COASEAN KEEP-AWAY: VOLUNTARY

TRANSACTION COSTSJordan M. Barry,* John William Hatfield,**

& Scott Duke Kominers†

Abstract

The Coase Theorem predicts that, if there are no transaction costs, parties will always contract their way to an efficient outcome. Thus, no matter which legal rules society chooses, “Coasean bargains” will lead to efficient results.

There are always some transaction costs. However, transaction costs are often thought to be low when there are no structural impediments to negotiation, such as large numbers of parties or barriers to communication. When these obstacles are not present, it is commonly assumed that the parties will achieve an efficient result through Coasean bargaining. We show that this assumption is incorrect.

* Visiting Professor, University of Michigan School of Law; Associate Professor and Herzog Endowed Scholar, University of San Diego School of Law. ** Associate Professor, McCombs School of Business, University of Texas at Austin. † Junior Fellow, Harvard Society of Fellows; Associate, Department of Economics, Harvard University; Associate, Harvard Business School; Research Scientist, Harvard Program for Evolutionary Dynamics; Associate, Center for Research on Computation and Society, Harvard School of Engineering and Applied Sciences. We thank Emily Keifer, Phoebe Allardice, Andres Almazan, Gary S. Becker, Giuseppe Dari-Mattiacci, Noah Feldman, Lee Anne Fennel, Jerry Green, Zachary J. Gubler, James J. Heckman, Dale Jorgenson, Louis Kaplow, Elizabeth Pollman, J.J. Prescott, Alvin E. Roth, Elaine Scarry, Amartya Sen, Steven M. Shavell, and E. Glen Weyl as well as attendees at the National Business Law Scholars Conference, the Southern California Junior Faculty Workshop, and the University of San Diego faculty colloquium for their helpful thoughts and comments. This project was generously supported by a summer research grant from the University of San Diego School of Law. Kominers gratefully acknowledges the support of the National Science Foundation (Grant CCF-1216095), the Harvard Milton Fund, a Yahoo! Key Scientific Challenges Program Fellowship, a Terence M. Considine Fellowship in Law and Economics funded by the John M. Olin Center at Harvard Law School, the American Mathematical Society, and the Simons Foundation.

1

In particular, we demonstrate that transaction costs can be high, even when there are no structural impediments to bargaining, because the parties themselves may intentionally create transaction costs. Intuitively, an individual may prefer the Coasean bargain that is struck when certain parties are excluded from negotiations. Accordingly, that individual will wish to create transaction costs that keep those parties—potentially including herself—away from the negotiating table. We show that there are many contexts in which the parties will choose to create these “voluntary transaction costs,” including environmental litigation, multilateral treaty negotiations, and creditor-debtor relationships.

Because of the prevalence of voluntary transaction costs, Coasean logic applies to a significantly smaller class of cases than has previously been recognized. This renders law very important: Legal rules provide the starting point for the parties' negotiation; we find that when the parties’ starting point is closer to the efficient result, they are more likely to achieve an efficient outcome through Coasean bargaining. This insight favors reasonable use rules and other legal rules that attempt to assign entitlements in an efficient manner. We also find that liability remedies are more likely to encourage efficient outcomes than injunctive remedies are.

I. INTRODUCTION...........................................................................3II. BACKGROUND ON THE COASE THEOREM..................................8III. VOLUNTARY TRANSACTION COSTS.........................................15

A. The Private Value of Transaction Costs for Others.............17B. The Private Value of Transaction Costs for Oneself............24C. Voluntary Transaction Costs and Inefficiency.....................28

IV. IMPLICATIONS FOR CHOOSING LEGAL RULES.........................43A. Structuring Legal Rights......................................................44

1. Legal Rules That Always Produce Efficiency...................442. The Value of a More Efficient Starting Point....................47

B. The Importance of Legal Remedies......................................571. Remedies in Low- and High- Transaction Cost

Environments.....................................................................582. Injunctive Remedies vs. Liability Remedies.....................60

C. Distributional Consequences...............................................65V. Conclusion...............................................................................66

2

I. INTRODUCTION

Legal scholars have cited Ronald Coase’s The Problem of Social Cost more than they have cited any other scholarly work.1

In that article, Coase put forth what has become known as the “Coase Theorem.”2 The Coase Theorem says, essentially, that if there are no transaction costs, individuals will always organize their behavior in a way that maximizes their combined well-being. The logic is straightforward: If Ronald has the legal right to do something, and George values that right more than Ronald, George can purchase the right from Ronald. When there are no transaction costs—that is, nothing impedes people from making and enforcing contracts—people will make contracts that transfer legal rights to the individuals who will use them most productively. These contracts, known as “Coasean bargains,” always produce an efficient outcome no matter how society initially assigns legal rights.3 Thus, legal rules will have no effect on individuals’ behavior or their total wealth; they can only affect the way in which wealth is distributed.4

Coase constructed the Coase Theorem as a reductio ad absurdum to highlight how important transaction costs are: If one accepts the premise that transaction costs are negligible, it follows that legal rules do not affect behavior. Coase thought that this

1 See Fred R. Shapiro & Michelle Pearse, The Most Cited Law Review Articles of All Time, 110 U. MICH. L. REV. 1483, 1489 (2012) (listing The Problem of Social Cost as having 5,157 citations in law reviews alone, over 40% more citations than the next-most-cited article). 2 The use of the term “Coase Theorem” to describe the insight of The Problem of Social Cost is usually attributed to George Stigler. Coase himself never stated a formal theorem; this has become increasingly problematic as the field of economics has become more technical and mathematically precise. See R.H. Coase, The Problem of Social Cost, 3 J.L. & ECON. 1 (1960); GEORGE STIGLER, THE THEORY OF PRICE (1966). In our Companion Paper, we formalize the Coase Theorem and formally prove our results. See Companion Paper, infra note Error: Reference source not found, passim. 3 See, e.g., Thomas W. Merrill & David M. Schizer, The Shale Oil and Gas Revolution, Hydraulic Fracturing, and Water Contamination: A Regulatory Strategy, 98 U. MINN. L. REV. 145, 210-11 (2013); Melissa B. Jacoby & Edward J. Janger, Ice Cube Bonds: Allocating the Cost of Process in Chapter 11 Bankruptcy, 123 YALE L.J. 862 (2014). 4 This is a slight oversimplification. Legal rules change transfer payments and, in certain specific circumstances, can affect final behaviors as well. See, e.g., Guido Calabresi & A. Douglas Melamed, Property Rules, Liability Rules, and Inalienability: One View of the Cathedral, 85 HARV. L. REV. 1089, 1095-96 (1972). We address this issue more formally in Part II. See footnote Error:Reference source not found, infra, and accompanying text.

