of a surplus-maximizing warranty that optimized of the buyer and given the characteristics seller. Bargaining power might affect the agreed-upon warranty in two directions. It might change the terms of the optimal warranty or it might across these considerations, lead the parties to deviate from the optimum in their agreement. In Section II.B that follows, we suggest that a shift in bargaining power might change the optimal nonprice terms through its effect on price. The change in price may have a wealth (or substitution) effect on a party's tradeoff between price and nonprice terms. It may also alter the of one or or moral hazard problems severity of the adverse selection of risk. Then, in both of the parties, leading to a new optimal allocation Section II.C, we identify several ways in which bargaining power can affect the efficiency of the nonprice terms. B. Bargaining Power Can Alter the Optimal Terms Nonprice In this Section, we explore how bargaining power can alter optimal nonprice terms.35 In the first instance, a shift in bargaining power puts 34 See the dignity et al., supra note 21, at 105 (noting the importance of preserving Lax & Sebenius, supra note 19, at 129 ("A number of studies suggest to his own weakness and the counterpart's a bargainer attributes his concession Fisher of the weaker that when party); strength, a blowup is likely."); Jeffrey Z. Rubin & Bert R. Brown, The Social Psychology of Bargaining and Negotiation 217 (1975) ("Pairs with equal [bargaining]power attainedhigh er joint payoffs than those with unequal Environmental Complex Negotiations: ory 247, 252, "In 269 addition (1997) (noting to removing in Nina Burkardt et al., Power Distribution power."); Does Balance Matter?, 7 J. Pub. Admin. Res. & The that power imbalance the negative factors tends to inhibit detailed above, successful negotia symmetrical power to creative, deal between the parties, to encourage feelings open parties good and remove the temptation to use force and threats." Robert S. Adler enhancing suggestions, in with Power Differentials & Elliot M. Silverstein, When David Meets Goliath: Dealing tions). tends Negotiations, In many the optimal 5 Harv. cases, (citations omitted). Negot. L. Rev. 1,18 (2000) the conditions that lead to a shift in bargaining power terms by changing, for instance, the volatility in the economic This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions might also change environment of The Effectof Bargaining Power 2012] pressure 1681 on the price term. This effect, in turn, can change the optimal of nonprice terms in the contract. The important insight here is design that any "price" term does not simply allocate surplus. A significant in bears on the and for change price surplus may provide opportunities surplus creation by the adoption of different nonprice terms. We explain how this might happen in two types of circumstances. First, it is helpful to recall that a contract term is part of the "quality" of the good or ser A party's willingness vice being exchanged. to purchase quality is a function of her wealth, just as is her willingness to buy the underlying or service. as increases because of a shift in bar Therefore, good price gaining power toward a seller, the buyer's wealth declines and so does her demand for various nonprice terms, such as an extended warranty or the right to sue in the buyer's own state courts.36 To a casual observer, it may seem like the seller is exercising its bargaining power by reducing the quality of the nonprice terms, whereas the seller is in fact responding to the wealth effect of a higher price on the buyer's demand for such terms. This effect may be at work in an employment contract or a venture in investment a capital ("VC") start-up enterprise. VC contracts contain financial terms (dividing the equity payoffs between the VC fund and the entrepreneur) and governance terms (for example, the VC fund's seats on the board).37 The financial terms can be thought of as "price" and the governance terms as "nonprice" terms. The governance terms are valuable tend to place in addressing problems of moral hazard, but entrepreneurs offsetting value on maintaining control of the fate of their the parties. For example, an increase in the volatility of economic conditions can lead to both a shrinking in the availability of credit and the value of covenants that constrain the borrow er's incentive to take risks. See infra note 41 and accompanying text. We set this possibility aside 36 in this Article There may because be other it does not follow substitution effects from the shift in bargaining from price changes. arising power. Professor Dennis Carlton customers tomorrow for today, so that con suggests may trade off consumption At least to the extent that the impetus to change prices shows a sumption shifts to tomorrow. some empirical studies have documented conditions, change in supply and demand price ri can be market-clearing devices in lieu gidity in some industries. Delivery lags, for example, of or in addition to price. Dennis W. Carlton, Equilibrium Fluctuations When Price and De in Lag Clear the Market, 14 Bell J. Econ. 562, 562 (1983) (finding that fluctuations as important to the equilibration of demand and supply as delivery lags are approximately are fluctuations in price); Dennis W. Carlton, The Rigidity of Prices, 76 Am. Econ. Rev. livery 637,638(1986). 37 See World: 82 (2003) Steven N. Kaplan An Empirical (describing & Analysis financial Per Stromberg, Financial Contracting Theory of Venture Capital 70 Rev. Econ. Contracts, and governance characteristics of VC Meets Stud. financings). This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions the Real 281, 281— Virginia Law Review 1682 [Vol. 98:1665 If the VC has more bargaining power than the entrepre own company. to sell a larger portion of the value the neur, entrepreneur is compelled the as well as to surrender control of, for instance, a of company, agree given number of seats on the board of directors. As power shifts to the in available of expansion capital, she can offer a entrepreneur because smaller share of equity to the VC fund for every dollar of capital invest wealth increases and the Her expected ed (a lower price for capital). marginal tradeoff between money and control changes as a result: in stead of reducing the share of equity she gives the VC, she would offer instead fewer outside seats on the board. This phenomenon may occur even in contracts between firms because benefits at their individual agents trade off monetary and nonmonetary another wealth. Consider on their individual different rates, depending contract in the stream of VC of the VC the limited partnership agreement fees and provides for the management funding: fund. This agreement interest that are paid to the venture capitalist. The venture capi the fund, including talist also enjoys private benefits from managing is costly to the in of these benefits and His pursuit perquisites. prestige carried vestors (limited partners) to the fund. To includes address these restrictions incentives, the on the decisions typically partnership agreement and activities of the venture capitalist. The tightness of these restrictions reflects, no doubt, the cost of the private benefits to the investors, but al so the tradeoff in the eyes of the venture capitalist between monetary and the value of such private benefits. In their study of compensation venture capital partnerships in the 1990s, Professors Paul Gompers and Josh Lerner observed that cyclical changes in the demand for and supply of venture capitalists may explain shifting contracting patterns.38 While of experienced the supply of capital varies, rigidity in the availability in market. When this imbalances causes venture capitalists periodic monetary returns are capital inflows are greater, the venture capitalists' higher. Gompers leads to dilution also of restrictions they do not explain 38 noted, however, that the increase also but on activities of venture capitalists; the mechanism by which this effect takes place.39 and Lerner & Josh Lerner, The Venture Capital Cycle 31-32,45^17 (1999). Gompers in the World of Ven Contract Design Id. at 25; see also George G. Triantis, Financial the irrelevance 68 U. Chi. L. Rev. 305, 319-21 ture Capital, why question: (2001) (asking not use their bargaining would venture capitalists power to capture a larger share of the 39 Paul monetary surplus from efficient contracting?). This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions The Effect of Bargaining 2012] 1683 Power explanation may lie in the shifting rates of substitution wealth effects we describe here. The caused by The effect of bargaining power through price on nonprice terms may also run through a second causal chain. Changes in price have an impact on the nature of adverse selection and moral hazard problems in some as lending or in plagued by these problems—such demand or excess sup not excess clear, leaving markets—may with insurers faced excess demand, may be reluctant ply. For example, to raise premiums for fear of driving out the lower-risk customers and transactions. Markets surance being left with a riskier pool. Nonprice contractual terms are commonly designed to mitigate these information problems by, for instance, screen A change in price can increase or de ing out the high-risk customers. the severity of these problems, and some nonprice terms become more or less valuable. Thus, changes in price can alter correspondingly the optimal use of these terms. In the warranty example, when a seller crease bargaining power and can charge more for a warranty, it is more likely to lose low-risk customers and attract a riskier pool of customers. Therefore, the rise in price may itself lead to a narrowing of the scope of the optimal warranty. gains A loan lems contract of adverse is a more powerful example because selection and moral hazard are widely the dual known. prob These and events of default. The typically include a set of covenants of the covenants may give the lender the right to accelerate the of the loan and if the borrower fails to pay the accelerated contracts violation maturity as amount, the lender may then enforce its claim against the borrower's restrict actions or such as the bor sets. Covenants some decisions, may rower's incurring new liabilities, selling assets, or making distributions Or, covenants may set tripwires that trigger default, in financial-ratio tests such as maximum debt-to-equity or interest to stockholders. cluding to-earnings ratios. Contracts vary in terms of the types of behavior that are restricted or the types of tripwires, as well as how close the tripwires are set to the borrower's current condition. Both the breadth and tight ness of covenants are matters of contract design.40 are sometimes regarded as "boilerplate" provi across contracts between packages vary considerably covenants Although covenant sions, 40 Since Contracting: scholarship instruments. work on debt covenants, Clifford Smith & Jerold Warner, On Financial An Analysis of Bond Covenants, 7 J. Fin. Econ. 117 (1979), a large body of the determinants in loan agreements and bond has investigated of debt covenants the classic See, e.g., Choi & Triantis, supra note 14 (manuscript at 2). This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1684 Virginia different lenders Law [Vol. 98:1665 Review A growing body of theoretical and em identifies firm-specific and market determi and borrowers. pirical finance scholarship The industry press also nants of the intensity and tightness of covenants. a connection between the and demand conditions suggests strong supply and covenant deals grew at a staggering pace patterns. Covenant-lite the first half of the decade until the onset of the financial throughout past crisis in 2007, and market observers attributed this growth to the excess loans collapsed in the supply of credit.41 The market for covenant-lite second half of that year and was followed by a period of more extensive that cove and tighter covenants during 2007-2009. Reports suggested of the excess supply of in nant-lite deals then emerged again because vestment funds, at least for higher-grade borrowers.42 The following re cent explanation by a partner at the law Wharton & Garrison is typical: firm of Paul, Weiss, Rifkind, (cov-lite) loans became widespread at the top of the last credit cycle before the 2007 credit crunch. During the credit crunch, Covenant-lite 41 Ana tures Lai & Steven Diminish M. Bavaria, The Leveraging 2 (2007), of America: available Covenant-Lite Loan Struc at http://s3.amazonaws.com/ In this report by zanran_storage/www2.standardandpoors.com/ContentPages/561145523.pdf. that: Standard & Poor's on the eve of the financial crisis, the ratings agency observed LBOs pace of private equity sponsored Strong loan market liquidity and the continued Recovery Prospects loans in 2007. Such favorable are driving a record volume of leveraged with growing investor demand from structured finance tors, combined "covenant-lite" bank facilities with weakened hedge funds, have allowed market vehicles loan fac and struc tures to emerge as the instruments of choice for many issuers. As the volume of lever facilities has in of covenant-lite an all-time high, the proportion aged loans reaches loans will revert to It remains to be seen whether leveraged creased tremendously.... There has already been structures when the credit cycle turns.... so far this year as market conditions begin to soften, with certain pushback without a robust covenant package. transactions unable to get through syndication more traditional some Id. 42 Wall St. J., Feb. 7, 2011, at Aneiro, Aleris Debt Sale: 'Covenant-lite,' See, e.g., Michael ... and has allowed has pushed the average junk-bond ("[D]emand yield down to 7.01% that govern new offerings."); or covenants, issuers to water down investor protections, Michelle Sierra Laffitte, IFR-Covenant-Lite Buyout Loans Return to U.S. Loan Market, Int'l C3 Fin. Jan. Rev., 31, 2011, http://www.reuters.com/article/2011/01/31/loans-covenant-lite are expected to try to ("As the market gets hotter, companies Kate in deals that were completed and slash covenants reduce recently...."); spreads of the Past, Covenant-Lite Loans Are Back but Investors Hope to Limit Mistakes Laughlin, idINLDE70U0T520110131 Fin. Times, where investors the covenant-lite covenant ldf-8d91 24, 2010, http://www.ft.eom/intl/cms/s/2/a242e5d0-f812-l loan market is for the most part a seller's environment, ("[T]oday's investors buying are flush with cash they need to put to work .... [S]ome deals are not solely loan investors, so in their hunt for high-yielding paper, Nov. 00144feab49a.html concerns are a low priority . . . ."). This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions The Effect of Bargaining 2012] 1685 Power from the market be however, new cov-lite loans largely disappeared cause lenders had greater market power to reject these types of bor rower-friendly deals.... [Smarting in 2010, cov-lite loans began reap pearing in the syndicated loan market. Borrowers can obtain cov-lite loans because of market dynamics. At the top of the last credit cycle, there was an oversupply of capital, and lenders competed for deals there was a from private equity sponsors and borrowers. Because greater supply of capital than there was demand to borrow capital, borrowers had more leverage to negotiate looser and more favorable terms, including cov-lite structures.43 on supply and demand, and the place great emphasis This of market is puzzling for the conven consequent power. tional law-and-economics position, which posits that bargaining power affects only price rather than nonprice terms such as covenants. If a cov These accounts balance and moral hazard selection by mitigating adverse problems, borrowers with bargaining power should be at least as eager to agree to them as when they lacked bargaining power. In fact, the very enant creates value ability to extract most of the surplus from a deal would give powerful borrowers a greater share of the surplus created by these terms. A more between elaborate story is needed, therefore, to explain the connection shifts in bargaining power and changes in contract design. costs internal to lending institutions might provide part of the Investment managers face pressures to meet targets for re explanation. turns. Where there is competition for relatively few debt securities, they Agency may be willing to sacrifice covenant protection for a higher yield. The risk may or may returns are immediately apparent while the consequent not reveal itself later. The financial crisis, however, drew dramatic atten tion to these risks, so the re-emergence of covenant-lite loans is not as In an alternative explanation, we suggest that the im explained. pact of bargaining power is mediated through an effect on price. that lenders acquire more bargaining power because exoge Suppose nous forces tighten the supply of credit. The first-order effect is to place upward pressure on interest rates. As noted above, the lower-risk bor easily rowers may exit, leaving a riskier pool. These borrowers also face incen tives to take greater risks in order to make borrowing at a higher rate and moral worthwhile. The prospect of exacerbated adverse selection 43 Eric Goodison, Covenant-Lite 37 (emphasis added). Loans: Traits and Trends, Prac. Law J., Sept. This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 2011, at 36, Virginia Law Review 1686 hazard would cause face of excess [Vol. 98:1665 to refrain from raising interest rates in the supply. The second order effect, of strict covenants and collateral, to discour lenders demand and to ration however, is that the value age and deter high-risk borrowers, would be greater. The optimal con tract design would have both tighter covenants and, probably, broader collateral as a result of the change in market conditions.44 This explana with empirical work that has found a positive correla rates of interest and covenant and tightness tion is consistent tion between market breadth.45 In sum, in relationships affected by asymmetric information (such as or lending insurance), price is an imperfect tool for adjusting for supply or demand changes, or shifts in bargaining power. As we have noted, the information problems of adverse se changes in price can exacerbate lection and moral hazard. Therefore, the parties can improve the effi ciency of their contract by using nonprice terms instead to shift value to the "stronger" party.46 As a result, the balance of from the "weaker" bargaining power might in fact affect contract design. C. Bargaining Power Can Lead to Inefficient Nonprice Terms In designing their contract, the parties might not be aware of the nonprice terms that maximize their surplus. Two factors are important in this regard. First, one or both of the parties must invest in the task of de information, considering alternatives, tailoring them signing: processing 44 We explore this effect in a companion paper that presents a model of this phenomenon and, in this light, examines existing data concerning cyclical changes in patterns of covenant and collateral. See Choi & Triantis, supra note 14. 45 of Leverage, Debt Ma Choice See, e.g., Matthew T. Billett et al., Growth Opportunities, Control 62 J. Fin. 697, 706-08 (2007); Greg Nini et al., Creditor turity, and Covenants, rela and Firm Investment (finding positive Policy, 92 J. Fin. Econ. 400, 400 (2009) R. Rob Michael interest rates and covenant breadth); Bradley & Michael tionship between Debt Covenants 13, 2004) 2-3, 28 (May (un erts, The Structure and Pricing of Corporate available at http://ssrn.com/abstract=466240); Zhipeng Zhang, published manuscript, Rights Recovery 2009) 46 Rates and Macroeconomic Conditions: The Role of Loan Covenants available at http://mpra.ub.uni-muenchen.de/17521/). (unpublished manuscript, A similar argument may be made in the context of corporate acquisition (Sept. 2, agreements. return to the seller, an exogenously induced reduction in price reduces the expected the value of cove the risk of moral hazard by the seller is greater and so is, correspondingly, clauses. We discuss material adverse nants and conditions for closing, including change When these clauses further below. See infra text accompanying notes 60-70. This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions The Effectof Bargaining Power 2012] 1687 to the parties' circumstances and innovating new solutions.47 While de the is to who invests in it, the incremental value of the party sign costly of this exter is shared by both parties.48 The consequence in the can be another is unless deal, the that, nality design redeployed and their be less efficient.49 will underinvest parties agreement may it en if a party has bargaining power because, for example, However, investment or the capacity for patience, it might have a better in joys a monopoly centive to expend the resources to develop value-increasing necessary it can capture most of the value.50 The adhesion nonprice terms because contracts of monopolists were the béte noire of the early academic such as Professor Friedrich Kessler, who thought that the commentators, nonprice terms in these contracts would be significantly less favorable to the law-and-economics scholars who Conversely, that they would be efficient and no different than in competitive markets.52 Contrary to both sets of schol the counterparties.51 followed suggested those produced ars, this Article or superior bargain suggests that those with monopoly are to the from innovative able ing power capture payoffs design. There their be in fact both different and fore, the terms of agreements may more efficient. Second, vantages ten need agreements in endowments create value ad by exploiting comparative and differences in preferences, and parties of to make such information during negotiations a of Yet, bargaining power in nego party's pursuit agreements possible. tiations can be antithetical to the creation of value.53 In the language of 47 See tion). 