3

conclusion was clearly incorrect. Accordingly, the premise—that transaction costs are negligible—must be false.5

As Coase himself recognized, real-world agreements always involve some transaction costs, so the Coase Theorem never applies directly. Nonetheless, it is widely agreed that the Coase Theorem’s predictions generally persist when transaction costs are low.6 Transaction costs are often thought to be low when there are no structural impediments to negotiation, such as very large numbers of parties or laws that make it difficult to form or enforce contracts.7

5 See, e.g., Robert C. Ellickson, The Case for Coase and Against “Coaseanism”, 99 YALE L.J. 611 (1989).6 See, e.g., Ian Ayres & Robert Gertner, Strategic Contractual Inefficiency and the Optimal Choice of Legal Rules, 101 YALE L.J. 729, 761-62 (1992); Richard R.W. Brooks, The Relative Burden of Determining Property Rules and Liability Rules: Broken Elevators in the Cathedral, 97 NW. U. L. REV. 267, 270 n.11 (2002) (“[E]fficiency is always achieved when transaction costs are sufficiently small . . . .”); Adam Chase, The Efficiency Benefits of “Green Taxes”: A Tribute to Senator John Heinz, 11 UCLA J. ENV’L L. & POL’Y 1, 7 (1992) (“When transaction costs are negligible, . . . parties will eventually contract with one another so that they optimize the allocation of [rights].”); Lee Anne Fennell, The Problem of Resource Access, 126 HARV. L. REV. 1471, 1499 (2013) (“It is standard to assume that in a low transaction cost world, property and markets are all we need.”); George S. Geis, Empirically Assessing Hadley v. Baxendale, 32 FLORIDA ST. U. L. REV. 897, 909 n.62 (noting that “if transaction costs are [sufficiently] small” then “everyone [will] contract[] around inefficient defaults”); Pierre Schlag, Coase Minus the Coase Theorem—Some Problems with Chicago Transaction Cost Analysis, 99 IOWA L. REV. 175, 205 n.96 (2013) (“When transaction costs are sufficiently negligible [the Coase Theorem’s predictions will hold].”); David Schleicher, City Unplanning, 122 YALE L.J. 1670, 1682 (2013) (“The assignment of [a] right should not matter if transaction costs are low, as Coasean bargaining . . . should ensure that we get to the optimal [result].”); Christopher Jon Sprigman et al., What’s a Name Worth?: Experimental Tests of the Value of Attribution in Intellectual Property, 93 B.U. L. REV. 1389, 1421 (2013) (“[W]hen transaction costs are small, default rules should have no effect.”); Todd J. Zywicki, A Unanimity-Reinforcing Model of Efficiency in the Common Law: An Institutional Comparison of Common Law and Legislative Solutions to Large-Number Externality Problems, 46 CASE W. RESERVE L. REV. 961, 999-1000 (1996). 7 See, e.g., Omri Ben-Shahar, Damages for Unlicensed Use, 78 U. CHI. L. REV. 7, 27-28 (2011) (arguing that transaction costs are low between parties who have already entered into a licensing agreement); Ariel L. Bendor, Prior Restraint, Incommensurability, and the Constitutionalism of Means, 68 FORDHAM L. REV. 289, 325-26 (1999) (“[I]n the field of intellectual property, transaction costs are generally low[, i]n part [because] the intellectual property interest in a given work tends to be found in the hands of one person or a limited number of people.”); Kenneth W. Dam, Some Economic Considerations in the Intellectual Property Protection of Software, 24 J. LEG. STUD. 321, 365 (1995) (arguing that transaction costs will be low in the software industry because the parties are businesses, there are many possible transaction structures, and the universe of

4

In this Article, we show that transaction costs can be high even when there are no structural impediments to negotiation. Individuals will frequently wish to create transaction costs, even when those transaction costs prevent the formation of an efficient Coasean bargain. Because these intentionally manufactured transaction costs are created by choice, and are not an unavoidable consequence of the circumstances in which the parties find themselves, we term them voluntary transaction costs.

potential contracting parties is limited and knowable); Alan Devlin, Restricting Experimental Use, 32 HARV. J.L. & PUB. POL’Y 599, 638-40 (2009) (arguing that transaction costs will be low for patent licensing agreements because negotiations are “between a small number of easily identifiable parties over a discrete topic”); Alan Grant & Emily Grant, The Bride, the Groom, and the Court: A One-Ring Circus, 35 CAPITAL U. L. REV. 743, 753 (2007) (arguing that transaction costs will be low between parties getting engaged because there are only two, easily identified parties); Henry Hansmann & Reinier Kraakman, Toward Unlimited Shareholder Liability for Corporate Torts, 100 YALE L.J. 1879, 1893-94 (1991) (noting that corporate shareholders may easily be able to create complex agreements when there are relatively few shareholders); Trotter Hardy, Property (and Copyright) in Cyberspace, 1996 U. CHI. LEGAL F. 217, 219-237 (arguing that transaction costs are low in cyberspace because it is easy for the parties to communicate); Keith Hylton, Duty in Tort Law: An Economic Approach, 75 FORDHAM L. REV. 1501, 1513-14 (2006) (arguing that transaction costs are low in the trespass context); Eugene Kontorovitch, Liability Rules for Constitutional Rights: The Case of Mass Detentions, 56 STAN. L. REV. 755, 769-70 (2004) (arguing that well-functioning markets make transaction costs low for purchases of corporate stock); Eugene Kontorovitch, The Constitution in Two Dimensions: A Transaction Cost Analysis of Constitutional Remedies, 91 VA. L. REV. 1135, 1165-66 (2005) (arguing that transaction costs of quartering soldiers are low in peacetime, but not in wartime, because there are well-functioning rental and construction markets); Robert Merges, Intellectual Property Rights and Bargaining Breakdown: The Case of Blocking Patents, 62 TENN. L. REV. 75, 76-78 (1994) (arguing that transaction costs are low with respect to patents licensing, as “(1) there are only two parties to the transaction, and they can easily identify each other; [and] (2) the costs of a transaction between the parties are otherwise low”); Thomas W. Merrill, Trespass, Nuisance, and the Costs of Determining Property Rights, 14 J. LEGAL STUD. 13, 13-15 (1985) (arguing that transaction costs are likely to be low in the context of trespass because there are only two parties and they are generally identifiable); Maureen A. O’Rourke, Rethinking Remedies at the Intersection of Intellectual Property and Contract: Toward a Unified Body of Law, 82 IOWA L. REV. 1137, 1148, 1185-86 (1997) (arguing that transaction costs are typically low with respect to copyright and trademark licensing transactions, as “there are few parties to the transaction” and contracting rules are not onerous); Alan Schwartz, A Contract Theory Approach to Business Bankruptcy, 107 YALE L.J. 1807, 1819-20 (1998) (arguing that transaction costs are low in renegotiations of commercial sales of goods because there are few parties and the legal requirements to modify a contract are not onerous); David E. Van Zandt, An Alternative Theory of Practical Reason in Judicial Decisions, 65 TULANE L. REV. 775, 821 n.202 (1991) (concluding that transaction costs are likely to be low in landlord-tenant negotiations because there are only two parties);

5

Intuitively, an individual may prefer that some transaction costs exist because the parties will strike different deals depending on who is seated at the metaphorical negotiating table. An individual may prefer the deal that will be struck if certain parties are excluded from negotiations over the deal that is struck when those parties participate. In such instances, that individual will wish to create transaction costs that keep those parties away from the negotiating table.

To take a simple example, an investor who owns a plot of pristine woodlands might wish to sell the land to a developer who wants to build on that land. A conservationist might prefer that the land remain in its natural state. If she can, she will create transaction costs that prevent the investor and the developer from coming to the negotiating table and striking a deal.8

Moreover, in many instances an individual may wish to create transaction costs in order to keep herself away from the negotiating table. In her absence, the remaining parties at the table will often reach an agreement that closely resembles the agreement that they would have made if she had participated in the negotiations. However, by not participating in the deal, the absent individual can avoid having to contribute any of her own resources to support it. In essence, creating transaction costs can allow an individual to hitch a free ride on the labors of others. This can leave her better off than if she had participated in the negotiations.