48 George 22, (Mar. to exchange 2011) Triantis, Standardization, (unpublished manuscript, and Innovation in Contract Modularity, on file with the Virginia Law Review Design Associa 2 There are countervailing For strategic reasons for writing the first draft of an agreement. it can lead the opponent to anchor on the proposed division of surplus and, more of the bargaining generally, on a perception range. Id. at 15. 49 The general point is well documented in contract theory. The hold-up exists problem example, when the parties cannot contract ex ante to share the value of the specific investment, for ex it cannot be verified in court. Where the investing party also has the bargain because ing power and can capture most or all of the surplus, it has the incentive to make the value investment. Oliver Hart & John Moore, See, e.g., Hart, supra note 29, at 32-33; increasing ample, Property See 51 See 52 See 53 The of the Firm, 98 J. Pol. note 47, at 12. Kessler, supra note 1, at 632-33. supra note 4. Rights Triantis, and the Nature Econ. 1119,1132 (1990). supra literature speaks of a fundamental tension between and creat negotiation claiming 245^6 the "Nego ing value. See, e.g., Lax & Sebenius, supra note 19, at 38^10, (discussing tiator's Dilemma," where it is individually rational for each party to claim value, but that this This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions Virginia Law Review 1688 [Vol. 98:1665 negotiation experts, shared information enables value creation while pri vate information promotes value claiming. To increase its share of the surplus, each party strives to conceal its own information and extract the private information of its counterpart. For example, a buyer might agree to a limited warranty in order to hide the fact that the good or service be is of great value to her. Faced with incomplete infor ing purchased mation, the seller may screen for the relevant information by offering a choice between a contract with a complete disclaimer and one with a full warranty. Low-valuing buyers, in contrast, may be eager to communi cate their relatively low valuations but have difficulty doing so credibly. dis by agreeing to a complete They may signal their low valuations claimer. Where the parties use nonprice terms to screen or to signal, as the case may be, these terms are likely to be inefficient. The danger of inefficient terms resulting from screening or signaling is well known.54 What is significant for our analysis is the more compli of bargaining power changes strat cated question whether the allocation egies and thereby the degree of inefficiency.55 In Part IV, we demon are most severe when strate with a numerical example that inefficiencies In there is significantly unequal bargaining power in either direction. the parties' ability to create value); Malhotra & Max H. Bazerman, Nego Deepak Brilliant Results at the Bargaining How to Overcome Obstacles and Achieve Robert H. Mnookin et al., Beyond and Beyond 6-7, 82 (2008); Winning: Negotiating constrains tiation Table Genius: 9 (2000). For example, the making of binding com Value in Deals and Disputes to third parties or threats to the opposing value, party may be helpful in claiming to create value. Lax & Sebenius, but constrains the flexibility and good will necessary supra note 19, at 245—46. 54 See supra note 53 and infra Part III and note 84. 55 for Professor Jason Johnston examines the effect of bargaining power on contracting to Create mitments the decision for breach: specifically, to accept or opt out of the rule in Had caps on damages Johnston demonstrates that where the seller is a 156 Eng. Rep. 145 (1854). ley v. Baxendale, this would its high-value buyer would hesitate to ask for higher caps because monopolist, its market power. For its inform the seller that there is more value to extract by exploiting in the the seller has the incentive to screen by price/quality discriminating, part, however, see in Subsection IV.B.l below. manner described Johnston, supra note 4, at 616, 636-37; Professors Ian Ayres and Robert Gertner also Ayres & Gertner, supra note 4, at 735-36. damages provision (given buyer type) would be agreed suggest that the efficient liquidated to if the market were competitive instead. Ayres & Gertner, supra note 4, at 742. In their on buyer type. Hence, once the of breach does not depend model, the seller's probability the seller's liability is damages, expected buyer and the seller fix the size of the liquidated of the buyer's loss from breach. This is not for all types of buyer, that is, regardless in Part IV. For instance, even when the size of for the nonprice terms we examine on the the warranty is fixed, the cost of serving a buyer under the warranty term depends on the characteristics of of claiming that warranty depends buyer type since the probability the same the case the buyer. This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions The Effectof Bargaining Power 2012] 1689 terms of market concentration, they are most severe either when there is a monopoly (excessive among sellers screening) or perfect competition for buyers (excessive signaling or cream skimming). The party motivat the inefficient ing nonprice term in this way is the one with the bargain ing power. The intuition is that the dominant party is willing to accept the consequent incremental loss of surplus in order to improve its share. lit The results are consistent with those from the industrial organization erature. In Part V, we show that when "evenly," each party has less of an ing or signaling, and the agreement fine more even bargaining power in some competition the parties share the surplus more incentive to engage in either screen they reach is more efficient. We de three ways. another seller In the first, we introduce (an entrant) to possibly by allowing In the second, the an compete against existing seller (an incumbent). the contracting parties to power of commitment is reduced by allowing the original contract with some chance. In the third, the sell to dictate the terms is curtailed by allowing the buyer to a counteroffer with some make delay. renegotiate er's power In each of these cases, we demonstrate the key result that the condi tions of more even bargaining power can mitigate or even eliminate the inefficiencies of screening and signaling through nonprice terms. While the three variations require different game theoretic presentations, they share a common theme. When one party deliberately imposes an ineffi cient nonprice term and leaves an unrealized surplus on the table, it pro vides a strong incentive to others to capture the unrealized surplus and In the the first that incentive is given variation, eliminating inefficiency. to a competitor (entrant); in the second, to the seller through renegotia tion; and in the third, to the buyer. In the process, we argue that it is im instance, by not introducing too portant to strike a proper balance—for much competition—so erage. The Appendix as not to give the other too much bargaining lev presents a more general model in which a social planner (mechanism designer) can choose what types of contract to offer to the buyer and the seller. The model shows that when the social plan ner only cares about the seller's or the buyer's welfare, the social plan ner will also impose inefficiency on the contract, but when the social planner cares about them more evenly, such inefficiency gets mitigated or eliminated. Finally, there is one more context in which bargaining power may have an effect on contract design: through two-staged negotiations This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1690 Virginia Law Review [Vol. 98:1665 agents. First, a term sheet, letter of intent, or similar document settles the price and other key terms. Second, a later negotiation (typically through lawyers) settles the remaining nonprice terms for the definitive contract. D. Bargaining Power in Two-Staged (Price-First) Negotiations term is unlikely to yield value to both parties. Therefore, logrolling is an essential element for creating value in bar in gaining. A buyer agrees to a lower quality of product, for example, Any given contract deal price. In commercial to create value in this exists when making, opportunity way parties can trade nonprice terms for adjustments in price. For this reason, return for an earlier delivery date or a lower the broadest it would optimal to leave the price term open until all other terms fail to reach the sur settled. Negotiations may nevertheless in the previous deal because of the obstacles discussed plus-maximizing have seem been Section. But fixing price at an earlier stage would nities for value creation. further limit opportu Although price terms are usually determined after the nonprice terms in commercial have been set, this is not always the case. For example, and loans, private equity investments, many corporate acquisitions, terms are agreed upon after the price is settled.56 In the first stage of ne often gotiations, the parties negotiate price and key nonprice provisions, with the signing without their lawyers.57 This stage typically concludes of of a document such as a term sheet, letter of intent, or memorandum which is not legally binding.58 The parties then turn over understanding, the second stage of negotiations to their lawyers to work out the details in a definitive contract, including representations and warranties, closing and termination rights. These terms are usually conditions, covenants, settled without adjustment to price. The parties would probably have an of these terms when they struck a price in the first stage expectation terms fall what is "market" at the time). If the second-stage (perhaps outside reopen to the parties may be compelled a range of these expectations, is not legally the price.59 Although the first-stage agreement 56 such as public or private offer One might contrast these deals with similar transactions, ings of securities, in which these terms do appear to be priced after they are settled. for Nego and Techniques of a Merger: Strategies See, e.g., James C. Freund, Anatomy 53-55 (1975). tiating Corporate Acquisitions 58 Id. at 59-60. 59 .. . bear primary responsi et al., supra note 53, at 129-35 See, e.g., Mnookin ("Lawyers the parties' preliminary understand concepts bility for translating into legally recognizable This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions The Effect of Bargaining 2012] Power 1691 binding, there would be nonlegal costs to allowing the deal to collapse after this point. This leaves lawyers with a meaningful space within which to bargain on behalf of their clients over nonprice terms. This arrangement leads to a peculiar process in the second bargaining stage between the lawyers, during which the two sides cannot use the for price term in their efforts to create value by logrolling.60 Consider, in the is set in a letter of a which price example, corporate acquisition and war intent before many of the terms—particularly, representations ranties, covenants, closing conditions, and termination rights—are nego Although the letter of intent is usually not binding, in these the for concessions price to compensate parties rarely adjust terms in either party's favor. In a collective effort to integrate the best tiated by the lawyers. field, the American Bar Association appoint practices in the acquisition ed a task force charged with producing and updating a Model Stock Pur it is particularly significant that chase Agreement.61 For our purposes, the report repeatedly emphasizes the divergent positions of buyers and the document as describes sellers. The preliminary note, for example, follows: Agreement has been prepared as a resource for a buyer's first draft of a stock acquisition agreement. In a buyer's first draft, the provisions generally favor the buyer and are not necessarily typical of The Model the final language in a fully negotiated agreement and consummated transaction.... Sellers usually will not agree to all the proposed pro visions, and their counsel can be expected to negotiate for language more favorable to them. The commentary identifies some sections of ing of their deal. In addition, that the clients may not have ancillary risks legal drafting involves identifying and allocating the problem of considered . . . .")• They later note, in discussing that "lawyers can waste the client's time and money by focusing on small or over-lawyering, Id. at 148. unlikely risks that do not justify contractual planning." 60 between terms can yield value, negotiation also warn Although logrolling specialists about the countervailing feasibility or danger of having too many issues on the bargaining To mediate between the benefits and costs of multi-issue Professor How negotiations, ard Raiffa proposed a process under which parties first agree to a simple deal and then ask table. their agents to improve it by incorporating other opportunities for logrolling. Howard Raiffa, 1 Negot. J. 9, 9-10 (1985); Post-Settlement see also Shell, supra note 28, at 186. Settlements, Raiffa labels the strategy "post-settlement settlement." Raiffa, supra, at 9. Similarly, the two described above offers a hybrid alternative stage process the details are preset and a complex in which negotiation 61 1 Am. Bar Ass'n & Acquisitions Comm. Mergers Stock Purchase & Acquisitions Agreement with Commentary, to both a simple negotiation where all terms are on the same table. on Negotiated Acquisitions, at v (2d ed. 2010) [hereinafter ABA, Comm.], This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions Model Mergers 1692 Virginia Law [Vol. 98:1665 Review the Model Agreement that are likely to prompt objections but most, if not all, provisions are negotiable.62 by a seller, The note lists three factors that may influence the scope and content of the ultimate agreement, the second of which is "the relative negotiating positions of the parties":63 Where the target is highly sought-after and there are competing offers, a seller may view some of the provisions of the Model Agreement as too aggressive or otherwise inappropriate.... On the other hand, if the target is financially distressed or the seller is otherwise in a weak bar gaining position, the buyer might be even more demanding it presents to the seller.64 in the draft We can illustrate the perspective of the task force through its com ments on two types of provisions: seller representations and the closing Seller repre condition requiring no Material Adverse Change ("MAC"). sentations are among the terms negotiated between the lawyers, and the must be true in order for the deal to close.65 The note de representations scribes the conflict between the parties: The buyer typically will ask the seller to bear most of the risk associ ated with discoveries that directly or indirectly relate to the target's business prior to the closing—issues that may be material to pricing the acquisition. The seller may counter that unknown contingencies are inherent in operating any business and should be borne by the owner of the business at the time they arise.66 under which Similarly, the definition of MAC sets the contingencies the buyer can walk away from the deal at closing. The comment states: Buyers generally prefer a broad MAC provision such as the one used in the Model Agreement. A broadly drafted MAC provision is thought to provide buyers with greater protection, as it gives buyers greater flexibility to terminate or renegotiate an acquisition agreement in the event 62 63 of unforeseen adverse events that are not described in the Id. at xi. Id. at xii. The other two factors are the size of the transaction and whether of another corporation. Id. subsidiary 64 Id. 65 The agreement survive may also provide that the representations 66 Comm., ABA, Mergers & Acquisitions supra note 61, at xi. closing. This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions the target is a The Effectof Bargaining Power 2012] 1693 agreement. . .. Sellers will want to minimize Buyers' ability to walk In this regard, Sellers away from or renegotiate the agreement.... will try to limit the definition of MAC to restrict the events or occur rences that could trigger the MAC condition.... One way Sellers may further narrow MAC provisions is by requesting exceptions ("carve outs") to the MAC The breadth bargaining conditions definition.67 of the MAC definition is perceived and business power. Lawyers were "seller-friendly" to be determined observed analysts more and contained by that MAC carve-outs when private equity firms were flush with funds before the financial crisis. Af ter the crisis, credit tightened and buyers gained bargaining power, lead ing to more "buyer-friendly" provisions with fewer carve-outs.68 Under what we have labeled the irrelevance proposition of bargaining this is Like many other terms, representations power, analysis puzzling. and warranties allocate risks and might be thought of as insurance prod ucts within acquisition For a variety of possible reasons, agreements. one party can bear the risk at lower cost, and the contract can create val ue by providing that this party will insure the other party against the risk, for a price.69 Both parties can be better off and therefore should agree to 67 68 Id. at 31-33. A report out of the Wharton Business office of the law partner in the Houston the competition for private . . . ." Knowledge@Wharton, Win the Compete, Intangibles prices, sellers School in 2007 quoted William Parish, Jr., then a firm of King & Spalding: "In addition to record equity deals is altering the terms for deals in favor of Private Equity Bidding Wars: When Capital-Rich Funds Deal wharton.upenn.edu/article.cfm?articleid=1721. industry risk from the scope of material 2-3 available at (2007), Parish cited, in particular, adverse change and the shortening http://knowledge. the carve-out of of periods during agree to indemnify buyers (from three years to a year or less). Id. In 2008, the law firm of Nixon Peabody a report of its review of acquisition dated published agreements from June 1, 2007 to May 31, 2008. Nixon Peabody Seventh Annual MAC LLP, Survey 2 available at (2008), http://www.nixonpeabody.com/linked_media/publications/MAC_ which sellers The report stated: "[WJhile the MAC definitional elements were slightly survey2008.pdf. narrower than in the prior year, we noted a decrease in the number of MAC ... exceptions the advancement of buyers' Id. at 4. The indicating bargaining power during this period." shift to more buyer-friendly terms is "likely due at least in part to a lack of credit available and sellers' understanding that they must decrease their expectations transactions, finance get a deal 69 MAC done." to to Id. clauses are understood to promote the following two objectives: first, the buyer's contingent option to terminate gives the seller the incentive to maintain the value of its assets between the time of the contract and closing (the moral hazard problem). the seller's Second, to grant such an option signals its information as to the financial and economic willingness condition subject of the target (the adverse selection At the same time, most MACs problem). to carve-outs—defined material changes that do not trigger such a termination This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions are op 1694 Virginia Law Review [Vol. 98:1665 that risk allocation, regardless of relative bargaining power. Yet, like the authors of the model agreement, practitioners frequently view these ne gotiations as zero-sum. While terms can be traded within this stage, the to trade off risk allocation inability value a price adjustment removes In this respect, bargaining terms because the outcome of the against from the table. significant potential power is a determinant of the nonprice second stage is constrained and predominantly distributional. Yet, even in this setup, the impact of bargaining power is complicat ed. Suppose that the parties agree to the price in the first stage and leave to the second stage the scope the deal under a MAC clause. er, then it might secure first stage. In the second of the buyer's option to walk away from If the seller has superior bargaining pow a higher price than it otherwise would during the stage, the parties' lawyers negotiate the carve outs from the buyer's MAC condition, among other terms. Suppose the seller derives bargaining power from its greater patience (the buyer faces more time pressure to have the deal signed). The seller's attorney can then present the buyer with a take-it-or-leave-it offer and get a more ex tensive carve-out from the MAC. The buyer may or may not be able to a on another nonprice term at this stage. A concession get countervailing buyer anticipating this in the first stage lowers its reservation price ac cordingly. Since the seller gets the greater portion of the bargaining sur plus, the buyer's anticipated exercise of the seller's power in the second stage in fact harms the seller (more than the buyer) in the first stage. Thus, the seller has the incentive to precommit to limit the scope of its bargaining power in the second stage, in order to secure a higher price. In the two-stage bargaining process, this may be difficult to do; the buy er will presume an unfavorable outcome in the second stage (of course, as noted above, within some range of expectations). As a result, the in the second stage may vary from "seller nonprice terms negotiated to on which party has bargaining friendly" "buyer-friendly," depending power. Nonprice terms may be inefficient as a result. The buyer's inability to observe private information held by the seller MACs. and "buyer-friendly" First, as may also explain "seller-friendly" tion. These downturn in such as the general exogenous contingencies, industry, that are typically outside the control and private knowledge in Contract See Albert Choi & George Triantis, Strategic Vagueness Design: We also discuss a of Corporate 119 Yale L.J. 848, 865-70 (2010). Acquisitions, the economy of the seller. The Case carve-outs describe or seller's third objective: facilitating Id. at 869-70. closing. renegotiation in case the deal turns out to be unattractive This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions before The Effectof Bargaining Power 2012] 1695 we described in the context of loan covenants, changes in the supply or for acquisitions can also exacerbate or mitigate the underlying moral-hazard and adverse-selection Given the target seller's problems. information as to its terms value, price private may function imperfectly as a means of adjusting for supply-demand imbalance or changing the demand of surplus. In particular, a price change may make moral hazard selection more severe. The parties may improve the efficien cy of their deal by using nonprice terms instead of shifting value from one party to the other.70 division or adverse Second, the asymmetric information about the target's value may en courage either the seller or the buyer to screen or signal, respectively. We demonstrate at greater length in Part IV that these actions can give rise to inefficient contract design, particularly at the extreme ends of the bargaining range at which either the seller or the buyer is the residual claimant of the surplus. the breadth of the MAC The party with such bargaining power may use clause to reduce the counterparty's information rather than to enhance the surplus by addressing the moral al advantage, hazard and adverse-selection obstacles. the seller could When the seller enjoys (com use different MACs to screen plete) bargaining leverage, different types of buyers. The buyer whose reservation value is less sen sitive to an external shock would be more willing to sign a narrow MAC than the buyer whose reservation value is more (or broader carve-outs) volatile. of different MACs with correspond By offering a combination the seller can better extract the surplus from the transaction. ing prices, A buyer with bargaining power might similarly use different MAC and to signal a more volatile reservation value, so as to price combinations pay a lower respectively, compromise demonstrate price for the target. These screening and signaling efforts, do not increase the size of the surplus. In fact, they may the MAC goals of controlling the seller's moral hazard. We in Section the efficient breadth V.C that the parties are more likely to agree to if their bargaining power is more even. of a MAC This is a rough hypothesis at this point, but it is at least an attempt to gain insight into the role of bargaining power in the design of these terms. some In sum, the bargaining rests on the power irrelevance proposition that the are risk neutral and that there are no infor premises parties mation 70 Choi imperfections & Triantis, or other transaction supra note 14 (manuscript costs. We have at 4). This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions suggested a Virginia Law Review 1696 [Vol. 98:1665 variety of ways in which are relaxed. assumptions introduction bargaining power may be relevant when these Each way can be elaborated beyond our brief in this Part and we leave this to future work. In Parts III, IV, and V, we examine in cases of asymmetric information in greater detail the impact information, particularly as to its reservation price and either tries to conceal private it or cannot readily reveal III. Product of bargaining power where one party has Quality it. Under Monopoly and Perfect Competition In the industrial literature in economics, a body of schol organization all else the equal, product quality offered by a arship analyzes whether, in is or different from that of a social monopolist perfect competition social welfare. This is relevant to our in planner seeking to maximize of the quality of the that sells a single good underlying product. Consider first the monopolist we at a single price in a given market. Under the standard assumptions, will sell a lower quantity than optimal, thus know that the monopolist quiry because the terms of a contract are elements loss. The reason is straightforward. The monopo creating a dead-weight in price brings additional list knows that an incremental decrease cus will of revenue consumers who were but at a loss from all other tomers, even at the higher price. While the second ing to make the purchase finds it cost effect would yield no loss in social welfare, the monopolist minimize To the effect of this revenue loss, the mo infra-marginal ly. nopolist charges a price higher than the marginal cost of production and serves fewer customers than its competition. When choosing the level of quality to be offered to its consumers, the chooses the quality according to the preferences of the mar monopolist ginal buyer—the buyer who is just indifferent between buying and not price. If the marginal buyer values an incre buying at the monopolist's mental increase in quality at least as much as the incremental cost to the then the monopolist improves the quality for everyone, but monopolist, not otherwise. If all buyers share homogeneous preferences for quality, then the monopolist the provides optimal quality of contract terms (albe if the marginal buyer has a it at a supra-competitive price). However, to pay for quality than the infra-marginal buyers, the the additional to purchase buyers will be compelled higher willingness infra-marginal if quality even though they would be unwilling to pay for it. Conversely, the infra-marginal customers would pay for an increase in quality, but the marginal buyer assigns an incremental valuation lower This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions than the in 1697 The Effectof Bargaining Power 2012] offers the higher quality to no one. Thus, cost, the monopolist if buyer preferences are heterogeneous, the quality offered by a monopo the list may be higher or lower than optimal.71 The social planner, on the of the average other hand, would base the quality on the preferences cremental as long as the preferences of the average buyer and the marginal buyer are not identical, the quality chosen by the monopolist social welfare. differs from that which would maximize buyer. Hence, The analysis so far assumes that the market can provide product with confer dif only one level of quality. Many types of contract provisions at the beginning of ferent values to different buyers. The three examples this Article illustrate this claim. A warranty is more valuable to a buyer who uses the good more frequently and intensively. concerned about sudden termination if it has made the franchise or if its location is more vulnerable A franchisee is more a large investment to short-term shocks. in A if it anticipates seller benefits from litigating jurisdiction more rather than fewer disputes over its performance, particularly those below, involving large monetary claims. In addition, as we emphasize in its home the cost to a contracting promisor is also likely to vary with the type of the effect of market or bargain buyer for similar reasons. In evaluating on contract terms rather than the ing power physical quality of products, not only should we assume heterogeneity but also that the market may offer more of preferences among buyers tailored products that cater to differing preferences. over quality and the When buyers have heterogeneous preferences the monopo can offer different price-quality combinations, monopolist list can increase its profits by discriminating its among buyers, on the knew each of price, quality, and contract terms. If the monopolist it would offer to each customer the quality and con buyer's preferences, basis maximize the surplus and charge a price that the monopolist to capture the entire surplus. If the monopo list had this information and discriminated, market power would not dis tort quality, and the irrelevance would be borne out. The proposition tract terms that would would allow buyer who places a higher value on warranty is offered an extended war ranty clause at a higher price, while a buyer who values it less purchases a limited warranty at a lower price. Better yet, even if two buyers place 71 See, Spence, e.g., Jean Tiróle, Monopoly, Quality, The Theory of Industrial and Regulation, 6 Bell 100-04 Organization J. Econ. 417,417-22 (1988); (1975). This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions A. Michael 1698 Virginia Law Review [Vol. 98:1665 the highest value on extended warranty, the monopolist will offer the same extended warranty to both buyers, but at different prices. Price and quality discrimination on the basis of contract might be effective because the can effectively prevent arbitrage quite monopolist where customers would sell their rights to high-value low-valuing buy ers. Warranties are often expressly nonassignable, for example, as are the franchises and purchase orders introduced earlier. The bigger prob lem facing the monopolist, rather, is that, in most cases, it lacks infor mation about its customers' individual It may nevertheless valuations. to more of the consumer attempt capture surplus through price discrimi and this lead to the of inefficient quality. nation, may supply If the monopolist cannot observe its customers' individual valuations, it might try to use variables related to their willingness to pay. For ex if tend to fewer ample, lower-valuing buyers buy products (for example, because of less wealth), the monopolist may charge a higher unit price for larger quantities, assuming it can prevent resale. Another method is to offer a range of products, or products of different quality, in order to smoke out the higher-valuing buyers.72 A monopolist seeking to maxim ize its profit may provide lower-than-competitive quality to all custom ers other than those who value quality the highest. By offering a menu of the monopolist to separates customers according price-quality options, their preference for quality.73 Whether the discrimination the lower-valuing 72 Professor high-valuation customers is by product quality may receive suboptimal or contract quality.74 terms, In fact, Tiróle offers the example of auto manufacturers extracting the surplus from who value luxury and prestige. He notes that the profit margins on cars and optional equipment are generally higher than those on basic cars and consumers top-of-the-line the existence of quality premia. He also suggests that this may lead equipment, suggesting the monopolist to offer too many products. Tiróle, supra note 71, at 158. 73Michael Mussa & Sherwin Rosen, Monopoly and Product Quality, 18 J. Econ. Theory The authors assume that the monopolist seller knows the general distribu 301, 301 (1978). tion of tastes and demands in the market, but cannot distinguish among buyers prior to sale a constant costs of producing prevent resale in other markets. They also assume costs of higher quality items. For a discussion of given quality and increasing marginal and its consequences, see David Besanko et al., Monopoly and Quali quality discrimination Effects and Remedies, 102 Q.J. Econ. 743, 743^4-4 (1987). For a discussion of ty Distortion: and cannot in healthcare, see Martin Gaynor, What Do We Know About Compe quality discrimination tition and Quality in Health Care Markets? 4-8 (Nat'l Bureau of Econ. Research, Working Paper No. 12301,2006). 4 See Besanko et al., supra note 73, at 749; Sherwin Rosen & Andrew M. Rosenfield, Ticket Pricing, 40 J.L. & Econ. 351, 352 (using example of intertemporal price discrimina This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions under 1699 The Effectof Bargaining Power 2012] some the monopolist its profits by conditions, may maximize in the customers order to extract the foregoing lower-valuing altogether, surplus from the high-demand buyers. For example, a monopolist might its with limited and discriminate product warranty giving by offering buyer an option to purchase additional warranty. By doing so, the can extract more surplus from those that place a higher value monopolist each on extended warranty. Yet, buyers who value warranties less than others end up with an inefficiently limited warranty or inefficiently broad dis claimer, in order to prevent the high-valuing buyers from pooling with them.75 When the market is perfectly competitive and the buyers' heterogene ous preferences are unknown to sellers, a different kind of inefficiency can arise. An often-cited example is that of an insurance market.76 Sup pose the insurance high chance buyers can be divided of suffering from an accident into two groups, one with a and the other whose accident probability is low. Apart from the differing chances of an accident, eve rything else is the same across the two groups, including the degree of risk aversion. Given that both groups are risk averse and assuming that the insurance sellers (companies) are risk neutral, the social-welfare solution is to provide both groups of buyers with full insur maximizing ance (without, for instance, deductibles or co-pay). When the insurance market is perfectly competitive but the insurance do not observe each buyer's risk propensity (risk characteris companies solution cannot be sustained, how tic), the social-welfare-maximizing ever. Suppose we start from the full insurance condition. A consequence of not being able that the insurance to observe each risk propensity implies tailored: both the individually consumer's cannot premium and low-risk consumers be will be paying an average premium. that the market is each company of given perfectly competitive, insurance will break even. The fering just average premium charged by the insurance companies will be just enough to cover the average ex high-risk Also, tion of tickets the revenues 75 Another to show that "catering to any subset of customer tastes that can be extracted from other groups in other classes"). example may be found in the industry that is airlines. The airlines restrict the price discrimination—the low-fare tickets in order to extract more of the surplus from would seem that an airline could provide this flexibility at many nate. 76 leisure Stiglitz travelers, it might refrain from doing & Rothschild, in one class constrains most notorious for its perhaps or cancel flexibility to change business travelers. Although it a lower so to protect cost its ability supra note 17, at 629. This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions than its value to price discrimi to 1700 Virginia Law Review will be equal to the expected or and low-risk there is, average payouts (both high-risk buyers), from the buyers' perspective, an indirect subsidy from the low-risk con sumers to the high-risk consumers. That is, the low-risk consumers are pected payouts. Because [Vol. 98:1665 the premium to all being charged a premium that is too high relative to their risk propensity while the high-risk consumers are being charged a premium that is too while each will make low; company money from low-risk consumers losing money to high-risk consumers. When the companies and the low-risk consumers realize this cross one of two will Either some will start subsidy, happen. companies things less than full deductible and/or insurance offering (with positive co-pay) with a lower premium consumers themselves, consumers, or the low-risk if they have the power to control the terms of the contract, will offer to share some of the risk in return for a lower premi um. And given that the companies were initially making profit on selling to attract the low-risk full insurance to low-risk consumers, they can design such a contract so as to keep the high-risk consumers away while making both the compa nies and the low-risk consumers better off. The first phenomenon is of in which companies skim the profitable "cream-skimming," of the while the second is called "inefficient market, signaling," segment in which the consumers signal their value to the market by taking costly ten called action (in this case, less-than-full insurance). Of course, once the low-risk consumers have been skimmed away by insurance to high the that are full companies, companies offering risk consumers will no longer break even, and the initial full insurance will fall apart. If there is an equilibrium at all, it will be one equilibrium some in which the low-risk high-risk insurance consumers insurance while the buy less-than-full full insurance, and the companies selling consumers purchase to either type will just break even by charging an actuarially fair premium.77 are For this type of "unraveling" to occur, at least three conditions pref important. First, the buyers in the market must have heterogeneous erences and those preferences must be private information for the buy ers. In the insurance market context, buyers had different risk propensi ties and that information was private. a buyer's Second, differing for must not to affect the only buyer's willingness pay quali preferences 77 See Choi & Spier, supra note 17, at 2-3, for a more under which equilibrium fails to exist. in-depth analysis of products ity and conditions This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions liabil The Effectof Bargaining Power 2012] 1701 ty but also the seller's cost of providing that quality to the specific buy er. In the insurance market, each buyer's risk propensity determines not to pay a certain premium, but also the sell only the buyer's willingness er's cost of providing insurance to the buyer. Even with the same payout amount, the high-risk buyer will be more costly to the insurance compa relevant for various ny than the low-risk buyer. This is particularly nonprice terms in contracts. Buyers may attach different values to dif ferent physical attributes of a product, but the seller's cost of producing a certain physical attribute is usually invariant to the type of buyer con suming the product. In contrast, nonprice contract terms, such as warran will not only command clauses, ty, termination, or dispute resolution different willingness to pay from a buyer but will also impose different the on the type of buyer that purchases costs on the seller depending product. or the inefficient signaling result also de Third, the cream-skimming in the market (vigorous pends on either the presence of many companies or the to dictate