For example, suppose that several nations convene a meeting to organize a peacekeeping action. An individual country

Developments in the Law of Cyberspace, Internet Regulation Through Architectural Modification: The Property Rule Structure of Code Solutions , 112 HARV. L. REV. 1634, 1636 & n. 13 (1999) (arguing that “transaction costs are extremely low” in cyberspace because the medium facilitates communication and the recording of data); Aaron T. Dozeman, Salinger and eBay: When Equitable Considerations Undermine Exclusivity, 21 DEPAUL J. ART, TECH. & INTELLECTUAL PROPERTY L. 323, 346-50 (2011) (arguing that transaction costs are low in negotiations to use copyrighted works because they only involve two parties); Daniel Kearney, Network Effects and the Emerging Doctrine of Cybertrespass, 23 YALE L. & POL’Y REV. 313, 338-39 (2005) (arguing that “transaction costs are likely to be low” in online interactions, even when there are many parties, because many interactions are automated and it is easy for parties to communicate). 8 This general scenario is quite common; for example, a similar dynamic is currently playing out with respect to a half-billion-dollar construction project in San Diego. See notes Error: Reference source not found-Error: Referencesource not found, infra, and accompanying text.

6

may decline to participate, even if it favors the mission and wants it to succeed: By declining to participate, it can avoid having to contribute its own resources to the project. So long as that country’s contributions are not essential, the mission will still go forward. Not participating may enable the country to reap the same benefits as it would if it participated, but at a lower cost.9

As we demonstrate, there are a large number of circumstances in which parties will wish to create voluntary transaction costs. The prevalence of voluntary transaction costs means that the Coase Theorem’s prediction—that people will negotiate their way to an efficient outcome—applies in many fewer scenarios than has previously been appreciated. In short, Coase was more insightful than perhaps even he realized.

We demonstrate that, when voluntary transaction costs are possible, the parties will typically achieve an efficient outcome under certain legal regimes, but not others. This makes the choice of legal regime increasingly important.

We explore two aspects of legal regimes that strongly affect the likelihood of an efficient outcome. The first is how the legal system initially assigns legal rights among actors. We identify some circumstances in which assigning legal rights to particular parties always encourages efficient Coasean bargains. More generally, we find that the parties’ likelihood of striking an efficient Coasean bargain depends on how the legal system initially assigns legal rights among the parties: The more efficient that initial assignment is, the more likely the parties are to strike an efficient Coasean bargain. This favors reasonable use rules and other legal rules that attempt to allocate rights in a way that resembles the efficient outcome.

The second key feature is the remedy available when a legal right is violated. In some instances, a protected party has the right to injunctive relief—that is, she can secure a court order that effectively forbids the illegal action. In other instances, the law merely provides a liability remedy; the protected party’s only

9 This dynamic applies to many other actions that require international coordination, including pollution reduction and trade sanctions. See notes Error:Reference source not found-Error: Reference source not found, infra, and accompanying text; see also Avinash Dixit & Mancur Olson, Does Voluntary Participation Undermine the Coase Theorem, 76 J. PUB. ECON. 309 (2000).

7

recourse is to collect damages from the party acting unlawfully.10

Our analysis suggests that liability remedies foster efficient Coasean bargains in a larger set of circumstances than injunctive remedies do. With respect to both legal rules and remedies, we find trade-offs between the likelihood that a particular legal regime will foster an efficient result and the amount of information that courts and legislators must have to implement that regime.

This Article proceeds as follows. Part II provides general background on the Coase Theorem and how it is commonly understood. Part III uses a series of simple examples to illustrate how voluntary transaction costs affect the Coase Theorem’s predictions. Specifically, it shows when and why parties will want to manufacture transaction costs for themselves and others, even when doing so creates inefficiencies. It also identifies real-world scenarios that are likely to exhibit dynamics that resemble our stylized examples. Part IV explores how the initial choice of legal rules and remedies can affect final outcomes, and the implications that this has for designing legal regimes. Throughout the paper, we include citations to a companion paper that contains formal proofs of our results.11 Part V concludes.

II. BACKGROUND ON THE COASE THEOREM

To understand the Coase Theorem, it is helpful to first define two concepts. First, a particular state of the world is efficient if it maximizes the combined welfare of all parties.12 An example helps illustrate this.

Suppose that Ronald has $2000 in cash and a car that he values at $1000. Suppose further that George has $2000 in cash, and that George values Ronald’s car at $2000. This state of the world is not efficient. If George were to buy Ronald’s car for $1500, both Ronald and George would be better off—Ronald 10 We use “injunctive remedies” and “liability remedies” to refer to what much of the literature refers to as “property rules” and “liability rules.” See Calabresi & Melamed, supra note Error: Reference source not found, at 1092 (starting this practice). We do so in order to sharpen the distinction between legal rules, which establish what actions are lawful, and the remedies available when those rules are violated. 11 See Jordan M. Barry et al., Voluntary Transaction Costs and the Coase Theorem, available at scottkom.com/article [hereinafter Companion Paper].12 More precisely, if it maximizes the combined utility of the parties. See HAL R. VARIAN, INTERMEDIATE MICROECONOMICS 54-55 (4th ed. 1996)

8

would go from having $3000 worth of utility to $3500,13 and George’s utility would increase from $2000 to $250014—raising their combined utility. Thus, the state of the world in which Ronald owns the car is inefficient, because Ronald and George’s combined utility is higher when George buys the car from Ronald.

Similarly, the state of the world in which George buys the car is efficient, as there is no way to reallocate resources between Ronald and George in a way that increases their combined utility. It is possible to make either Ronald or George better off by moving money between them—in other words, increasing or decreasing the price that George pays Ronald for the car—but the combined utility of Ronald and George will remain the same.

Second, we must define a transaction cost. Scholars have used this term to encompass a wide range of phenomena, but, at its broadest, it includes anything that impedes a transaction.15 This would include, for example, the time it takes the parties to negotiate a deal, the time it takes to locate the relevant interested parties, and the cost of enforcing an agreement.16 It has also been applied to circumstances that complicate the bargaining process, such as when the parties have asymmetric information17 or parties must coordinate their efforts to achieve a common goal.18

We are now able to state the Coase Theorem: If there are no transaction costs, and legal rights are clear, then no matter which legal rules society adopts, the parties will reach an efficient

13 Ronald started with $2000 in cash and a car he valued at $1000, giving him a total of $3000 of value. Now he has $3500 in cash, making him $500 better off. 14 George originally had $2000 in cash. He now has $500 in cash and a car that he values at $2000, making him $500 better off. 15 See, e.g., Calabresi & Melamed, supra note Error: Reference source notfound, at 1094-95. 16 See, e.g., Tun-Jen Chiang, The Reciprocity of Search, 66 VANDERBILT L. REV. 1, 21-22 (2013) (discussing circumstances in which it may be difficult to determine whether a party has breached an agreement).17 See, e.g., Ian Ayres & Eric Talley, Solomonic Bargaining: Dividing a Legal Entitlement to Facilitate Coasean Trade, 104 YALE L.J. 1027, 1035 (1995) (“This Article focuses on a specific type of transaction cost: private information.”); Calabresi & Melamed, supra note Error: Reference source notfound, at 1094-95 (defining transaction costs to include imperfect knowledge). 18 See, e.g., James E. Krier & Stewart J. Schwab, The Cathedral at Twenty-Five: Citations and Impressions, 106 YALE L.J. 2121, 2135 (1997) Kevin Werbach, Supercommons: Toward a Unified Theory of Wireless Communication, 82 TEX. L. REV. 863, 953 n.385 (2004).