control or the terms of the competition) buyer's ability In other words, the market is heavily skewed in favor buyers and they have all the bargaining power against the sellers. some reason, there isn't as much competition among the sellers terms of trade are to control the restricted, one buyers' ability contract. of the If, for or the would suspect that the inefficiency result may be mitigated or even disappear. In fact, the distortions caused by a monopolist in perfect competition seem to rely heavily on the condition that the market is very one-sided. Yet, we do not yet have a very good idea of what may happen when the In Part V, we attempt to demonstrate how when the market conditions disappear provide a market is more even handed. such distortions could more even playing field to contracting parties and, in the process, the gap between the irrelevance proposition and the practitioners' standing of the importance IV. Effect of Uneven of bargaining Bargaining bridge under power. Power Under Asymmetric Information one seller and one buyer contract over the sale of a product. Suppose How much the buyer values the product (her reservation value or will ingness to pay) and how much the product costs the seller to produce depend on two factors: buyer type and product quality. Starting with quality, the higher the quality, the more costly it is for the seller to pro to pay. For instance, if qual duce, but the higher the buyer's willingness This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1702 Virginia Law Review [Vol. 98:1665 ity is represented by the warranty that comes with the product, a more extensive warranty will impose a higher cost on the seller but will also increase the maximum the buyer would be willing to pay for the prod uct. Similarly, in the seller's if the contract obligates the buyer to resolve dispute only state (or grants the franchisor a broad termination right), such a restrictive forum selection (or a broad termination) clause will re duce both the seller's (the franchisor's) cost and the buyer's (the fran to pay. A forum selection clause that restricts liti willingness state (or a broad termination right) can also be gation to the seller's of as thought providing low quality to the buyer (the franchisee).78 chisee's) How the buyer values quality and how much it costs the seller a certain level of quality will also depend on various buyer much to produce specific factors. in A more extensive user will value an improvement A user will also more than a less user. more frequent warranty frequent impose a higher warranty repair cost on the seller. Similarly, a litigious buyer may place a higher value on the right to bring a lawsuit in her home jurisdiction, and for the same reason this right is more expensive for the seller to provide. Finally, a franchisee with whom the franchisor is more likely to terminate the relationship will value restrictions on the right more than the franchisee who is likely to in a longer relationship with the franchisor. We aggregate these under the rubric of "buyer type."79 We assume also that each franchisor's engage factors termination buyer type's valuation of the incremental value of the contractual provi sion is correlated with its respective valuations of the basic product.80 To succinctly represent these ideas, assume that the product can be levels of quality: high or low. High a higher cost on the seller but also increases the quality product imposes to pay. In addition, the buyer can be of two different buyer's willingness types: type 1 or type 2,81 where the probability that the buyer is of type 1 manufactured 78 at two different cost of other types of terms may be much less affected by buyer type, such as a 739 (showing that stipulates damages. See Ayres & Gertner, supra note 4, at 736-37, dam that a competitive market leads to more efficient contracting around default liquidated The clause ages provisions). 19 of a used car, known This is qualitatively similar to the story of how the true condition value but also how much the buyer only to the seller, affects not only the seller's reservation is willing to pay for the car. See George A. Akerlof, The Market for "Lemons": Qualitative and the Market Mechanism, 84 Q.J. Econ. 488,489-92 (1970). Uncertainty 80 more frequent and intense use of As mentioned earlier, it may be that a type-1 buyer's the basic product leads her to value more both the product and the contractual warranty. 81 while the seller type is Here we assume that there are two potential types of consumer fixed. This is done to simplify the analysis. The main implication of assuming no private in This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions The Effectof Bargaining Power 2012] 1703 is 9 G (0,1) and of type 2 is 1 — 6. For most of the analysis, we will as sume that Q = 1/2: each buyer type is equally likely. Both types of buy er value high quality more than low quality, but depending on quality level, the type-1 buyer is willing to pay more for the product, and it is also more costly for the seller to provide the quality for that buyer type. We can think of the type-1 buyer as the more frequent user of the prod uct and the type-2 buyer as the less frequent, casual consumer. Since the type-1 buyer is more likely to make a warranty claim, warranty is valued of any given more by the type-1 buyer. At the same time, provisions level of warranty to the type-1 buyer are also more costly for the seller. The following table summarizes the monetized values and costs that de pend on both buyer type and product 1: Production Table Costs Type-1 Product and Reservation Buyer Values Type-2 Buyer Low $190 High $250 $170 High $200 $70 $120 $100 $150 $50 $120 $70 $130 Low Quality Reservation quality. Value Production Cost Surplus Note that when the quality of the product is low (for example, warran ty is limited, forum is restricted to the seller's state, or broader termina tion right to the franchisor) the type-1 buyer is willing to pay up to $190 while the type-2 buyer is willing to pay up to $170 for the product. For the seller, it costs $70 to offer low quality to the type-1 buyer and $50 to the type-2 buyer. The values and costs for the high-quality product extensive warranty, no restriction on forum, or narrower (for example, termination right) are analogous.82 If we define social welfare as the buyer's willingness to pay minus the seller's cost, by assumption, the numbers in Table 1 imply that to max formation profit will on the seller's be reduced power. 82 By assumption, crease in reservation posed to $30 ($250 versus side is that when to zero. This will the buyer has all the bargaining power, the seller's not be true when the seller has all the bargaining the type-1 buyer not only has a higher marginal value of $60 when switched from low quality for the type-2 and $190 but also to pay (in willingness to high quality, as op to pay for quality willingness higher absolute this assumption in Although may be reasonable many settings, if it does not hold, it may be the case that the party with all the bargaining power might offer nonprice terms that are inefficiently high rather than inefficiently low. $200 buyer) versus $170). This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1704 Virginia Law Review [Vol. 98:1665 imize social welfare, the seller should provide high quality to both types of buyer.83 By doing so, the surplus of $150 is realized from the type-1 buyer and $130 from the type-2. In expectation, the maximum expected social welfare is 9 x $150 + (1 — 0) x $130. When 9 = 1/2, this is to $140. we assume that the buyer and seller realize zero if there is no sale. These are the parties' utility and profit, respectively, outside reservation values. respective equal Finally, A. The Irrelevance Proposition Under Complete Information When both the buyer and the seller are fully informed of each other's and costs, regardless of the distribution of bargaining power, they will negotiate to achieve the surplus-maximizing result. They will choose the quality level to maximize the total surplus from the transac values tion while working out the bargaining issue through price. For example, the seller has all the bargaining power. The seller, knowing to suppose which type of buyer she is selling the product, will sell high-quality product to both types of buyer but charge two different prices: $250 to the type-1 and $200 to the type-2.84 By engaging in perfect price dis its expected profit by capturing all of crimination, the seller maximizes the surplus: $150 er for an expected from the type-1 buyer and $130 from the type-2 buy — if profit of 9 x $150 + (1 9) x $130. Conversely, the buyer has all the bargaining power, the type-1 buyer will offer to purchase high quality at $100, and the type-2 buyer will offer $70 for high quality. The expected buyer surplus will be equal to the maximum social welfare of 9 x $150 + (1 — 9) x $130. The only difference be tween these two polar cases will be the price at which the parties reach an agreement. More generally, if we let A G [0,1] denote the fraction of the surplus that the seller captures in equilibrium (or the seller's relative bargaining power vis-á-vis the buyer), while both types of buyer will purchase high quality product, the equilibrium price for the type-1 buyer is $100 + 83 The assumption that a single level of quality (high quality in the example) maximizes the surplus from both buyer types is not important but is used to simplify the analysis. In the we provide a model in which optimal qualities differ based on buyer type. Appendix, 8 To make sure that the buyer will make the purchase, the prices have to be slightly less than the buyer's reservation for example, and $199.99. This type of tie $249.99 value, will be common the numerical breaking throughout examples, assume that when the buyer (or the seller) is indifferent between ing or not selling), the buyer will purchase (the seller will sell). and, for simplicity, we will and not buying (sell buying This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions The Effectof Bargaining Power 2012] 1705 Note that as A ris and, for the type-2 buyer, $70 + A($130). A($150), = es, so do the equilibrium 1, denoting full bargaining prices. When A for the seller all the power (or giving surplus to the seller), prices equal the buyer type's respective to pay. Similarly, when A = 0 willingness (when the buyer has all the bargaining power), the prices equal the sell er's respective costs. Table 2: Equilibrium Under Symmetric Information A e [0,11 Type-1 Buyer Type-2 Product Equilibrium Quality Equilibrium Price Consumer Producer Total Surplus Surplus Surplus High $100 + A($150) (1 — A)($150) A($150) $150 $70 + A($130) (1 — A)($130) A($130) $130 This result yields the bargaining of complete strong assumption tive of their relative Buyer High bargaining power irrelevance proposition and symmetrical information: power, contracting under the irrespec parties will always terms and work out choose the efficient, surplus-maximizing nonprice the bargaining power issue only through price. As we will see in the next of bargaining infor section, the combination power and asymmetric mation leads to inefficient nonprice terms. B. Private Information on Buyer Type that the buyer knows how much she is willing to pay for quality (which type she is), but the seller does not: when the seller meets the buyer, the seller only knows that the buyer is type-1 with probability 6 or type-2 with probability 1 — 0.85 Under this assumption, the alloca tion of bargaining power determines whether the parties will agree to an Suppose efficient quality of nonprice term. As in the complete, symmetric infor 85 We assume that it is the seller, not the buyer, who lacks the relevant knowledge about the buyer's preferences. This assumption seems realistic since, presumably, the buyer knows more about her preferences than the seller. We can flip the assumption and let the seller be aware of the buyer type while the buyer is ignorant of her preferences, but this will not results. In reality, private information will run on both sides: while the change the qualitative she wouldn't know much about buyer would know more about her preferences necessarily the seller's cost structure. We stay away from such complication to keep the analysis tracta ble. This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1706 Virginia Law Review [Vol. 98:1665 mation case, we first start with two polar cases: when either the seller or the buyer has all the bargaining power. In the next Section, we turn to the more complicated of "even" distribution of bargaining examples power.86 1. Dominant Seller the seller has all the bargaining power. In this Subsection, Suppose we represent the bargaining power as the ability to make a take-it-or leave-it offer to the buyer, without competition, which cannot be subse If the seller were to provide high quality to both quently renegotiated. buyer types, the seller would not be able to charge two different prices for high quality since she would not know which type of buyer she was dealing with. Unless the price is so high to keep the type-2 buyer from purchasing at all (p > $200), both types of buyer will simply choose the offer with the lower price. With that constraint,87 the profit-maximizing price the seller can offer for high quality and still be able to sell to both If the seller were to charge a higher price, the types of buyer is $200. would not buy, and lowering the price would only increase type-2 buyer the buyer's surplus and reduce the seller's profit. With $200, and when both buyer types accept the offer, the seller's expected profit is $200 — 86 in this Section is known as bargaining games with pri is re as soon as the informed party's private information models, of vealed to the uninformed agreement or complete convergence party, there is an immediate with Private Information, & Robert Wilson, Bargaining posterior beliefs. See John Kennan 31 J. Econ. Literature 45, 45-50 for a survey of this class of bargaining games. With (1993), vate The class of models information. we present In these that analyzes in the game theory literature, there is a different strand of models bargaining of the informed party's in with "non-convergent" priors, in which even after the revelation about the state of the world, or no immediate formation, there is no immediate agreement Without a of players' Yildiz, Bargaining convergence posterior beliefs. See, e.g., Muhamet 71 Econometrica Prior—An Immediate Common 793, 808-09 Theorem, (2003). Agreement cases, whether the parties will agree immediately (whether there will be an ineffi will depend a lot on how fast each player will be able to update her cient delay in agreement) For the sake one remains after a communication). beliefs (e.g., how optimistic or pessimistic of tractability, we do not deal with this latter, important strand of literature. 87 the seller might If the proportion of high-value buyers is sufficiently high (0 > 13/18), In those a single contract that is attractive only be able to increase her profit somewhat by providing is using to a type-1 consumer: high quality at a price of $250. In our model, the monopolist Alan Schwartz that terms to screen buyers. Professor both the price and nonprice suggests when trade is uncertain ex post and the buyer is privately informed of the surplus, the mo to screen buyer (down-payment) may use the initial price and liquidated damages are similar to as being the We can think of which warranty, damages, types. liquidated with Contract Price Discrimination term in our model. See Alan Schwartz, "nonprice" nopolist Terms: The Lost-Volume Problem, 12 Am. L. Econ. Rev. 394, 413-16 (2010). This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions The Effectof Bargaining Power 2012] 1707 — for certain, and {0 x $100 + (1 0) x $70}. The seller will get $200 will incur an expected cost of 0 x $100 + (1 — 0) x $70. When 0 = 1/2, the expected cost is $85, and the expected profit is $115. When the seller cannot identify the buyer type, the seller's ability to limited. The problem capture surplus from the type-1 buyer becomes with offering high quality to both buyer types is that, although it is so from the seller's cially optimal, it is not profit-maximizing point of view. At $200 for high quality, the type-1 buyer realizes a surplus of and this represents a foregone opportunity (an opportunity cost) to the seller been able to identify the buyer type, she would have been able to engage in perfect price discrimination and earned an $50 the seller. Had additional $50 from the type-1 buyer. Due to the buyer's private infor even mation, though the seller has all the bargaining power, the seller is letting the type-1 buyer enjoy a significant amount of surplus. When faced with such information obstacles, a menu of offers with different levels by making stead of offering high quality at $200, the seller can do better of quality. Suppose, in the seller makes the following menu of offers: (jpvq{) = ($230, h) and (p2, q2) = ($170,/). That is, the buyer is given a choice between purchasing high-quality product at or low-quality Each type of buyer, presented $230 product at $170. with such a choice, will choose whichever maximizes her surplus. For the type-2 buyer, since she is willing to pay only up to $200 for high the first offer is quality, clearly unattractive. With respect to the second for low quality, she would be to pay $170 offer, given her willingness willing to choose that option, although she will earn no surplus. When p2 is slightly below $170, the type-2 buyer will choose the second op tion. What about for the type-1 buyer? If she were to accept the second offer, since she is willing to pay up to $190 for low quality, she would realize a surplus of $20. Similarly, if she were to accept the first offer, her surplus would be also $20. Again, when p1 is slightly below $230, the type-1 buyer will choose the first option. When the buyer type is thus separated, the seller will enjoy a larger expected profit. Recall that when the seller was offering high quality to both buyer types at $200, the seller's expected profit was 9 x $100 + — The seller's (1 0) x $130. expected profit, when the buyer type is — — separated with the menu, is 9 x ($230 $100) -I-(1 9) x — — = x x 9 $130 + (1 $120. Compared to the previ ($170 $50) 0) ous case, the seller is grabbing a higher fraction of the surplus from the versus $100) while sacrificing some profit with re type-1 buyer ($130 This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1708 Virginia Law [Vol. 98:1665 Review versus $130). So long as the chances of spect to the type-2 buyer ($120 not too the are small, making such a tradeoff will facing type-1 buyer = make sense for the seller. When 0 1/2, for instance, the seller's ex The following table will increase from to $125. $115 pected profit compares Table the two outcomes. 3: Equilibrium ing Power Comparison When Seller (p,q) = ($200, h) e = 1/2 Has All the Bargain (Pv Ri) = ($230, h) Cp?,q?) = ($170,0 $20 Type-l's Surplus $50 Type-2's Surplus $0 $0 Seller's (Expected) $115 $125 (Expected) $140 $135 Profit Total Surplus An important point about the example is that, even though offering high quality to both buyer types is socially optimal, the seller is deliber suboptimal quality for the type-2 buyer. Such reduction ately choosing in quality stems from the seller's desire to exercise its bargaining power and maximize profit. When the seller can dictate the terms of the trade, claimant of the transaction. the seller becomes the de facto residual the seller knew which buyer type she was facing, she was able to capture all the contractual surplus by selling the same high-quality prod uct at two different prices. When the seller cannot engage in such perfect due to lack of information on buyer type, she is in price discrimination When inefficiency in the transaction to extract more of the buyer's surplus. In the current example, by offering low-quality product with a price that is sufficiently unattractive to the type-1 buyer (but at tractive to the type-2 buyer), the seller can induce the buyer to "reveal" clined to introduce the type-1 her type and is better able to reduce, albeit not completely, contractual however, surplus for the buyer's surplus. In the process, type-2 buyer is inefficiently reduced. suppose we equate Using the product warranty as an example, low with "limited" and with "extensive" quality warranty quality high war ranty and let the type-1 buyer serve as the frequent user of the product and the type-2 buyer as the infrequent user. When the seller was offering $200 for the product with extensive warranty, the extensive user was en This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions The Effectof Bargaining Power 2012] 1709 a surplus of $50, but both types of user were able to enjoy the so her cially optimal level of warranty. When the seller wants to maximize profit, instead of offering the product with extensive warranty at $200, joying the buyer a choice: the seller gives the buyer can purchase the product with limited warrantyat $170, but by paying an additional $60, she can get an extensive warranty. With these choices, the casual user will not find it worthwhile to pay $60 to obtain the extensive warranty while the extensive user will. The seller will increase her expected profit from $115 to $125 and reduce the type-1 buyer's surplus from $50 to $20, but the type-2 buyer will be stuck with an inefficiently limited warranty. The fact that the seller's bargaining power is playing an important role can also be demonstrated using the following thought experiment. due perhaps to regulation, that the seller cannot charge more Suppose, of this cap, if the seller than $210 for the high-quality product. Because she would have to leave a larger sur were to price-quality discriminate, the seller's power plus for the type-1 buyer. At the same time, because of extracting surplus from the buyer is more limited, her incentive to in troduce inefficiency to the type-2 buyer is also reduced. To see this, if the seller were to offer two different contracts to separate buyer types, she would now offer(pv q{) = ($210, h) and (p2< R2) = ($170,1). Be cause bargaining power, her expected profit is is no higher than the profit the seller could gen of the limit on the seller's reduced to $115, which erate by offeringboth types high quality at $200. If the price ceiling is between and $210, the seller no longer has any incentive to en discrimination. This example demonstrates that the $200 gage in price-quality incentive to produce quality distortion depends crucially on the party's ability to extract surplus from the other—in other words, its relative bar gaining power. 2. Dominant Buyer The quality distortion in the previous example resulted from the sell er's desire to minimize the buyer's rent. It is natural to ask whether shift power to the buyer would correct the distortion. Unfortu the buyer introduces a different kind however, fully empowering nately, of distortion to the transaction. We turn to the case in which the buyer ing bargaining has all the bargaining power and the buyer is allowed to make a take-it or-leave-it offer (without competition from other buyers or an opportuni This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions [Vol. 98:1665 Virginia Law Review 1710 to the seller.88 Given the parameters in our example, ty to renegotiate) to having offer is equivalent make a take-it-or-leave-it the buyer letting number of sellers in the market, in which a large perfect competition the most attractive offers and the fully-informed buyers choose them.89 among When the buyer has all the bargaining power, the seller's equilibrium make This result is in contrast profit will be reduced to zero (in expectation). to the previous case where the seller, with full bargaining power, was eliminate the type-1 buyer's surplus. The reason unable to completely for the difference information stems about that while the buyer has from the assumption her preferences, the seller does not. There is private no seller "type" that is kept hidden from the buyer in our analysis. When the buyer has all the bargaining power, the buyer will be able to, in equi librium, capture the entire surplus from the transaction. If social welfare were to be maximized, both types of buyer should offer to purchase high = 9 x cost of production: p quality at a price equal to the average $100 + (1 — 0) x $70, which is equal to $85 when 6 = 1/2. The problem with this solution, however, is that the type-1 buyer is receiving a great benefit by paying a price that lies below the production while the type-2 buyer is paying a price cost for that type (p < $100) higher than the production cost (p > $70). The type-2 buyer indirectly subsidizes the type-1 buyer, and the type-1 buyer captures more than the — while the type-2 buyer from the transaction ($250 p > $150) surplus — when the the < less warranty example, ($200 p $130). Using gets seller offers the product with an extensive warranty at a single price may equal the average cost of servicing both types), the infre the more frequent users of the product. users will be subsidizing quent health insurance contracts. When an in be that of Another example may (which company charges an identical premium (with imperfect screen to both the healthy and the less healthy consumers, the healthy con ing) the less healthy. sumers will be subsidizing surance break this indirect subsidy and enjoy Can the type-2 buyer somehow a larger surplus? The fact that the seller cannot identify buyer type and 88 in that case, is often of inefficient quality to some segment of consumers, problem rather than "inefficient as "cream-skimming," See, e.g., Choi & Spier, su signaling." pra note 17, at 27. 89 offer is of allowing the buyer to make a take-it-or-leave-it Indeed, the theoretic approach not significant. Even if the buyer still gets to make a take-it-or-leave-it offer, so long as there The noted are multiple sellers and no entry barrier, the equilibrium presented below will hold. This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions The Effectof Bargaining Power 2012] that the type-2 buyer is subsidizing type-2 buyer will have an incentive would 1711 the type-1 buyer implies that the to make a differentiating offer that make better off both her and the seller. Consider this deviation: of offering high quality at price equal to 9 x $100 + (1 — 0) x $70, suppose the type-2 buyer offers to purchase low-quality product at price $51. If the seller were to accept this offer, knowing that this is coming from the type-2 buyer, the seller would realize a profit of $1, as instead opposed to just breaking even. For the type-2 buyer, making this (unilat a surplus of is better, since by doing so, she realizes eral) deviation $119, as opposed to $115 (assuming 9 = 1/2). When the type-2 buyer thus realizes that she is paying too high a for the to indirect subsidy to the type-1 price high-quality product (due she has an incentive to herself to get a better deal. Of buyer), separate course, when the type-2 buyer thus deviates, the seller will no longer break even by serving only the type-1 buyer at the average cost price — (p = 9 x $100 + (1 0) x$70). fall apart and the only possible The initial (pooling) equilibrium will equilibrium is for the type-1 buyer to of fer (pv q{) = ($100, h) and the type-2 buyer to offer The seller will break — even when serving both types at the re prices. The type-1 buyer would not want to mimic the type-2 buyer. If she were to do so, her surplus would only decrease from $150 to $140. Likewise, the type-2 buyer would not want to mimic the type-1 ($50,/). spective buyer since that would reduce her surplus from $120 the equilibrium results. lowing table summarizes Table 4: gaining Equilibrium Power 9 = 1/2 Surplus Type-2's Seller's (Expected) Surplus Profit Surplus When (p,q) = ($85, fi) Type-l's Total Comparison (Expected) $165 $115 Buyer to $100. Has All The fol the Bar (PvRi) = ($100, ft) (P?,qt) = ($50,0 $150 $120 $0 $0 $140 $135 When the buyer has all the bargaining power, the buyer will deliber offer ately suboptimal contract terms so as to increase her gain. The rea son stems from the bargaining power and the temptation to signal her This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions Virginia Law Review 1712 [Vol. 98:1665 type to capture a bigger surplus. When the seller cannot distinguish be tween buyer types, the socially optimal equilibrium might force certain buyer types to subsidize by paying a higher price than justified by the = cost. When the 0), production buyer has all the bargaining power (A because she is the de facto residual claimant, such a cross-type subsidy for her, creating an incentive for her to engage in costly, but As the buyer's bargaining signaling. strength decreases (A an incentive to separate herself through inef gets higher), she has less of ficient signaling because she will not be able to capture the full benefit is a burden inefficient, from doing so. V. Effect In the previous of More two polar Even Bargaining Power the party with all the bargaining examples, inefficient terms to capture more of the con power deliberately imposes tractual surplus. In this Part, we explore whether these terms would change if the bargaining power were more evenly distributed. But, what of bargaining power? We sug would constitute a more even allocation seller might face a threat of First, a monopolist gest three possibilities. future competition if its contract terms are inefficient, but not otherwise. the monopolist Second, may be tempted to renegotiate after the initial sale, to profit from the surplus created by removing the inefficiency. We examine these cases in Sections Y.A and V.B, below. Third, in Section in in a bilateral monopoly, V.C, we present an analysis of bargaining which the parties may trade offers and counteroffers. In this game, bar gaining power can be adjusted by varying the rate at which the payoffs from future agreements are discounted to the present.90 In the Appendix, we present a more general model that does not rely on any specific bar gaining protocol but does allow the mechanism designer (social planner) over buyer's and to implement the solution based on her preferences seller's welfare. 90 In the complete, share of bargain case, reflecting a more "even" symmetric information A 6 [0,1]. In the parameter and was done by adjusting ing power was fairly straightforward the presence of private information, this is not as straightforward, the equilib partly because as to who rium tends to be sensitive to the structure of the bargaining Assumptions process. gets to make an offer first, whether the offeree can make a counteroffer, and how much delay the equilibrium of the game. This, in turn, there is between offers can matter in determining makes it more difficult to make a strong generalization about the effect of bargaining power on nonprice terms. The ideas of how deviations duce or eliminate three variations, therefore, are meant to illustrate the main following offer bargaining models can re from the simple take-it-or-leave-it the inefficiency. This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions The Effectof Bargaining Power 2012] 1713 A. Threat of Competition between In this first variation, after the initial period of negotiation the buyer and the seller, the seller (now called the incumbent) will face a competitor (called the entrant) in the market with some delay. The intro duction of competition has two important implications. First, it will keep in incumbent will be the incumbent's check so that the pricing power more im unable to extract as much surplus from the buyer. Second, will also diminish or eliminate the incumbent's portantly, competition to impose inefficient nonprice terms on the buyer. This is be an inefficient term offers a profit opportunity for the entrant. incentive cause When an entrant sees an inefficient term, the entrant will recognize that not all the potential surplus is being realized. The entrant will compete with an efficient term and induce the buyer to breach the contract with and the incumbent. The possibility of breach will make discrimination imposing inefficient nonprice terms more difficult for the incumbent. To represent these ideas more formally, we take the previous seller offer game and turn it into a two-period take-it-or-leave-it competi = the buy nature determines with (t 0), tion/entry game delay. Initially er type and only the buyer observes the type. In the first period (t = 1), the seller) makes an offer to the buyer with the incumbent (previously, out knowing the buyer's type. As in the seller take-it-or-leave-it game, the incumbent can either make a single/pooling offer ((p, q)) or a menu of offers ((pi,/i), (p2, 0)- The buyer either accepts or rejects. After the = 2) with some buyer's action, the game moves to the second period (t G To we use a discount rate of 8 delay. represent delay, [0,1]. If any of the contractual surplus is realized in the second period, rather than in the first, all payoffs that come from the second period surplus are multiplied ("discounted") by 8. A higher 8 implies that the second period payoff is less discounted vis-á-vis the first period payoff and this provides less of an incentive for the players to reach when 8 is low, the players an agreement in the first period. will have a stronger incentive to Similarly, reach an agreement in the first period. When <5 = 1, as an extreme case, the parties are indifferent between realizing a payoff in the first or in the second period. There is no cost in and the buyer would be happy to simply wait until an entrant ap in the market, making the game identical to the one in which the pears offer to the seller. That is, buyer was able to make a take-it-or-leave-it with a higher 8, there is more robust competition between the incumbent and the entrant. If 8 = 0, on the other extreme, having the option of be delay This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1714 Virginia Law Review [Vol. 98:1665 ing able to wait for the second period becomes useless. The players must reach an agreement in the first period if they are to realize any surplus, making the game identical to the one in which the seller was able to make a take-it-or-leave-it the second offer. With a lower S, potential competition in to the buyer and the incumbent has a the entrant. The discount factor 5, hence, means less period stronger upper hand vis-á-vis also determines the degree of competition between the two entities. In the second period (t = 2), a competitor (the entrant) appears in the market. If the buyer rejected the incumbent's offer in the first period, the incumbent and the entrant will make competing offers to the buyer in the offer in the first period. The buyer's rejection of the incumbent's second period implies that the buyer's type remains unknown to both the in cumbent and the entrant.91 When two sellers thus compete for a single is for both buyer whose type is unknown, the unique Nash equilibrium sellers to make an identical menu of offers: ((px, h), (p2< 0) = (($100, h.),($50, /)). The equilibrium will be the same as the one in which the buyer was making a take-it-or-leave-it offer to the seller. With between two sellers is strong only two types of buyers, competition to create a (but inefficient) enough perfectly competitive equilibrium. If the buyer accepted the incumbent's offer in the first period, how the entrant's appearance in the market will affect the equilibrium depends on the efficiency of the incumbent's nonprice term. When the incum bent's realized nonprice term is efficient, since all the potential surplus is being by the incumbent and the buyer, the entrant cannot offer any set of terms to successfully lure the buyer away from the incumbent. Hence, the initial contract between the incumbent and the buyer will stand. When the incumbent's term is inefficient, on the other hand, nonprice the entrant can successfully ficient nonprice term. Even induce the buyer's breach by offering an ef if the buyer has to pay the incumbent expec tation damages,92 the presence of a residual surplus implies that the en trant can still make both the buyer and itself better off through breach. 91 It is also that only one type of buyer accepts the offer while the other does not. possible will reveal the buyer's type to the incumbent and the entrant. This type of separation is dealt with through refinements, are not dealt with in detail. which, due to their complexity, This The equilibria are constructed to survive the refinements. presented in all three variations 92 The optimal response when faced with entry, is to set liquidated dam by the incumbent, This will allow the incumbent to ages at an amount higher than the expectation damages. extract more rent from the buyer-entrant duo. See Philippe to Entry, 77 Am. Econ. Rev. 388, 396-97 tracts as a Barrier sume that the court would honor such a penalty clause when & Patrick Bolton, Even if we were (1987). all three parties are aware Aghion This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions Con to as of the The Effectof Bargaining Power 2012] 1715 For instance, suppose the incumbent makes a menu of offers to the in in the first buyer period, ((p1( h), (p2, Q), and the buyer self-selects accordance with her type. In the second period, the entrant, after observ that the ing type-2 buyer has chosen the low-quality contract, will selec contract to the type-2 buyer. Since there is tively offer a high-quality $10 of residual surplus from switching the type-2 buyer from low quali even when the type-2 ty to high quality (surplus of $130 versus $120), ~ $50 to the has to of incumbent, buyer pay expectation p2 damages there still is enough to make both the entrant and the type-2 buyer better off. When the entrant thus attempts to induce the type-2 buyer to breach contract, the incumbent will respond by also offering high quality to the type-2 buyer.93 The result will be that the type-2 buyer will the initial be able to obtain the residual p2 — $50. high-quality surplus of $10 product at a price of $70 and capture while the incumbent's profit remains all at 1 represents the potential outcomes of the competition/entry After nature makes its the selection game.94 (at top of the tree), the in cumbent (the seller) makes an offer to the buyer, which the buyer either Figure or rejects. The bottom numbers represent the (expected) surplus For simplicity, captured by the buyer and the incumbent, respectively. the payoffs to the entrant are not shown, and the second period actions accepts are folded into the payoffs. The dashed curve represents the assumption relevant based on type, setting ineffi values, which is the result when the buyer separates not prevent the type-2 buyer from obtaining would high liquidated damages high cannot commit not to renegotiate the liqui quality in the second period when the incumbent dated damages clause. 93 When the incumbent himself offers the high-quality contract in the second period, this ciently will lead to modification or renegotiation of the initial contract. We will analyze the renego in more detail with the third variation. it So, for the sake of distinction, possibilities that the type-2 buyer will breach the initial contract with the in might be easier to suppose cumbent and purchase high quality from the entrant at $70. 94 Note that the diagram already partially reflects both pooling and separating of equilibria the game. It is not the usual extensive tree form representation of the game. This will be true tiation for all the tree diagrams in the Article. In a true extensive tree representation of the game, for action by the buyer, the seller will form a belief that assigns probabilities of <r e [0,1] on the buyer being type-1 and I — a of being type-2. The equilibrium concept we are using here is known as Perfect Bayesian-Nash See Robert Gibbons, Game Equilibrium ("PBE"). each Theory concept. for Applied Economists 149-52 (1992), We also apply the Cho-Kreps "intuitive for an easy exposition of this equilibrium criterion" to rule out any unreasonable off David M. Kreps, Games and Stable Signaling belief. See In-Koo Cho & the-equilibrium 102 Q.J. Econ. The 179, 201-04 Equilibria, (1987). portant role in the third variation in Section V.C. intuitive criterion will This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions also play an im 1716 Virginia Review [Vol. 98:1665 is making the offer in the first period, the in not know which node she is at, that is, she does not know the incumbent that when cumbent Law does the buyer type. First, note that when the buyer rejects the initial offer from the in cumbent (represented written next to them), by branches with "Reject" the competition between the incumbent and the entrant in the second pe that the buyer captures all the surplus. The type-1 buyer a surplus of $150 (multiplied by 8 due to delay) while the type-2 buyer will realize a surplus of $120 (multiplied by 5). Second, whet) the incumbent induces the type-2 buyer to accept low quality in the first period (represented branch that follows the by the "Accept" riod ensures will realize of offers for the type-2 buyer), both the entrant and the incumbent will offer high quality to the type-2 buyer in the second period and in the initial duce the type-2 buyer to breach (or anticipatorily repudiate) menu The type-2 buyer will switch to the high-quality contract while the of p2 — $50. The type-2 buyer, in the expectation damages paying process, will capture the residual surplus of $10 (multiplied by S to rep — $50. resent delay), while the incumbent's profit stays at p2 contract. Figure 1: Competition/Entry Game with Breach Nature /2S0 - p\ Ip-iooJ /5l50\/250 - Pi\/S150\/200 K o iU-iooJl - p\ /S120\A70~ o /1 p —70J v 0 A p2 + SlO\fSl20\ P.