9

result.19 This does not mean that legal rules do not matter, because they will have distributional effects—that is, legal rules can make particular parties better or worse off than they would be otherwise—but the choice of legal rules will not affect efficiency.20

The intuition behind the Coase Theorem is as follows: If society chooses legal rules that are not efficient, then, by definition, there are other ways of allocating rights among the parties that would make them better off as a group.21 If the parties can make themselves better off as a group, then they can make transfers between themselves so that each individual member of the group is better off.22 In other words, if the parties can strike a deal that grows the economic pie, everyone can have a bigger slice. Thus, in any inefficient state of the world, it is in every party’s interest to move to an efficient state. And, because there are no transaction costs—that is, there is nothing to impede the parties from making any deal that is in their interests—the parties will make and implement an agreement to move the world to an efficient state. These efficiency-creating agreements are commonly referred to as Coasean bargains.23

One of Coase’s classic examples24 illustrating his theorem involves a railroad track that runs next to a wheat field. 25 The railroad’s locomotive shoots off sparks. If these sparks hit the wheat, the field will burst into flame, which is extremely costly 19 The version of the Coase Theorem we state here depends on the presence of a numeraire good (money) that all parties value equally, as well as an assumption that parties can borrow this good, i.e., that the parties are not credit-constrained. Without these assumptions the Coase Theorem guarantees that the world will reach a Pareto efficient result, but not necessarily one that maximizes the combined utility of all parties. See, e.g., PAUL MILGROM & JOHN ROBERTS, ECONOMICS, ORGANIZATION AND MANAGEMENT (1992). We suppress these technical issues for clarity and ease of exposition: Throughout this Article, we assume that the difference in the monetary value that each actor assigns to different outcomes is finite and independent of the actor’s wealth, and that all actors have access to enough cash to support any relevant Coasean bargain. 20 RICHARD A. POSNER, ECONOMIC ANALYSIS OF LAW 63-64 (8th ed. 2011).21 Id. at 7.22 This follows from the assumptions made in footnote Error: Reference sourcenot found, supra. 23 See sources cited in note Error: Reference source not found, supra.24 See Coase, supra note Error: Reference source not found, at 29-34. Our example differs slightly from Coase’s. We return to variations on this example throughout this Article. 25 We revisit variations on this basic theme throughout this Article. See Parts III.A-C, IV.A-B, infra.

10

(that is, inefficient). There are two possible solutions: (1) the railroad can refrain from running its locomotive, or (2) the farmer—let’s call her Alice—can plant her wheat far enough back from the train track to ensure that sparks will not land on her wheat and ignite it. 26 Assume that there are no transaction costs.

First, assume that the law gives farmers the right to not have sparks on their land; in other words, assume that farmers have the right to prevent railroads from running any locomotives that shoot sparks onto their land. Suppose that running the train is worth $500 to the railroad and that Alice will lose $1000 if she has to plant back from the track. In these circumstances, the railroad will not run the train and Alice will plant near the track. This is efficient, so the parties have no incentive to strike a different deal.27

Now suppose that we change the legal rule, so that railroads have the right to shoot sparks off of their locomotives. This would be inefficient, as the railroad would be willing to forgo running the train in exchange for a payment of $500 or more, and Alice would be willing to pay up to $1000 to be able to plant wheat closer to the track. In this scenario, Alice and the railroad have an incentive to strike a deal in which Alice pays the railroad not to run the train so that she can plant closer to the track. As long as Alice pays the railroad more than $500 but less than $1000,28 both parties will be strictly better off.29 We assume here, and throughout the paper generally, that the parties at the table split their joint gains

26 It is often tempting to see one party as the victim of the other’s behavior, but this is frequently counterproductive. Externalities in the Coasean sense are inherently symmetric: If the farmer has the right to have no sparks on his land, then the railroad does not have the right to emit sparks from its locomotive. Giving the farmer the legal right harms the railroad and vice versa. The right to engage in an activity is equivalent, from an economic perspective, to a right to be free from harm. See ROBERT H. FRANK, THE DARWIN ECONOMY 90-95 (2011).27 Recall that the benefit that the railroad would receive from running the train is less than the harm that Alice would suffer from planting farther back from the track. Accordingly, changing the status quo will not produce net gains. 28 Equivalent arrangements are possible in which both parties make payments; for example, Alice could pay the railroad $3000 and the railroad could pay Alice $2200. For simplicity and clarity of exposition, all of the cash payments described in this example and throughout this paper are net payments. 29 Alice pays less than $1000 for a benefit worth $1000. The railroad loses a benefit of $500 but receives more than $500 in cash. Thus, both parties are better off.

11

evenly. We adopt this assumption merely to simplify our exposition; we note that it is not necessary for our analysis. Thus, Alice pays the railroad $750, leaving both her and the railroad $250 better off than they would be if the railroad ran its train past Alice’s field.30

Observe that in both of the cases discussed above, the railroad did not run its train and Alice planted wheat right up to the railroad track. Even though the legal rule was different, the parties reached the same efficient result in both scenarios. The legal rule in these scenarios is much like the car in the prior example—an item that some parties value more than others, and which the parties can trade among themselves. Note also that the choice of legal rule does have distributional effects—the railroad is $750 better off in the second scenario than the first,31 and Alice is $750 better off in the first scenario than the second.32

Now suppose that the facts are exactly as described above, except that the railroad reaps a benefit of $2000 when it runs its train instead of $500. If the law protects farmers from having sparks on their land, Alice would be willing to plant further back from the track, as long as she was paid $1000 or more to do so.33

At the same time, the railroad would be willing to pay up to $2000 to be able to run its train.34 Once again, Alice and the railroad will have an incentive to strike a deal. So long as the railroad pays Alice more than $1000 and less than $2000 to plant her crops further back from the track, both Alice and the railroad will be better off. Assuming once again that Alice and the railroad split the gains from the transaction, the railroad will pay Alice $1500, leaving both parties $500 better off.35

30 The parties’ transaction creates $500 of joint gains. Each party gets half of that: Alice pays $750 and receives $1000 in benefits; the railroad gets $750 and loses $500 in benefits. 31 In both scenarios, the railroad does not run the train, but in the second scenario it receives a $750 payment from Alice.32 In both scenarios, Alice plants wheat up to the track, but in the first she makes a $750 payment to the railroad.33 Planting back from the track would cost Alice $1000 in forgone wheat, so she must receive a cash payment of at least this amount to surrender her right. 34 Running the train gives the railroad a $2000 benefit, so the railroad is willing to pay up to $2000 cash to acquire the right to run the train. 35 The parties’ bargain creates $1000 in net benefits. The railroad pays $1500 and gets a benefit of $2000 from running the train. Alice gets $1500 and loses $1000 in forgone wheat. Thus, each party nets $500, one half of the gains

12

If the law gives railroads the right to shoot sparks off of their locomotives, the railroad will run the train and Alice will plant away from the track. This is an efficient outcome, so the parties will not have any incentive to strike further deals.36

Once again, in both of these cases, Alice plants back from the track and the railroad runs its train. This is an efficient result because the cost to Alice of planting away from the track ($1000) is less than the benefit to the railroad from running the train ($2000). There are distributional consequences to the choice of legal rule, however; Alice is $1500 better off in the first scenario compared to the second,37 and the railroad is $1500 better off in the second scenario compared to the first.38

All of these examples assume that there are zero transaction costs (as does the Coase Theorem itself). Because nothing impedes Alice and the railroad from making deals, they can easily and completely circumvent any inefficient legal rule and reach an efficient result.

Assuming that there are no transaction costs is a tremendous assumption. In reality, there are always some transaction costs39—for example, it will always take at least a small amount of time to work out a deal40—but, when transaction costs are low, the basic logic of the Coase Theorem still applies.41 On the other hand, when parties face large transaction costs, the legal rule—and not the relative benefits of different activities—may determine how parties behave.42 In such cases, legal rules may

produced. 36 The railroad would lose $2000 if it lost the right to run its train, while Alice would only gain half of that from being able to plant wheat closer to the track. Thus, there are no gains to be had from trade. 37 In both scenarios, Alice does not plant wheat up to the track, but in the first she receives a $1500 payment from the railroad. 38 In both scenarios, the railroad runs the train, but in the second it does not make a $1500 payment to Alice. 39 See, e.g., POSNER, supra note Error: Reference source not found, at 65.40 Id.41 Id.42 For example, suppose again that running the train produces benefits of $500 for the railroad, and that if the train does not run $1000 of additional wheat can be planted near the track. However, assume that $1000 in lost profit is spread among 200 farmers whose fields border the train track, each of whom loses $5. The choice of legal rule is likely to be important, because it will probably be very difficult to coordinate the 200 farmers to act together and raise enough

13

control whether the parties reach an efficient outcome, as well as how value is distributed among the various parties.

When assessing the magnitude of transaction costs standing in the way of an agreement, commentators have generally focused on structural features of the negotiations. For instance, scholars have frequently considered the number of parties,43 how easily the parties can communicate with each other,44 whether the parties have the same information,45 whether the agreement being negotiated has simple terms,46 and how easy it is for a party to enforce an agreement after it is made.47 When these transaction costs, which we term “structural transaction costs,” are low, commentators have often accepted the Coase Theorem’s predictions.48

money to pay off the railroad not to run its train. Similarly, if the railroad gets a benefit of $2000 from running the train, it will be costly for the railroad to track down all of the farmers and make a deal with each one to plant wheat farther back from the track. Parties also may not reach a deal if they have a strained or contentious relationship (which can often happen after protracted litigation). See Ward Farnsworth, Do Parties to Nuisance Cases Bargain After Judgment? A Glimpse Inside the Cathedral, 66 U. CHI. L. REV. 373 (1999) (examining 20 nuisance cases that produced injunctions and finding that parties did not bargain after any of them). 43 Commentators often assume that transaction costs are likely to be low when there are few parties involved, and high when there are many parties involved. See, e.g., POSNER, supra note Error: Reference source not found, at 77; Zywicki, supra note Error: Reference source not found, at 961-62. However, bilateral monopolies can encourage strategic behavior, which is itself a transaction cost. See POSNER, supra note Error: Reference source not found, at 77-78. 44 See, e.g., Hardy, supra note Error: Reference source not found; Mark A. Lemley, Shrinkwraps in Cyberspace, 35 JURIMETRICS 311, 323 (1995). This might be violated if, for example, the parties do not speak the same language, either literally or figuratively. See, e.g., William Hubbard, The Competitive Advantage of Weak Patents, 54 BOSTON COLLEGE L. REV. 1909, 1948 (2013). 45 The greater the extent to which the parties have asymmetric information, the higher that transaction costs are usually expected to be. See, e.g., Ayres & Talley, supra note Error: Reference source not found, at 1035; Calabresi & Melamed, supra note Error: Reference source not found, at 1094-95. 46 The logic is that agreements with more complicated terms take more time and effort to negotiate and draft. Thus, these agreements involve larger transaction costs. See, e.g., Devlin, supra note Error: Reference source not found, at 638-40; Hansmann & Kraakman, supra note Error: Reference source not found, at 1893-94.47 If it is relatively simple for a party to determine whether the agreement has been honored and to be made whole for any damages she suffers as a result of her counterparty’s non-compliance, transaction costs are expected to be lower. To the extent that more costly monitoring actions are necessary, transaction costs are expected to be higher. See, e.g., Hylton, supra note Error: Referencesource not found, at 1513-14.48 See, e.g., sources cited in footnote Error: Reference source not found, supra.

14

Structural transaction costs are undeniably important; large structural transaction costs can prevent the parties from making agreements. However, structural transaction costs are not the only impediment to bargaining. As we demonstrate in Part III, even when there are no structural transaction costs, in many instances the parties themselves will intentionally generate transaction costs. Thus, the zero-transaction-cost assumption is an even more herculean assumption than is commonly recognized.

III. VOLUNTARY TRANSACTION COSTS

The Coase Theorem, and the basic logic underlying it, have generally been explored and understood through a few famous examples.49 These examples are designed to have low transaction costs, and they generally resemble the examples described in Part II. They involve exactly two parties. The parties’ identities vary— a farmer and a railroad;50 a doctor and a confectioner;51 a farmer and a rancher;52 etc.53—but they are almost always neighbors who 49 There is also a second category of examples, designed to show how the logic of the Coase Theorem breaks down when transaction costs are high. These examples typically feature a large number of parties who are difficult to coordinate. Common examples include motorists disturbing patients by honking their horns as they drive by a hospital; traffic congestion; and overfishing of global fisheries. However, the most common version involves a factory emitting pollution that affects a large number of homeowners. See, e.g., POSNER, supra note Error: Reference source not found, at 78-79; Jodi Beggs, The Coase Theorem, ABOUT.COM (last visited Jan. 13, 2004) (providing a more modern spin by discussing a wind farm whose turbines produce noise that bothers nearby households). Because there are so many homeowners, organizing them into a unified group that can jointly take action is assumed to be impossible. If the factory has the right to pollute, but doing so is not efficient, the residents will not be able to coordinate and collectively buy off the factory. Conversely, if each resident has the right to prevent the factory from polluting, the factory will not be able to arrange a purchase of that right from each resident. Thus, in these examples, Coasean bargains are deemed impossible, and the legal rule controls the outcome. Cf. Alan Schwartz & Joel Watson, The Law and Economics of Costly Contracting, 20 J.L. ECON. & ORG. 2, 3 (2004) (noting that “contracting and renegotiation costs are . . . commonly assumed to be either very high or very low” and “explor[ing] the middle ground”). 50 See, e.g., Robert C. Ellickson, Of Coase and Cattle: Dispute Resolution Among Neighbors in Shasta County, 38 STAN. L. REV. 638 (1986).51 See Sturges v. Bridgman, LR 11 Ch D 852 (1879). 52 See, e.g., Graciela Kuechle, & Diego Rios, The Coase Theorem Reconsidered: The Role of Alternative Activities, 32 INT’L REV. L. & ECON. 129, 129-30 (2012).53 See, e.g., Coase, supra note Error: Reference source not found, at 8 (discussing a building that obstructed currents of air, impairing a windmill’s operation); id. (discussing a building that cast a shadow on a nearby hotel’s sunbathing area); id. at 37 (giving an example involving one neighbor whose

15

must resolve a question of conflicting rights.54 Does the confectioner have the right to use his kitchen as he sees fit, even though it makes the doctor’s examination room noisy, or does the doctor have the right to quiet in his examination room, even though it means the confectioner cannot use his kitchen?55

These examples are useful, but they are not representative of the broad class of problems to which the logic of the Coase Theorem is often applied. In particular, they do not consider scenarios in which there are more than two parties. In such circumstances, parties will often have incentives to create transaction costs.56 These voluntary transaction costs confound the mechanism underlying the Coase Theorem, rendering its logic applicable to a smaller class of circumstances than has previously been recognized.