-50 This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions A o ) The Effectof Bargaining Power 2012] 1717 What will be the equilibrium of this competition/entry game? In short, the threat of having to face a competitor in the second period induces the incumbent to both lower the offer price and not impose an inefficient = 0.5, for instance, the nonprice term on the type-2 buyer. When <5 the to make a pooling offer of is for incumbent unique equilibrium = (p, q) ($140, h) and for both types of buyer to accept the offer in the first period. The equilibrium price is substantially lower than the type-2 to pay for high quality ($200). The type-1 buyer buyer's willingness will realize a surplus of $110 and the type-2 buyer will realize a surplus of $60. The incumbent, when 6 = 1/2, will realize an expected surplus of $55. Compared to the game where the incumbent was able to make a offer to the buyer, even though the game ends in the first period, the buyer enjoys both a larger surplus and efficient nonprice terms. Table 5 summarizes the outcome of the game when S = 0.5. take-it-or-leave-it Table 5: Equilibrium e = 1/2 Type-l Buyer's Surplus Type-2 Seller's Buyer's Profit Surplus Total (Expected) The of Competition/Entry Surplus Game When S = 0.5 (p,q) = ($140, h) $110 $60 $55 $140 reason why the incumbent offers a lower price to the buyer is straightforward. When the buyer is aware of the potential competition in the second period, the buyer becomes unwilling to accept a high price offer in the first period. The type-1 buyer, for instance, knows that if she were to wait until the second period, she would be able to obtain high quality at a price of $100 and realize a surplus of $150. Delay imposes some cost, so that the surplus of $150 from the second period, when S = 0.5, is equivalent to an immediate, first period surplus of $75. For the type-1 buyer to accept the high-quality offer from the incumbent in the first period, given that the type-1 buyer is willing to pay up to $250 for high quality, the price must be $175 or lower. Since the incumbent, when endowed with the power to make a take-it-or-leave-it offer without competition, was offering $230 to the type-1 buyer, the type-1 buyer is already enjoying an additional (potential) surplus of $55. Similar logic also applies to the type-2 buyer. What is more interesting and somewhat less intuitive is (1) why the incumbent is disinclined to offer low quality to the type-2 buyer and (2) This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1718 Virginia Law Review [Vol. 98:1665 why the type-2 buyer is unwilling to signal its type to the market by re jecting the incumbent's pooling offer. To better understand the underly for the still makes a moment, that the incumbent ing logic, assume, menu of alternatives for the buyer, ((Pi,/i), designed to induce (p2,0), the type-1 buyer to choose high quality and the type-2 buyer to choose low quality. When the incumbent was able to make a take-it-or-leave-it offer to the buyer, the incumbent only had to make sure that the terms of the low-quality contract that the incumbent offered were sufficiently un attractive to the type-1 buyer. This constraint was satisfied when the — — p2. With the threat prices were chosen to satisfy $250 px > $190 of competition, the incumbent now also needs to worry about trant's offer to the type-2 buyer in the second period. the en in the second period means that even if the type-2 buyer Competition had chosen low quality in the first period, the type-2 buyer would still be able to obtain high quality through breach when an entrant emerged. is that this will not only improve the welfare of the more importantly, will also lessen the incentive of the type-2 buyer, but, to with the high-quality contract in the first period. stay type-1 buyer the low-quality contract that choosing The type-1 buyer now realizes What is interesting not necessarily mean she will be stuck with low quality. Rationally and correctly expecting that the price of high quality offered to the type 2 buyer will be quite attractive in the second period (due to competi does tion), the type-1 buyer is less inclined to choose high quality at a rela tively high price in the first period. If the incumbent still wants to sepa rate the buyer types, therefore, the incumbent will have to give a larger makes to the type-1 buyer, and a large price concession price concession more competi discrimination less attractive. When the market becomes tive (when 8 rises), discrimination becomes even less attractive from the incumbent's perspective. When the market gets too competitive (when 8 is too close to 1), the would want to offer high quality to incumbent however, although both types of buyer, the buyer, particularly the type-2 longer want to choose high quality. As 8 gets larger, high-quality, pooling offer gets closer to the expected both buyer types ($85 when 9 = 0.5), and the type-2 that she is indirectly subsidizing the type-1 buyer, will buyer, would no the incumbent's cost of serving buyer, knowing have a stronger to signal her type to the market. With too much competition (8 breaks apart and the close to 1), the high-quality, pooling equilibrium market will revert back to an equilibrium that is similar to the one in incentive This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions The Effectof Bargaining Power 2012] 1719 which the buyer was able to make a take-it-or-leave-it offer to the seller: the type-1 buyer acquires high quality at a price close to the cost of serv ing that type while the type-2 buyer receives low quality at a price cor to its cost.95 For the pooling equilibrium a moderate level of therefore, maintaining competition range) is important. responding to be sustained, in the middle (5 B. Renegotiation In the previous uted by allowing take-it-or-leave-it more evenly distrib who are making competition among offers to the buyer. Another important source of bar example, for some bargaining power was sellers gaining power is the power not to renegotiate or modify the terms. When one party is endowed with the power to make a take-it-or-leave-it offer, the offeror has an incentive to deliberately introduce inefficiency in the a larger share of the surplus. The presence of such in however, implies that, if the bargaining parties have a chance efficiency, to renegotiate the terms, they would be willing to make a welfare The ability to extract the maximal share of sur improving modification. hopes of capturing plus through the introduction of inefficiency is sensitive to the assump tion that the party with bargaining power could commit not to renegoti ate the terms that were previously agreed upon. An important aspect of renegotiation is that such possibility is particu larly salient and relevant to contract terms, such as warranty, termina rather than other aspects of the transac tion, choice-of-forum clauses, tion, such as the product's physical attributes. It may be fairly easy for the parties to renegotiate over such contract terms either before or even after the product has been sold. On the other hand, physical attributes of a product tend to be immutable: once the seller has decided on the phys ical attributes, for instance before introducing the product to the market, 95 What we mean by the pooling equilibrium no longer being sustained is that the pooling in which both types of buyer acquire equilibrium, high quality at a single price, no longer satisfies the Cho-Rreps "intuitive criterion." See Cho & Kreps, supra note 94. With a large 8, only the type-2 buyer will have an incentive to reject the incumbent's offer, and the mar ket (the incumbent and the entrant) correctly the incumbent's offer. There is some welfare believes loss, that it is the type-2 buyer that rejects the type-2 buyer will have to wait since until the second the transaction. As S approaches one, this equilibrium period to consummate to the Nash equilibrium in which the buyer was making a take-it-or-leave-it offer converges to the seller. This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1720 or once change When Virginia the product has been the attributes. Law Review [Vol. 98:1665 sold to the buyer, it is often impossible to the contracting parties cannot commit not to renegotiate the to the one-period terms, this will introduce two important modifications take-it-or-leave-it offer models. First, since the parties will voluntarily renegotiate the terms when the nonprice will mitigate the inefficiency. terms are inefficient, such rene In equilibrium, the parties are terms (either at initial formation gotiation more likely to adopt efficient nonprice or through renegotiation). Second, even when only one party has all the power to make offers, both in the initial formation and the renegotiation stages, the lack of commitment implies that a bigger share of the surplus will have to be shared with the counter party because it makes the initial discrimination more difficult. To understand these points more clearly, let us go back to the exam the where seller had the power to make a one-time take-it-or-leave-it ple offer to the buyer. When the seller could commit not to renegotiate the terms, the seller was able to separate the buyer types and extract the maximal surplus from the buyer. Once the buyer types self-select, on the other hand, the buyer type is revealed to the seller: the seller knows for certain which type has accepted which offer. In the one-shot game, even after knowing the buyer type, the full commitment implies that the par ties will go ahead and execute the inefficient, low-quality contract with respect to the type-2 buyer. If the seller cannot commit not to renegotiate the terms, on the other when the hand, type-2 buyer accepts the offer with low quality, the sell er will attempt to renegotiate the terms so as to capture the residual sur plus. Although this will be better for both the seller and the type-2 buy the er, when the type-1 buyer expects that the seller will renegotiate the no have an incentive contract, low-quality type-1 buyer may longer to stay with the high-quality contract. The type-1 buyer now may want to mimic the type-2 buyer by choosing the low-quality contract in the first stage, hoping that she will be able to get the high-quality contract, at a lower price. To achieve separation, the seller through renegotiation, will have to leave a larger surplus to the type-1 buyer. Contractual sur will be more shared with the even the seller is plus evenly buyer though the only one making offers. To present these ideas more formally, suppose, as in the previous var iation, we have a two-period bargaining game with delay but with only the seller making offers in both periods. Like before, at t = 0, nature se This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions The Effectof Bargaining Power 2012] 1721 lects the buyer type, and in the first period (t = 1), the seller makes an offer to the buyer. If the buyer rejects, the game moves to the second pe riod (t = 2) with delay (discount factor 5), where the seller gets a se cond chance to make the offer in the second an offer to the buyer. When period, since there already the seller is to make is an agreement, the sure that both parties will get more from the re negotiated contract than what they are entitled to receive under the ini tial contract for renegotiation to be successful. The following diagram seller will have to make represents Figure the possible 2: Possible scenarios Scenarios of the game. of the Renegotiation Game Nature Type 1 Offers {&)yS ((Pv^UPi'lz))/ \ / \ Buyer/ / Accept Pi j l A ~ci(9i) J \ 2 (1 - 0) ~~^~~-\SeUer Seller^-""" /\ ^s\Type Reject Seller Makes Second Offer with delay S offers 7\ V(A>?.X(P2.?2>) (/ Accept \ Buyer / Seller Offers to Renegotiate with delay S \ Reject Seller Makes Second Offer with delay S To construct an equilibrium of this game, first hypothesize that, in the first period, the seller makes a menu of offers, ((px, h), (p2, /)), to the in accordance with her buyer and the buyer self-selects (and accepts) This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1722 Virginia Law Review [Vol. 98:1665 type.96 In the second period, with respect to the type-1 buyer, since high quality has been agreed upon, the seller knows that there is no gain from With respect to the type-2 buyer, on the other hand, since renegotiation. the buyer has agreed to purchase low quality (i.e., has accepted (p2, /)) in the first period, the seller knows that there is a surplus to be captured from renegotiation. Under the initial agreement, the type-2 buyer ex — pects to realize a surplus of $170 p2 and the seller expects to realize a — of Since $50. profit p2 selling high-quality product to the type-2 buy er generates a larger surplus, the seller will offer to renegotiate tract by making a renegotiation offer of (p2, q2 = h) where — p2 > $170 p2, that is, the buyer's surplus from renegotiation the con $200 — must be at least as large as that from the initial contract. Having the power to dic tate the terms of renegotiation, the seller will offer p2 = $30 + p2 to the type-2 buyer in the second period. Moving back to the first period, when the seller offers a menu of con the type-1 buyer knows that if she were to tracts, ((p1( h), (p2,0), choose the contract with low quality, (p2, I), with the discount factor of offer of S, in the second period, the seller would make a renegotiation < where So as the of the $200. (p2, K), p2 long price high-quality prod uct is higher than $200, that is, p1 > $200, the second con choosing tract and waiting for the renegotiation becomes attractive. To prevent the type-1 buyer from doing so (and pooling with the type-2 buyer), the seller has to make a bigger concession on p1. That is, the seller has to — — ensure that $250 — px > (1 — 5) x ($190 p2). p2) + <5 x ($250 When 6 = 0.8, the profit-maximizing set of prices for the seller is = $206, = $170, and = $200.97 The p1 p2 p2 following table summa rizes the results. 96 9 > 1/4, offering a menu of contracts to the buyer is It is easy to show that whenever more profitable for the seller than making a pooling offer. 97 in the industrial organ This result is similar to what is known as the "Coasean dynamic" izations literature. Professor Ronald Coase that if a durable-goods conjectures monopolist cannot commit not to lower its price, the monopoly rent and the deadweight loss will disap and Monopoly, 15 J.L. & Econ. This is 143, 143^44 Coase, (1972). pear. Ronald Durability because after selling to only a subset of consumers at a mo (with high reservation values) will attempt to satisfy the residual demand (the consumers who nopoly price, the monopolist value the good more than the cost of production but less than the initial monopoly price) by consumers lowering its price. If the initial high-reservation-value expect this, they will simp ly wait for the lower price. Our story is similar but different since we are more concerned with renegotiation of an existing contract, rather than forming new contracts with other sets of consumers. and decreases in our game, the price of the high-quality Nevertheless, product starts at $206 to $200 in the second period. See also Faruk Gul et al., Foundations of Dy This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions The Effectof Bargaining Power Table 6: Equilibrium ate (S = 0.8) When Seller Commit Not to Renegoti ((Pi, h), (p2,/)) = (($206, h), ($170,1) 9 = 1/2 (p2,q?) = ($200,h) Type-1 Buyer's Surplus $44 Type-2 Seller's Buyer's Profit Surplus $0 Total Cannot (Expected) Surplus $117 $139 to the game where the seller was able to make a one-time Compared take-it-or-leave-it offer, by taking away the power of commitment not to both the total surplus and the type-1 buyer's surplus have renegotiate, the total surplus from $135 to $139 and the type-1 buyer's surplus from $20 to $44. We can think of S as being inversely related to the power of commitment. With a larger 8 (a weaker commitment pow increased: er), we can expect a bigger surplus for the buyer and larger efficiency while, with a smaller S (a stronger commitment power), the equilibrium will produce a smaller surplus for the buyer and lower efficiency. In deed, when 5 = 1, with the seller having no commitment power, the profit-maximizing strategy for the seller is to offer ($200, h) in the first and serve both types of buyer. With no commitment power, inef when 5 = 0, we come back to the ficiency disappears. Conversely, in which the seller is able to make a one-time take-it-or-leave-it game period offer to the buyer, with maximal inefficiency.98 C. Bilateral Negotiation way of reducing the bargaining power gap is by giving both a chance to dictate the terms of the contract (in sequence). parties Sup pose, similar to the competition/entry game, the bargaining game con Another sists of two periods. But rather than having a competitor enter the mar ket in the second period or the seller renegotiate, we let the suppose and the Coase 39 J. Econ. Theory 155, 169 (1986) Monopoly Conjecture, (showing that monopolist rent and deadweight loss disappear as the time interval between offers goes to zero). 98 S (or reducing Note that in this renegotiation the seller's game, increasing bargaining increases This is partly due to the fact that the seller still retains power) always efficiency. namic the power to make a take-it-or-leave-it offer in the renegotiation stage. If the buyer were S too much would create a signaling offer, increasing making the renegotiation inefficiency. The model, then, will be qualitatively similar to the first and the third variations. This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1724 Virginia Law [Vol. 98:1665 Review buyer make a counteroffer to the seller. In the first period (t = 1), the seller makes an offer to the buyer, which the buyer can either accept or the game ends on the seller's reject. If the buyer accepts, proposed terms. If the buyer rejects, on the other hand, the game moves to the se cond period. In the second period (t = 2), the roles are reversed and the buyer gets to make an offer to the seller. If the seller accepts, the game ends on the buyer's proposed if the seller rejects, the terms, whereas ends with no trade and both Unlike the game parties getting nothing." the and competition/entry renegotiation games, however, game moves to the second period firstperiod. only when an agreement has not been reached in the As in the previous games, we assume that delay is costly for both par ties. Since the buyer has the last chance of dictating the terms of the transaction, if the buyer were to make the second period offer immedi ately after rejecting the seller's offer, the game would collapse to the one that allows the buyer to make a take-it-or-leave-it offer to the seller. Similarly, if the buyer never gets to make an offer in the second period, the game will be identical to the one in which the seller was able to make a take-it-or-leave-it offer to the buyer. To keep the respective par in check ty's bargaining power (and allow both parties a chance to dic tate the terms of the transaction), we again multiply the payoffs from the second period by a discount factor of 8 G [0,1]. As before, the higher the Ö, the less costly the delay in reaching an agreement (or the more pa tient the parties become) and the more bargaining power the buyer has. The discount factor plays the dual role of providing the parties an incen 99 We can reverse the roles and let the buyer make the initial offer and the seller the subse results will not change. There are (at least) two quent offer (with delay), but the substantive other ways of representing more "even" One is through a "flip-a bargaining assumption. in which the outcome coin" mechanism, of a coin flip will get to make a take-it-or-leave-it offer to the other party. In that game, however, no matter who comes to be the offeror, the offeror will act as if she had all the bargaining will contain power. Hence, the equilibrium the inefficiencies that are identified in the previous models. The other is by allowing both parties to make simultaneous that setting, when the buyer's a double auction mechanism. In offers, that is, by imposing bid price is larger than the seller's ask price, trade is executed whereas if the buyer's bid price is lower than the seller's ask price, price in between takes place. See Kalyan & William Under In Samuelson, Chatterjee Bargaining 31 Operations Res. 835, 838 (1983). full bargaining solu Information, complete Although tions have not been worked out yet, in that model it is likely that the set of inefficient equi at some no trade libria cannot be ruled out. This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1725 The Effectof Bargaining Power 2012] an agreement early and also distributing the bargaining to dictate the terms) between the parties.100 power (the power The following ways that the game could figure represents possible = 0), nature determines the buyer type. In the first play out. Initially (t = the seller makes an offer. Given that the seller does not 1), period (t tive to reach know the buyer's type, the seller can either make a pooling offer, (p, q), If the seller were to choose the lat or a menu of offers, ((pv h), (p2,1)). ter, the buyer can either reject the entire menu or choose one of the of If the buyer were to reject the offer, the game moves to period. In the second period, given that this is the last chance fers in the menu. the second an agreement and that the buyer has all the bar the offers that are iden buyer will make type-dependent gaining power, tical to the ones the buyer used when the buyer could make a take-it-or for the parties to reach leave-it offer: (($100, ft),($50, Z))-10' In the figure, as before, the se cond period actions are folded into the payoffs for simplicity. The pay offs are discounted by <5 in case the buyer rejects the seller's offer in the first period. 