We begin by demonstrating why parties might wish to create transaction costs for others, before turning to scenarios in which they may wish to create transaction costs for themselves.57

rabbits eat another neighbor’s crops). 54 See Calabresi & Melamed, supra note Error: Reference source not found, at 1101 (describing parties in an example illustrating the Coase Theorem as “inevitably neighbors”); Ellickson, supra note Error: Reference source notfound.55 See Coase, supra note Error: Reference source not found, at 2.56 In the two-party case, voluntary transaction costs should not arise because it is in neither party’s interest to create them. See Companion Paper, supra note Error: Reference source not found, at thm 1. 57 For ease of exposition, we generally restrict our focus going forward to a particular class of transaction costs that are easy to understand and model. In the examples that follow, and for much of the remainder of this Article, we use the term transaction costs to refer to costs that a party must pay to make any deal at all; they are, quite literally, the cost of engaging in a transaction. Cf. Farnsworth, supra note Error: Reference source not found (considering the transactional barriers created by unhappiness about bargaining over a particular topic, or with a particular party). Our definition fits comfortably within the heart of the definition of transaction costs usually used in the literature, though that definition encompasses other phenomena as well. See footnote Error:Reference source not found, supra, and accompanying text. Moreover, we note that our definition captures more of these other phenomena than might initially be apparent. For example, a party may be able to eliminate the problem of asymmetric information by conducting her own costly investigations. Similarly, coordination costs are essentially the time, effort, and other resources required to assemble and organize collective action. One must commit those resources in order to achieve the desired deal; thus, there is a cost to engaging in such transactions as well.

16

We illustrate both of these dynamics by revisiting, and tweaking slightly, our example from Part II.58

A. The Private Value of Transaction Costs for Others

There are many instances in which one or more individuals will wish to create transaction costs for other parties. For example, this dynamic arises whenever the following two criteria are met59: First, there is at least one “dispensable party”—a party whose participation in the Coasean bargain is not necessary to achieve an efficient result. Second, there is a dispensable party whose preferred outcome is an inefficient one. More precisely, the dispensable party prefers the Coasean bargain that would arise if certain individuals (the “opposition”60) did not participate in the bargaining process over the efficient Coasean bargain that would arise if everyone participated.61

To see why parties often wish to create transaction costs for others, it is helpful to consider two separate possibilities. First, suppose that the parties strike a Coasean bargain under which the dispensable party does not receive any cash payments.62 By assumption, there is some other Coasean bargain that (1) the dispensable party would like better, and (2) would arise if the opposition were excluded from the negotiating process.63

Accordingly, the dispensable party will wish to create transaction costs that keep the opposition away from the bargaining table and bring about her preferred outcome.

58 Throughout this Article, we ignore mixed-strategy equilibria for two reasons: First, it simplifies our analysis. Second, we primarily focus on the efficiency of various outcomes. Our examples involve unique efficient outcomes; thus, mixed-strategy equilibria are inefficient. 59 These conditions are sufficient, but not necessary. See footnotes Error:Reference source not found-Error: Reference source not found, infra, and accompanying text (discussing a scenario that does not satisfy these conditions and in which parties wish to create transaction costs for others). 60 Cf. FRANK, supra note Error: Reference source not found, at 90-95. 61 See Companion Paper, supra note Error: Reference source not found, at § 2.1. 62 Again, the relevant measure is net payments. See note Error: Referencesource not found, supra.63 See text accompanying note Error: Reference source not found, supra.

17

On the other hand,64 suppose that the parties strike a Coasean bargain that gives the dispensable party a cash payment.65

In this scenario, the opposition has an incentive to exclude the dispensable party from the bargaining table66: By definition, the dispensable party is not needed to achieve efficiency.67 Therefore, the remaining parties can still produce the efficient result in her absence. This means that the size of the economic pie will remain the same if the dispensable party is excluded from the bargaining table. However, if the dispensable party is excluded from the bargaining process, she will no longer receive any payments.68 In other words, the opposition can take the payments that would have gone to the dispensable party and divide them among themselves instead. Consequently, the opposition wants to create transaction costs that keep the dispensable party away from the bargaining table.69

An example helps illustrate these points. Consider once again a railroad whose track runs next to farmer Alice’s wheat fields. The railroad’s locomotive emits sparks, and if these sparks land in Alice’s wheat fields, the fields will ignite, which is extremely costly. As before, we assume that there are only two ways to avoid this problem: (1) the railroad can refrain from running its train, at a cost to the railroad of $500, or (2) Alice can plant her wheat far enough back from the track to ensure that errant sparks will not ignite her fields, at a cost to her of $1000.

We now depart from our previous example by adding a new person, Zelda. Zelda is a conservationist who owns undeveloped, forested land that also abuts the train track. The sparks emitted by the locomotive ignite dry brush on Zelda’s land. 64 Note that, between the first and the second case, we cover all possibilities; the parties’ Coasean bargain must either (1) provide for expected cash payments to the dispensable party, or (2) not provide such payments. 65 In other words, assume that the cash payments that she receives from the other parties to the Coasean bargain are larger than the cash payments that she must make to other parties pursuant to the Coasean bargain. 66 More precisely, it is not just the opposition who has this incentive. Every other party to the Coasean bargain will also wish to exclude the dispensable party and capture some of the cash payments that the dispensable party would have received. 67 See text accompanying note Error: Reference source not found, supra.68 See Companion Paper, supra note Error: Reference source not found, at 4 n.5. 69 The precise incentives of each of the remaining parties to exclude the dispensable party depend on how the payments that would have gone to the dispensable party are divided among the remaining parties.

18

However, from Zelda’s perspective, these sparks are a positive; the regular train schedule produces many small fires that help clear away dead underbrush and prevent large and dangerous forest fires from developing on her property.70 If the train does not run, Zelda can accomplish the same result by doing controlled burns on her land, at a cost to her of $200.71 Thus, unlike Alice, Zelda reaps a $200 benefit when the locomotive runs.72

Suppose that the law gives the railroad the right to emit sparks. If there are no transaction costs, Alice and the railroad will make a Coasean bargain in which Alice pays the railroad not to run its train. This would produce an efficient result—not running the train costs the railroad $500 and Zelda $200, but it provides Alice a $1000 benefit.73

However, even if there are no structural transaction costs, one or more of the parties will always have an incentive to manufacture transaction costs. The dynamic that we describe below applies regardless of how the parties divide up the gains from bargaining; however, it is easiest to understand in the context of a specific numerical example. We therefore assume, as before, that the parties divide the gains produced by their Coasean bargain evenly among themselves.74 Accordingly, Alice will pay $600 to the railroad and $300 to Zelda, totaling $900, and will be able to grow an additional $1000 worth of wheat. This leaves each party $100 better off than they were before they struck their Coasean bargain.75

70 Cf. U.S. FOREST SERV., MANAGING WILDLAND FIRES: PRESCRIBED FIRE, available at fs.fed.us/fire/management/rx.html, last visited Feb. 23, 2014 (describing how, without periodic fires, certain ecosystems can become unhealthy for flora and fauna and dangerous for people).71 Cf. id. (describing how the U.S. Forest Service engages in prescribed fires to ensure that certain lands do not go too long without a fire). 72 Many other similar examples can easily be constructed. For instance, Zelda could be a bed-and-breakfast owner whose guests like to see the train go by.73 The total cost of not running the train is $700 ($500 to the railroad and $200 to Zelda), which is $300 less than the $1000 benefit to Alice. 74 See text accompanying notes Error: Reference source not found and Error:Reference source not found, supra. 75 The parties’ transaction produces $300 net gains, so splitting those gains means giving each party a $100 benefit: Alice pays a total of $900 and receives $1000 more wheat. Zelda receives $300 and loses a $200 benefit. The railroad receives $600 and loses a $500 benefit.

19

Consider the Coasean bargain that would be struck without Zelda’s participation, however. These circumstances are essentially identical to those of an example discussed in Part II: Alice and the railroad will split the $500 joint gains that they receive from the transaction, leaving each $250 better off.76 Thus, both Alice and the railroad are $150 better off if Zelda is excluded from the Coasean bargain. Accordingly, they would each be willing to pay up to $150 to create transaction costs that prevent Zelda from participating in the bargaining process.

It is worth emphasizing that this dynamic does not depend on the precise way that the parties divide the surplus: Zelda’s participation is not necessary to achieve an efficient result. Thus, if the Coasean bargain provides for any payment to Zelda, either Alice, the railroad, or both are better off when Zelda is excluded from the bargaining process. Hence, either or both Alice and the railroad will create transaction costs for Zelda if possible.

The only circumstance under which neither Alice nor the railroad will want to create transaction costs that keep Zelda away is when Zelda does not receive any cash payments. However, in such circumstances, Zelda will want to create transaction costs that prevent Alice and the railroad from negotiating with each other: If Alice and the railroad cannot negotiate, the railroad will run its train. This will give Zelda $200 more in benefits that she receives if Alice and the railroad negotiate a deal that does not provide Zelda with a cash payment.77

Putting this all together, it will always be the case that either (1) Alice and the railroad have an incentive to keep Zelda away from the negotiating table, or (2) Zelda will have an incentive to keep Alice and the railroad from negotiating with each other.78 Even if there are no structural transaction costs, overall

76 See footnote Error: Reference source not found, supra, and accompanying text. 77 If the railroad does not run the train and Zelda gets no cash payment, she nets zero. If the railroad runs the train, Zelda gets a $200 benefit. 78 This situation has some similarities with, but is distinguishable from, the problem of the empty core, which is well documented in the context of the Coase Theorem. See, e.g., Yakar Kannai, "The Core and Balancedness", in 1 HANDBOOK OF GAME THEORY WITH ECONOMIC APPLICATIONS, 355–395 (Robert J. Aumann & Sergiu Hart, eds., 1992); Olga N. Bondareva, Some Applications of Linear Programming Methods to the Theory of Cooperative Games, 10 PROBLEMY KYBERNETIKI 119 (1963); Lloyd S. Shapley, On

20

transaction costs may still be large because the parties will create voluntary transaction costs if they can.

It is not difficult to come up with a bevy of real-world scenarios in which an individual prefers to exclude other parties with opposing interests. For instance, at auctions, bidders often wish to exclude other would-be bidders in order to keep sale prices lower.79 Similarly, producers are typically happy to increase their market power by preventing competitors from having access to their customers.80

More interestingly, the chief goal of many lawsuits is to impose transaction costs on parties who are attempting to bring about a result that the plaintiff does not like. These lawsuits’ purpose is to make it harder for the defendants to strike certain bargains. Often, these lawsuits challenge the procedure through which the defendants made the deal in question.81 Even when

Balanced Sets and Cores, 14 NAVAL RESEARCH LOGISTICS Q. 453 (1967); see also Varouj A. Aivazian & Jeffrey L. Callen, The Coase Theorem and the Empty Core, 24 J.L. & ECON. 175 (1981); Ronald H. Coase, The Coase Theorem and the Empty Core: A Comment, 24 J.L. & ECON. 183 (1981); Varouj A. Aivazian & Jeffrey L. Callen, The Core, Transaction Costs, and the Coase Theorem, CONST’L POL. ECON. (2003).79 See, e.g., Robert C. Marshall & Michael J. Meurer, Bidder Collusion and Antitrust Law: Refining the Analysis of Price Fixing to Account for the Special Features of Auction Markets, 72 ANTITRUST L.J. 1 (2004); Katherine Sawyer, Mobile Investor to Plead Guilty in Foreclosure Auction Bid-Rigging Scheme, TIMES-PICAYUNE, Apr. 30, 2012, available at blog.al.com/live/2012/04/ mobile_investor_to_plead_guilt.html.80 Agreements among competitors to divide territory geographically is illegal under the Sherman Act. See, e.g., U.S. DEP’T OF JUSTICE, ANTITRUST RESOURCE MANUAL § 8. 81 For example, environmental groups sometimes bring lawsuits to halt proposed construction projects on the grounds that the developer has not undertaken certain procedural steps, such as conducting legally mandated studies or allowing sufficient opportunity for public comment. Similarly, parties sometimes challenge the validity of new regulations issued by a government agency on the ground that the agency did not follow the proper procedure, such as giving notice to the public and allowing an appropriate amount of time for the public to submit comments. See, e.g., Electronic Privacy Information Center v. Dep’t of Homeland Security, 653 F.3d 1 (D.C. Cir. 2011) (ruling that the Department of Homeland Security violated the Administrative Procedures Act when it decided to screen airline passengers through advanced imaging technology instead of magnetometers and remanding for further proceedings); Lucia Graves, Groups Against Keystone XL Oil Pipeline Sue to Block Construction, HUFFINGTONPOST.COM, Dec. 5, 2011, huffingtonpost.com/ 2011/10/05/groups-sue-to-block-construction-of-keystone-xl_n_996075.html; Suzanne Stevens, Environmental Groups Expand Lawsuit to Block Tesoro Oil Terminal, SUSTAINABLE BUS. OREGON, Nov. 19, 2013, available at sustainablebusinessoregon.com/articles/2013/11/environmental-groups-expand-

21

these suits succeed, the defendants can usually still proceed with their deal, so long as they first engage in the required steps. In some instances, the added cost and delay is sufficient to prevent the project altogether.82 Other times, the plaintiff is able to extract concessions from the defendants in exchange for dropping its suit.83 This dynamic resembles the scenario in which conservationist Zelda attempts to create transaction costs that keep farmer Alice and the railroad from striking a deal.84

The parallels are easiest to see in the context of a specific example. One of the biggest current issues in San Diego politics is a $525 million proposed expansion of the city’s waterfront convention center.85 In October, 2013, San Diego port and city officials (the “San Diego officials”) received a necessary approval for the project from the California Coastal Commission (the “CCC”).86 The San Diego Navy Broadway Complex Coalition (“SDNBCC”) then filed suit in an attempt to block the expansion project from going forward.87 The SDNBCC’s legal challenge is based on alleged procedural failings in the CCC’s approval process.88 The SDNBCC opposes the expansion because it