100 Another a chance way to think about to make a counteroffer. S is that it represents the probability that the seller is given If there is a chance that the consumer will simply walk out of the store when the seller rejects the consumer's offer in the first period, the seller's ability to make a counteroffer should be reduced by that probability. 101 This results from refinement of the off-the-equilibrium beliefs. When 8 > 5/6 « 0.83, we can show that this will always be true. When the buyer rejects, as an off-the-equilibrium the offer of (p, h) or ((p1( h), (p2l /)), the seller correctly assigns probability 9 to deviation, the possibility that the offer is coming from a type-1 consumer. And, therefore, knowing this, in the second in signaling period, the type-2 buyer will have to engage by offering low belief (by the seller) will satisfy the Cho-Kreps "intui quality. This out-of-the-equilibrium See Cho & Kreps, supra note 94. When 8 < 5/6, however, this will no long er be true. With respect to the pooling offer, (p, h), the seller will assign the probability of 1 that the rejection is coming from the type-2 buyer. Similarly, when the buyer rejects the tive criterion." menu is done of offers, ((p1; h), (p2,/))> the seller by a type-2 buyer when 5 < 2/3. will assign the probability of 1 that the rejection This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1726 Virginia 3: Possible Figure Scenarios Law Review of the Alternate [Vol. 98:1665 Offer Game Nature Type2 (1-6) Type 1 (0). Seller Seller Offers Offers Offers \((p».h).{Pi.O) \«Px.h),(Pz, 0) (p,h)/ Type-2 Type-1 \ Buyer Buyer VReject (250-p\ lp-ioo/ \Reject ^Reject Accept, Accept/ fSlS0\(2S°-PA v o J Ipi - loo; Accept/ \Rqect Accept/ fSl50\ f200-p\fsi20s.n70-p2\ k o Mp-70/1 0 Jvp2~5o; (S1Z0\ I o J How will the equilibrium change from the one-period game? Assume and start that S — 0.9. To construct an equilibrium, we move backwards from the second period. As we have noted, in the second period, assum ing that the buyer has rejected the seller's offer in the first period, the buyer will behave as if she has the power to make a take-it-or-leave-it offer to the seller in a one-shot game: the type-1 buyer will offer ($100, h) while the type-2 buyer will offer ($50, Z). The seller will ac cept both offers, rendering $120 to the type-2 buyer. a surplus of $150 to the type-1 buyer and Moving back to the first period, whether or not the seller would want to make a pooling offer, (p, q) — (p, h), or give the buyer a menu of op she will have to ensure that the type-1 buyer will tions, ((px, h), (p2,0), while the type-2 buyer will at a surplus of S x ($150) a surplus of ö x ($120). Furthermore, if she were to offer a she has to further make sure that the type-1 buyer is better off at least realize least realize menu, choosing (p1( h) selecting (p2,0 102 rather than (p2,1), and the type-2 buyer is better off rather than (p1( /i).102 Under the first option, the best — — the pooling 5150,200 offer, the seller must satisfy p < min{250 5120}. — < 250 — 5150; the separating offer, the seller has to satisfy (1) (2) p2 — 170 and (3) pi — p2 < 60. Intuitively, 5120; having to satisfy a larger number of constraints, ceteris paribus, usually implies that the seller's profit will be lower. With With This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1727 The Effectof Bargaining Power 2012] = 0.9, is possible price the seller can offer for the high quality, when 5 $92. Under the second option, the optimal set of offers for the seller is (($113, h), ($53,0). In both cases, the buyer will accept. The first, pooling offer, option will allow the seller to capture a surplus of $7, while the second, a menu offer, option will allow the seller to realize a the seller will choose to make a profit of $3.5. Clearly, in equilibrium, pooling offerof ($92, h) and the game will end in the firstperiod with is better off since, had the game pro period, the seller would have made no profit. The better off since the buyer does not need to wait for the se out any inefficiency. ceeded to the second The seller buyer is also cond period, that is, there is no delay in contract formation. Table 7: Equilibrium When e = 1/2 Type-l Buyer's Surplus Type-2 Seller's Buyer's Profit Surplus Total (Expected) Parties Alternate in Offers (p,q) = ($92, h) $158 $108 (S — 0.9) $7 Surplus $140 An important factor that determines the characteristics of the equilib = factor Ö. If Ö for the rium is the discount 0, instance, buyer's ability to and the two make an offer in the second period is of no consequence to letting the seller make a equivalent game becomes = if S there is no cost in delay in offer. Similarly, 1, an agreement (or the buyer gets to make a counteroffer for cer period bargaining take-it-or-leave-it reaching tain), and the buyer will fully exercise her right to make the last offer: identical to the monopsony the bargaining bargaining game becomes scenario.103 As 5 gets smaller, the buyer's last-shot power becomes more and the seller will be able to get a larger share of the surplus. So long as <5 does not get too large or too small, the above pooling equi librium can be sustained and the first best can be achieved. diminished One intuitive way of thinking about this bargaining game is by recog that the nizing how much pricing restriction is imposed by 8. Suppose parties are still playing the two-stage bargaining game but with complete and symmetric information over the buyer's type. In that scenario, as we have seen earlier, the parties will choose the optimal nonprice terms to 103 two games (monopolist and perfect competition, The previous can be thought of as special cases of this more general game. or monopsonist, This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions games) 1728 Virginia Law [Vol. 98:1665 Review that the contractual surplus. At the same time, the assumption the buyer can make a counteroffer only with some delay introduces an important check on the buyer's bargaining power. To see this, with symmetric information, the buyer, in the second maximize product with prices of either stage, will offer to purchase high-quality This allows the buyer to realize a or on her $100 $70, depending type. respective profit of $150 and $130 in the second period. The seller, ex pecting this outcome from the second period, will offer, in the first peri od, respective prices that make the buyer just indifferent. To the type-1 — $100 > buyer, the seller will offer p1 with high quality, such that pl while to the type-2 buyer, the seller will offer p2 such that 8 x $150, — $70 >5x p2 $130. If we let p1 = $100 + 8 x $150 and p2 = her profit, we get + 8 x $130, which allow the seller to maximize — $30 + 8 x $20. Note that as 8 gets smaller, so does the dif P2 Pi ference between pt and p2. When 8 — 1, the buyer can fully extract all case. However, when 8 < 1, surplus from the seller as in the monopsony $70 ~ the buyer no longer buyer's has unlimited decreases, pricing power power in setting prices. And as the so does her incentive to impose ineffi ciency. VI. Bargaining The motivation Power, Discrimination, of this Article is to address and Legal Policy what we identified at the theory of bargaining power's effect on contract how does bargaining is Our descriptive: purpose predominantly design. the influence terms? way, we have commented Along nonprice power and under different the of the outcomes on power allocations efficiency outset as the irrelevance this naturally leads to the question bargaining or market inefficiencies. in correcting that legal institutions are so to mitigate these problems of the role of the law We believe unlikely to have the information needed that the cure may be worse than the disease.104 Nevertheless, before con of our analy as to the implications cluding, we offer some observations sis on the policing of bargains under contract law. a court may re Under the common law doctrine of unconscionability, contract term or the entire contract by fuse to enforce an unconscionable either modifying or voiding the contract. The doctrine requires not only but a defect in the bargaining process ("procedural unconscionability"), to the vulnerable unfavorable also a term that is harsh or unreasonably 104 See Craswell, supra note 4, at 33. This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions The Effectof Bargaining Power 2012] 1729 While gross inequality of unconscionability").105 is often mentioned as a factor to bargaining power contributing proce dural unconscionability, it is rarely sufficient on its own.106 Unless the party ("substantive imbalance amounts to duress, undue influence, or incapacity, courts typ further defects in bargaining, especially a finding that the ically require weaker party also lacked the opportunity to read or understand the harsh term. Courts do not interfere with commercial contracts based solely on a procedural concern with unequal bargaining power.107 When parties are rational and fully informed about the terms of their concerns are less severe because at least each agreement, distributional 105 The terms "substantive coined originally Code—The by Professor unconscionability" Arthur Leff. and were "procedural unconscionability" A. Leff, Unconscionability and the L. Rev. 485, 487 (1967). The formulation of Arthur New Clause, 115 U. Pa. Emperor's Court of Appeals for the District of Columbia Circuit in Williams v. Walker-Thomas Furniture Cir. 1965), is often cited: "Unconscionability has Co., 350 F.2d 445, 449 (D.C. to include an absence of meaningful choice on the part of one of generally been recognized the parties together with contract terms which are unreasonably favorable to the other party." the U.S. In addition to noting the imbalance in bargaining the fact power, the court also emphasized that the terms were written on the back of the order form in fine print and in language that was difficult to understand. Id. at 448-49. 106 Both the Restatement of Contracts and Article 2 of the Uniform Commercial (Second) Code adopted similar interpretations of the doctrine. According to the Restatement, "[a] bar the parties to it are unequal in bargaining gain is not unconscionable merely because posi tion .... But gross inequality of bargaining favor power, together with terms unreasonably able to the stronger party" may support a finding of unconscionability in the bargaining Restatement of Contracts the high-water process. (Second) § 208 cmt. d (1979). Perhaps mark of the concern over the imbalance of bargaining power on its own was the famous case of Henningsen v. BloomfieldMotors, Inc., 161 A.2d 69 (N.J. 1960). The court struck down a that was not specifically but warranty disclaimer brought to the attention of the consumer, the main thrust of the court's focused on the concentration of bargaining opinion power in Id. at 92, 94. Since then, the occasional court has based a finding of on bargaining procedural See, e.g., Gianni unconscionability power. Sport Ltd. v. Gantos, Ct. App. 1986) (upholding the lower court's determi Inc., 391 N.W.2d 760, 762-63 (Mich. nation that a cancellation clause benefitting a large retailer against a small independent man the automobile industry. ufacturer was unconscionable); Shell Oil Co. v. Marinello, 307 A.2d 598, 601 (N.J. 1973). most do not. Comment 1 to § 2-302 of the Uniform Commercial Code However, suggests that bargaining of op power is not by itself enough: "[t]he principle is one of the prevention of allocation of risks because of supe pression and unfair surprise . . . and not of disturbance rior bargaining U.C.C. power." § 2-302 cmt. 1 (2011). 107 U.C.C. v. Caterpillar, 1995 § 4.9 (5th ed. 1996); see, e.g., Coursey Inc., No. 94-1348, WL 492923, at *3 (6th Cir. Aug. 6, 1995) ("Unconscionability is rarely found to exist in a commercial & Eng'g Inc., v. Lewis setting."); Cnty. Asphalt, Welding Corp., 323 F. Supp. 1300, 1308 commercial (S.D.N.Y. 1970) ("[I]t is the exceptional setting where a claim of will be allowed But see Campbell ...."). Soup Co. v. Wentz, 172 F.2d 80, where the bargain was one-sided and op (refusing specific performance unconscionability 83 (3d Cir. 1948) pressive). This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1730 Virginia Law [Vol. 98:1665 Review party is better off than without an agreement. While one could imagine a policy striving to achieve more even sharing of the surplus, another con cern is with cases in which the exercise of bargaining power undermines that inefficiently The analysis in Part IV demonstrates terms can persist even between sophisticated parties when the in or the in seller engages buyer engages screening signaling, particular This result underscores that the ly when bargaining power is unequal. value-creation. one-sided current judicial and scholarly skepticism as to the earlier adhesion and the lack of meaningful choice is exaggerated. concern over it re Indeed, that a menu of terms may itself be evidence of a problem. When a seller screens, it may impose harsh terms on one set of buy monopolist veals ers (the low-value warranties to buyers) by, for example, disclaiming to the other group (high-value them while offering broad warranties If courts were to be more aggressive in policing buyers). bargaining be particularly vigilant when faced with a discrimi Sometimes the monopolist's nating monopolist. screening is obvious, in which case the court should scrutinize the nonprice terms of the lower power, they should quality contract: purchase for example, a warranty disclaimer with an option to an extended warranty. as Part IV also demonstrates, the absence of bargaining Conversely, power on the part of the seller does not resolve the danger. Inefficient Of course, in cases in might occur in perfect competition. signaling the low which the buyer has all the bargaining power (monopsonist), value to her buyer may propose inefficient terms that are unfavorable self, to avoid being pooled with the high-value buyer.108 If that were the case, since it is the buyer's exercise of bargaining power that is causing more seller-friendly terms, it will be difficult for the buyer to claim the After all, the sympathy of the court in a review for unconscionability. gains from the inefficient signaling accrue to the buyers rather than the competitor sellers. Therefore, mandatory terms might be a superior solu tion. On the other hand, it is difficult to know what a court should do with If the law compelled the monopolist to offer only a sin such a contract. 108 tion This about result, derived from the numerical the buyer's relative reservation not had a higher marginal buyer only in Part IV, is sensitive to the assump example we have assumed that the values. Throughout, will to pay but also higher absolute willingness reservation val ingness to pay for quality. If the type-2 buyer were to have a higher absolute ue vis-á-vis the type-1 buyer, when the buyer has all the bargaining power, the buyer will demand quality that is inefficiently high. See supra note 82. type-1 This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1731 The Effectof Bargaining Power 2012] gle contract (rather than a menu of contracts or options) in order to pre vent inefficient screening, we observed earlier (and show in the more in model the that this Appendix) may not improve welfare in the general of heterogeneous preferences requiring different nonprice terms. establish a mandatory minimum quality and the court Alternatively, may refuse to enforce anything less—for example, by prohibiting warranty disclaimers or requiring a period of notice before termination. Of course, face as others have argued, any attempt to mandate quality levels for specific classes of buyers must also deal with the informational limitations of the lawmaker, whether regulator or court.109 Even assuming that a court can set a floor at the optimal level for the buyer that values quality the least, however, the discriminating monopo list or a perfectly competitive market may still offer lower-than-optimal quality to intermediate classes with preferences between the lowest and highest or, perhaps even worse, may decide not to serve consumers who quality the least. As a variation to the numerical example, if there are three types of buyer and three different levels of optimal quality, set ting the minimum quality standard may improve the welfare for the con value sumer who values the quality the least, but it may not do anything for who also purchases inefficiently low-quality prod uct. If the quality minimum gets too high, the monopolist or the compet itive market may decide not to serve the lowest-type consumer, thereby the middle consumer generating an even greater inefficiency.110 price regulation, an instrument that courts have been reluctant Perhaps to invoke,111 might be more effective: courts might impose limits, both 109 The doctrine of unconscionability has been criticized on this score. See, e.g., Craswell, Another relevant critique is that the supra note 4, at 1066-67. supra note 4, at 33; Schwartz, will lead stronger parties to extract court's refusal to enforce one-sided nonprice provisions their rents by raising the price, leaving weaker parties worse off than if the original contract had been enforced. not feasible extract that this is e.g., Kennedy, supra note 19, at 619. We note, however, where the monopolist is already maximally utilizing price terms to Since a price term is more efficient for rent, as in our numerical example. See, in situations consumer extracting rent, our result seems not unreasonable. 110 See Besanko et al., supra note 73, at 750-51, for a demonstration who desire a moderate level quality threshold (1) will leave consumers and (2) may induce the monopolist els of quality. 1'1 Courts seem to be reluctant to altogether stop serving to strike down the boundaries difficulty of determining sor Farnsworth notes two other possible of "fair" reasons: likely to be known and understood by the weaker struck down as being unconscionable (i.e., being consumers of how a minimum of quality unaffected who desire low lev of the terms, at least partly because markets. Profes prices in noncompetitive (a) the price, being much more salient, is party and (b) even if the contract price is price outside the bounds of "fair" This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions price), the 1732 Virginia Law Review [Vol. 98:1665 upper and lower, on price for the product. At least in theory, this can work because once the monopolist's incentive to extract consumer sur plus (or buyers' incentive to engage in signaling) is kept in check, its de sire to quality discriminate will disappear or at least be substantially mit is prevented from igated. From the numerical example, if the monopolist above for she can no longer profit $220 charging high-quality product, from inefficient and will, instead, offer both types price discrimination of buyer high quality at $200. Also, if the minimum price of $70 can be maintained when the buyer had all the bargaining power, the type-2 buyer can no longer benefit by deliberately choosing a low-quality prod uct. Price regulations, lenge of institutional of course, immediately run into the familiar chal competence—particularly, identifying the proper limits in any given market. These considerations suggest that a general structural policy to pro mote competition might be preferable to contract regulation. Our analy sis, however, demonstrates that shifting bargaining power to buyers may overshoot the objective if it gives buyers too much power. The policy of in this sense may backfire. As was demonstrat empowerment ed through the numerical example, the buyers themselves may choose to consumer create a signaling equilibrium that leaves some of them with inefficient when are aware of the possibility of indirect subsidy to oth quality they er buyers. This is true even in markets that appear to courts as competi tive where power the valuation of customers is private. The interior range of is golden in this context: a more even distribution of power and sharing of rents can mitigate the inefficiency. The to achieve this balance is an interesting question worthy of allocation bargaining mechanism further investigation. Conclusion For over forty years, law-and-economics has been work scholarship ing on the premise that bargaining power is irrelevant to the design of nonprice contract terms. Practitioners have the opposite understanding in contrast, give much weight to the balance of power, whether market concentration, imbalances, stemming from supply-and-demand or negotiating skill. This Article takes the first steps to bridge theory and and, practice by identifying the conditions court will be obliged in many cases to substitute Contracts § 4.28, at 306-07 (4th ed. 2004). under a fair price which in its place. bargaining E. Allan This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions power Farnsworth, 2012] The Effectof Bargaining Power 1733 might affect contract design. We analyze at some length one set of ex the effect of relative bargaining power where one party has planations: information about its reservation price. We also identify a variety private of other possible and leave to future research the further explanations unpacking of the bargaining power irrelevance proposition. This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1734 Virginia Law A General Appendix: [Vol. 98:1665 Review Model capturing the notion of bargaining power in the presence of asymmetric information is difficult. In this setup, we take a more re duced-form approach by letting A denote the fraction of the equilibrium Accurately the social planner surplus the seller captures in expectation. Suppose function a social-welfare (the mechanism designer) wants to maximize (the objective function), which is constructed on the weighted average of con the buyer's and the seller's expected profit, subject to various straints. represent the fact that although the social planner can the terms of the trade, she is still constrained by (1) the lack of information and (2) the inability to "force" both the seller and the buyer The constraints dictate to participate. That is, although the buyer knows his type, the social does not observe buyer type and the social planner has to guar planner antee both the buyer and the seller at least zero profit, their respective outside reservation value. The social planner's problem can be written as follows:112 ~ ~ ciO?i)) + C1 0)(P2 MaxÍP,q) A (9(Pi CzOfe))) + (1 0)(v2(q2) pa) + (1 A)(0(v1(q1) subject - p2)) to Pi - Ci(£7i) P2 -c2(<72) > 0 ^ 0 ^i(<7i) ~Pi ^ 0 v2(q2)-V2 ^ 0 ^ vi(<fc)-P2 ViM-Pi Pi v2(<?2) P2 ^ v2(<h) 112 the social planner will induce the buyer to that, in equilibrium, program assumes a different contract based on his type: Pi =*=p2 and q1 =/=q2. If this is not feasible, that is, Pi = p2 or qt = q2, then we can rewrite the program by defining p = min(p1(p2) the objective function subject and q = max(q1, q2) and letting the social planner maximize The choose by (p, q). In theory, this formula only to the first four constraints, where (p¡, q{) is replaced with 0(p — c1(q)) + (1 — tion presents a possibility of replacing the first two constraints — > will but we the easier, 0, ignore 0)(p c2(q)) thereby making implementation problem this possible relaxation. This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions The Effectof Bargaining Power 2012] We will impose al utility(v[(q) the usual assumptions 1735 of strictly diminishing margin > 0 but v['{q) < 0) and (weakly) increasing marginal c" (q) > 0). Furthermore, to make the problem more interesting, we will assume that the type-1 buyer's marginal utility is always higher than the type-2's marginal utility (v[(q) > v2(q)), that the is and that the cost of is, satisfied, serving single crossing property cost (c[(q) > 0 and the type-1 buyer is strictly higher than the cost of serving the type-2 > c2(q) Vc/. To ensure interior solutions, we'll also as buyer: c^q) sume that c¿(0) = 0, Vi > 0, and v¡ » 0. function is a weighted average of the The objective (or social-welfare) surplus expected profit (weighted by A) and the buyer's — 1 As A the will more em A). gets larger, equilibrium put (weighted by A is smaller, the buyer's surplus phasis on the seller's profit, and when function can be rewritten as becomes more important. The objective seller's 0((1 - A)y1(g1) - Acx(qj + (2A - l)px) + (1 - 0)((1 - A)v2(q2) - + (2A — 1 )p2). Three special cases are worth separate consid = When A to maximizing the 1/2, the problem is equivalent conventional social surplus (or consumer surplus plus producer surplus, Note that when A = 1/2, the price terms disappear equally weighted). from the objective function. When A — 1, the problem is identical to that Ac2{q2) eration. of a monopolist who, with the power to make a take-it-or-leave-it offer, tries to extract as much surplus as possible from the buyer and maximize her profit subject to satisfying the buyer constraints. When 1 = 0, the offer to buyer, for instance, by being able to make a take-it-or-leave-it the seller, is maximizing her surplus subject to making sure that the sell er at least breaks even.113 When A e (0,1), the bargaining power is dis tributed between the two parties. 113 We can also think of this model as representing a solution to Nash bargaining in which A represents the seller's relative bargaining the parties bar power. Under that interpretation, solution must respect each gain ex ante, without knowing the buyer type, but the bargaining ex post participation the buyer's incentive con and, in particular, party's compatibility straints. attempt That is, even after they have worked out a solution, one or both of the parties can to renegotiate when the solution does not give the party more than what the outside or the other option dictates. Although it may be tempting to suggest that the equilibri option um with A e (0,1) can be replicated to by "flipping a coin" that gives more (or less) chances make a take-it-or-leave-it offer to the seller (or the consumer), this will not be true. Such a model will produce an equilibrium that will be a convex combination of two polar equilibria, and the inefficiencies at each pole may be simply averaged out without getting any closer to the optimum quality provision. The equilibrium under our setup will be different produced from such combination. This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1736 Virginia Law [Vol. 98:1665 Review Turning to constraints, the first four constraints represent the seller's and the buyer's participation constraints, making sure that they, at least, realize zero surplus in equilibrium. As we will see shortly, in equilibri um, each respective buyer type will receive different quality level prod uct. The seller, therefore, must be able to break even for respective buy er type and each type of buyer should be able to realize more than her outside to be zero. The last two constraints option, which is assumed the incentive conditions: each type must represent buyer's compatibility (at least weakly) prefer to choose the contract that is intended for the type. Since the social planner does not observe buyer type, the social planner must offer a menu of contracts for the buyer to self-select.114 With these able in the first best, if the social planner were buyer type, regardless of the weight (A) she assigns to the respective marginal utility would profit, in equilibrium, assumptions, to observe the seller's be equated with the marginal cost: v[(jqf) = c¡(q¡). Even if X = 1, for planner will still equate the marginal utility to the = but let and p2 = v2(ql). marginal px vx(.q{) Similarly, when = A 0, the social planner will choose the optimal qualities for both types and set p1 = c1(q¡) and p2 = c2(q2). For simplification, we will assume that the marginal cost of producing for the type-1 consumer isn't too instance, the social cost much larger than the marginal cost of producing for type-2, so that the first best requires the type-1 consumer to purchase a higher quality and at first the that, best, product, q{> q2, type-1 buyer's utility is > than the higher type-2 buyer's utility, v1 (qf) v2(q2). Figure 4 pre sents an example in which q{ > q2. In addition, to make the problem in — that v^q?,) > stated, we assume teresting, unless otherwise c2(ql) — The that when the v-tiql) c^ql). assumption implies respective quali ties are offered at marginal cost, the type-1 consumer will prefer to pur 114 Note that, in our setup, there always is a positive surplus from trade: the probability that the buyer's valuation is larger than the seller's cost is equal to 1. If it is uncertain whether a values and costs, surplus exists and if the parties are privately informed of their respective the mechanism becomes more complicated. In particular, if the values problem we may run into the (strong) inefficiency result of Professors Satterthwaite: there will be no mechanism that realizes all positive surplus. costs design are uncorrelated, erson and and My See B. Myerson & Mark Satterthwaite, Efficient Mechanisms for Bilateral Trading, 29 J. When the values and costs are correlated, as in our example, Theory 265, 273 (1983). all possible easier since the mechanism can use one realizing gains will become designer information. See Jacques Crémer & Richard P. party's report to learn about the other's Roger Econ. Under Uncertainty for a Discriminating McLean, Optimal Strategies Selling When Demands Are Interdependent, 53 Econometrica 345, 346 (1985). This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions Monopolist The Effect of Bargaining 2012] chase the product intended for the type-2 tion will result with marginal cost pricing. Figure 4: An Illustrative Power buyer. 1737 That is, adverse selec Example A. Case 1: A > 1/2 When A > 1/2, the social planner cares more about the seller's profit — than the buyer's surplus. From the objective function, 6>((1 — A)v1(q1) ~ tei(<h) + (2Ä !)Pi) 1 )p2), + C1 - 0)((1 - ^)v2(<72) - 2{q2) + (2A holding everything else constant, when A > 1/2, higher prices strictly increase the value. As an extreme case, when A > 1/2, the social welfare functionbecomes 6{p1 — the problem + (1 — 0)(p2 — ^(^2)) to a monopolist trying to maximize a menu of contracts to a profit by offering potential customer. becomes identical and her A> PROPOSITION 1: Suppose 1/2. The social planner will imple ment q1 = ql and q2 < q2- In equilibrium, the type-2 consumer realizes zero consumer while the consumer will realize positive surplus type-1 -» As A 1, q2 decreases. surplus. This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions 1738 Law Virginia [Vol. 98:1665 Review PROOF: When A > 1/2, in equilibrium, the participation constraint for the type-2 consumer (the fourth constraint) and the incentive com for the type-1 consumer patibility condition (the fifth constraint) will — — = 0 = If we and let bind:y2(^2) p2 —p1 vx(q2) p2~ = = and + and substitute these P2 ^2(^2) Pi vi(fli) vi(Q2) ^2(^2) into the social welfare function,we get 6(A(v1(q1) — q^)) + (2A — ~ ~ 1)0*2(92) vi(Qz))) + (1 9)A(v2(q2) c2(<?2))- When we maxim ize the objective - 0(2A From function with respect to qt and q2, we get l)(l?2(i?2) the first equation, will set the quality planner We can rewrite the second v'Mt) - = 0 c'iOji)) + (1 0X^2(92) Ö(^'i(£?i) - - C¿(<fe)) = 0 that qx = q{. That is, the social for the type-1 consumer at the optimal level. equation as = c'2^2) ~ it is clear y~Tq (2A ~ iX^Ofe) - vi(qfz)) — < 0, in order to satisfy the equality, we must v2(q2) v[(q2) have q2 < q2. The type-2 consumer will receive suboptimal quality in Furthermore, as A -> 1, the right-hand side of the equality equilibrium. Since gets smaller, making it necessary to reduce q2 more to satisfy the equali ty. Q.E.D. In equilibrium, the type-1 consumer is able to realize a positive sur = = and + v2(q2)), known as plus (since p2 ^2(^2) v^q-J vt(q2) the "informational rent" in the literature. Since the social planner cares more about the seller's profit than the buyer's surplus (A > 1/2), any that is taken away from the seller imposes an opportunity cost: surplus one dollar from the buyer to the seller increases the overall transferring social welfare when A > 1/2. Hence, to reduce the surplus captured by the type-1 consumer, the social planner introduces inefficiency on the Furthermore, as the social planner cares more and type-2 consumer. more about the distortion the seller's profit vis-á-vis the buyer's gets larger: the social planner imposes surplus, the size of even less favorable terms on the type-2 buyer. This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions The Effect of Bargaining 2012] Power 1739 Figure 5 represents the optimal set of contracts when the social plan ner cares only about the seller's profit (A = 1). The dashed curve repre sents the type-1 consumer's utility shifted down to cross at the optimal (p2,ij,2)- Note that while the type-2 nated, the type-1 consumer realizes consumer's surplus has been elimi some positive surplus. Had the so cial planner chosen the first best qualities (q{, q2), surplus captured by the type-1 consumer would have been much larger. To reduce that sur plus, the social planner reduces the quality offered to the type-2 con A marginal reduction in quality for the type-2 consumer produc es the benefit of being able to charge a higher price to the type-1 consumer and realizing a smaller profit from the type-2 consumer. At sumer. optimum, Figure the benefit will be set equal 5: Optimum when to the cost. A = 1 B. Case 2: A < 1/2 With A < function 1/2, the seller's profit gets smaller weight in the objective to the buyer's surplus in the social welfare function. compared From the objective function, ö((l — A)v1(q1) — Ac^qJ This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions + (2 A — 1740 Virginia Law Review + (1 - 0)((1 l)Pi) ~ ~ X)v2{q2) [Vol. 98:1665 - 1 Ac2(q2) + (2Ä )p2), holding everything else constant, when A < 1/2, higher prices strictly lower the value. As an extreme case, when A = 0, the objective function becomes — 0(fi(i7i) — Pi) + (1 0)(y2(q2) — p2) and the problem becomes to a monopsonist equivalent trying to maximize a take-it-or-leave-it offer to the seller. making her profit, or a buyer PROPOSITION 2: Suppose A < 1/2. The social planner will imple ment qx — ql and q2 < q2. In equilibrium, the seller realizes zero profit while the buyer captures all the surplus. As A -* 0, q2 decreases. In equilibrium, two constraints) will bind: PROOF: the seller's zero profit conditions (the first — = 0 and = 0. In ad (<7X) p2 c2(q2) dition, with the assumption of Vi(q2) - c2(q*2) > vx(q{) — cx(q{), the consumer's incentive condition compatibility (the penultimate — ~~ = constraint) will also bind: Pi vi(fo) P2- When we use the three equalities to simplify the welfare function, and set up a Lagrangian with the type-1 consumer's condition, binding incentive compatibility we get type-1 L(q,n) = (1 - A){d(v1(ql) - qCqj) + (1 - 0)(v2(q2) - c2(q2))} ~ where ~ p is the Lagrangian to (qvq2,p),weget (1 - qOZi) multiplier. - Vx(q2) + c2(q2)) When we maximize with respect = 0 (1 A)0(1 fi)(v[(ch) ci(<?i)) = 0 1 + A)( 0)(v¿(q2) c2(q2)j ß(v[(q2) c'2(q2)) ~ = 0 ViCqJ q(i?i) ViOfe) + c2(q2) To satisfythe firstequality (with p > 0), we must have q1 = q{. When the second equality is rearranged, viM = = because Suppose <72 í?2 Then, will be violated. To restore equality Kfo) - cá(<?2)) = > ^2(^2) ^(^2) c2(q2), the equality, we must have This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions the q2 < The Effect of Bargaining 2012] Power 1741 q*2.115The type-1 consumer will receive product with optimal quality but As the type-2 consumer will receive suboptimal quality in equilibrium. A -» 0, the right-hand side of the inequality gets smaller, further necessi tating the reduction of q2. Q.E.D. Figure 6 represents the optimal set of contracts when the social plan the consumer surplus (A = 0). Compared to the ner wants to maximize eliminated and the social previous case, the seller's profit is completely planner allocates able to eliminate formation: producer the seller does therefore, cannot realize Figure to the buyer. The social planner is the seller has no private in surplus not have any informational and, advantage the entire surplus 6: Optimum because any "informational when rent." A = 0 At the same time, the social planner does have to worry about keeping consumer from choosing the contract intended for the type-2 — — > ^i(íí) the assumption that v^q^) Under c2(q2) with zero Ci(<7Í), had the social planner chosen the efficient qualities the type-1 consumer. 115 Starting from q2, as we decrease q2, v[(q2) — the gap between v2(q2) c2(q2) thereby decreasing decreases and v[(q2) at a higher c2(q2). This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions rate than v2(q2) 1742 Virginia Law Review = seller profit, (p2,q2) (c2 the contract intended choosing [Vol. 98:1665 Q2.)> the type-1 consumer would prefer for the type-2 buyer. And, when both types pool on (p2,q2) = (c2 (<7l)' not only is the equilibrium ineffi cient (thereby reducing the surplus that could have gone to the type-1 a negative profit, since (in expectation) buyer), the seller also realizes the price is lower than the average cost of serving both types of buyer. Hence, to keep the seller in the market while preventing the type-1 buy er from pooling with the type-2 buyer, the social planner has to reduce the quality below the efficient level offered to the type-2 buyer. C. Case 3: X = 1/2 When seller's the social surplus, ~ 1/2 (flOiOh) planner assigns equal the objective (social + (1 - e)(v2(.q2) CiM) to the buyer and the function becomes welfare) weight - c2(q2)))- Note that the from the objective function since they only affect price terms disappear the distribution of the surplus. When the social planner cares equally about the buyer's and the seller's welfare, even though she does not di buyer type, we can show that the social planner will al implement the first best, that is, she will not introduce any distor rectly observe ways tions. A = 1/2. PROPOSITION 3: Suppose the first best: ( qt = qI, q2 = q2) The social planner implements Let q1 = q{ and q2 = q2. We just need to find the set of prices (p1( p2) that satisfy all the constraints. This can be done in the fol manner. First, if q{ — q2, this is easily achieved by letting lowing = that q{ > q2. Let px = Second, suppose Pi p2 G (cx(ql),v2(q2)). — of and <5 = v^ql) Given the simplifying assumption v2(q2) v2(q{). > 0. Given the single crossing condi that Ö we know — v2(qD ^2(^2)' PROOF: tion, v[(q) > v2(q), and the condition that q{ > - S < ^1(92) v2(q2). Once we let p2 6 (v^ql) found the solution. level, the social - q2, we must have 8,v2(q*2)), we have the respective qualities are set at the optimal cannot do any better. Q.E.D. Since planner the proof. The dotted curve repre Figure 7 graphically demonstrates sents the type-1 consumer's utility shifted down to cross at v2(ql). Sup = and sets p2 at anywhere be v2(q2), pose the social planner sets p1 has no tween v2(q2) and the dotted curve. The type-1 consumer This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions The Effect of Bargaining 2012] Power 1743 (p2, q2) since that would make her strictly worse off. the Similarly, type-2 consumer has no incentive to choose (plt q{) since that would give her zero utility while (p2< q2) gave her a strictly positive utility. Both types of consumers realize a strictly positive surplus and are incentive to choose offered When A > 1/2 or the optimal level of quality, respectively. can be varied the social planner because 1/2, quality continuously, will again impose some inefficiency in the market. However, as A ap That 1/2, the size of the inefficiency will gradually disappear. proaches is, q2 -» q*2 as A -* 1/2. A < Figure 7: Candidate First-Best Solution when A = 1/2 This content downloaded from 149.142.112.3 on Wed, 12 Feb 2014 14:25:38 PM All use subject to JSTOR Terms and Conditions