lawsuit.html; Ty West, Lawsuit Filed to Block Northern Beltline Permit, BIRMINGHAM BUS. J., Oct. 25, 2013, available at bizjournals.com/birmingham/ news/2013/10/25/northern-beltline-lawsuit-filed.html. 82 Cf. Lucian Arye Bebchuk et al., The Powerful Antitakeover Force of Staggered Boards: Theory, Evidence, and Policy, 54 STAN. L. REV. 887, 904 927-29 (2002) (finding that, when anti-takeover devices extend the time required to acquire a company to over one year, acquisitions do not happen); Jordan M. Barry & John William Hatfield, Pills and Partisans: Understanding Takeover Defenses, 160 U. PENN. L. REV. 633 (2012) (exploring the implications of this dynamic). 83 See, e.g., MAYER BROWN, U.S. ENVIRONMENTAL LAWS INCREASINGLY LEAD TO LITIGATION CONCERNING THE SITING AND OPERATION OF RENEWABLE ENERGY PROJECTS (2010), available at mayerbrown.com/files/Publication/.84 See Part III.A, supra. 85 See, e.g., Padma Nagappan, Convention Center Expansion Plans Mired in Controversy, DAILY TRANSCRIPT, Aug. 30, 2012; James R. Riffel, Filner, DeMaio Debate Convention Center Expansion Financing, KPBS.ORG, Oct. 3, 2012, kpbs.org/news/2012/oct/03/filner-demaio-debate-convention-center-expansion-f/. 86 Lori Weisberg, Commission OKs Convention Expansion, SAN DIEGO UNION-TRIBUNE, Oct. 10, 2013, available at utsandiego.com/news/2013/Oct/10/coastal-commission-convention-expansion-plan-bay/.87 Roger Showley, Convention Center Lawsuit Filed, SAN DIEGO UNION-TRIBUNE, Nov. 25, 2013, available at utsandiego.com/news/2013/nov/25/convention-center-expansion-lawsuit/. 88 See id. (identifying challenges to the approval process, including the failure to consider alternative projects, according insufficient opportunity for public comment, the CCC’s failure to issue a recommendation in a specific form at a

22

believes that the expansion will reduce the public’s access to the San Diego Bay and have adverse effects on the environment.89

This example tracks our farmer/conservationist example well.90 The SDNBCC resembles conservationist Zelda, the CCC resembles the railroad, and the San Diego officials resemble farmer Alice. Left to their own devices, the San Diego officials and the CCC would proceed with the expansion project,91 just as Alice and the railroad would reach an agreement not to run the train.92 The SDNBCC, like Zelda, prefers the outcome when the parties do not bargain93 over the bargain that will otherwise occur.94

Accordingly, the SDNBCC filed its lawsuit and created transaction costs for the other parties. And, by creating those transaction costs, the SDNBCC may be able to extract concessions from the other parties95—essentially equivalent to cash payments in our stylized example.96 Many other real-world examples also fit this general mold.97

Finally, note that the other parties’ incentives to exclude a dispensable party stem from the cash payments that the dispensable party receives. The larger those payments are, the more incentive the other parties have to kick out the dispensable particular time, and some CCC members’ failure to disclose communications in a timely fashion). 89 See Weisberg, supra note Error: Reference source not found. 90 In analogizing to the farmer/conservationist example discussed above, we do not mean to imply that the proposed expansion is efficient. We take no position on that issue. 91 See Weisberg, supra note Error: Reference source not found (discussing the negotiations between the CCC and San Diego officials). 92 See footnotes Error: Reference source not found and Error: Reference sourcenot found, supra, and accompanying text. 93 That is, the legal default rule. See footnote Error: Reference source not found, supra, and accompanying text.94 See footnote Error: Reference source not found, supra, and accompanying text.95 For example, SDNBCC would likely prefer a prior proposal, which included a large pedestrian bridge into the expansion. This option was rejected on the grounds that it would be too costly, but that decision could potentially be reversed. See Weisberg, supra note Error: Reference source not found. 96 See Part III.A, supra.97 A similar story could be told about the same project with respect to organized labor instead of SDNBCC. Labor groups originally opposed the project and threatened litigation, but changed their position after reaching an agreement with the project’s contractor. See Lori Weisberg, Convention Center Labor Pact Ignites Controversy, SAN DIEGO UNION-TRIBUNE, Nov. 12, 2012, available at utsandiego.com/news/2012/Nov/12/convention-center-labor-pact-ignites-controversy/.

23

party and divide up the payments that would have gone to her. Note also that the dispensable party is likely to receive larger payments when the parties at the table divide up the gains from their agreement more equitably.98 Hence, the incentives to exclude dispensable parties from negotiations altogether become stronger as the parties at the bargaining table allocate the benefits of the transaction among themselves in a more egalitarian manner.

B. The Private Value of Transaction Costs for Oneself

There are also scenarios in which it is in a party’s best interest to create transaction costs that keep her from participating in any Coasean bargain. Consider once more our prior example with a railroad that borders farmland. However, unlike our previous example, suppose that there are now two farmers, Alice and Bob, whose wheat fields abut the train tracks. As before, not running the train will cost the railroad $500, and it will cost Alice and Bob $1000 each to plant their crops away from the tracks.

Assume that the law gives the railroad the right to emit sparks from its locomotive. In the absence of any agreement, the farmers will be forced to plant back from the tracks. However, the Coase Theorem predicts that, as long as there are no transaction costs, this will not happen: Because planting back from the tracks costs Alice and Bob more than it would cost the railroad not to run the train, Alice and Bob should pay the railroad not to run the train. Assuming once more that the parties evenly divide the gains produced by their agreement,99 Alice and Bob each pay the railroad $500,100 leaving each party $500 better off than she would be in the absence of an agreement.101

So far, having three actors in our example instead of two has not had much effect. The basic transaction is the same: The

98 The dispensable party is receiving cash payments, even though her participation is not necessary to achieve efficiency; thus, she is receiving a share of the gains, even though she did nothing to create those gains. 99 See text accompanying notes Error: Reference source not found and Error:Reference source not found, supra. 100 The exact split is not important to our conclusions; it simply facilitates and clarifies the analysis somewhat. 101 The railroad loses $500 by not running the train, but receives $1000 cash from the farmers, for a net profit of $500. Alice and Bob each pay $500, but earn $1000 in additional profit from growing additional wheat.

24

farmer buys off the railroad, just as in our original example.102 The only difference is that now there are two farmers who team up to buy off the railroad instead of one farmer acting alone.

Now suppose that Bob faces high transaction costs: In order to be party to any agreement, Bob must pay $2000. The most that Bob can hope to benefit under a Coasean bargain is $1500, the net value produced from not running the train.103

Because that is less than the $2000 cost that Bob must incur to make any agreement, it is never in Bob’s interest to participate in any transaction. Bob should never come to the negotiating table.

Taking Bob out of the picture leaves just Alice and the railroad. Because Alice’s gains from planting additional wheat ($1000) exceed the benefit that the railroad receives from running the train ($500), Alice should still buy off the railroad without Bob’s involvement: As long as Alice pays the railroad more than $500 and less than $1000 not to run the train, both she and the railroad will be better off.104

Three things are worth noting here. First, even without Bob’s involvement, the parties have resolved the legal entitlements question in the same way; the train does not run in either scenario.

Second, even though Bob has not participated in the negotiation, he still benefits from its results: By not running its train, the railroad protects both Alice’s and Bob’s fields from sparks and enables them both to plant more wheat; the same track and the same trains run (or do not run) by both farms. Thus, Bob will still pocket $1000 in profits from growing wheat near the track. Bob’s lack of involvement in the negotiation has not made the bargaining outcome any less favorable to him.

Third, by not participating in the agreement, Bob has avoided having to make any payments to the railroad. This leaves

102 See footnotes Error: Reference source not found-Error: Reference source notfound, supra, and accompanying text.103 If he receives any more than $1500, at least one of the other parties to the agreement must be paying out more money than she receives in benefits from the transaction. Accordingly, that party would be better off rejecting the Coasean bargain. 104 Applying the same criteria as in our prior example, we would expect Alice to pay the railroad $750, leaving both her and the railroad with a gain of $250 from the transaction.

25

Bob $500 better off than he was in the previous example, in which he grew the same amount of wheat but paid $500 to the railroad.

Thus, the $2000 transaction costs that Bob faced have left him $500 better off. And, although Bob benefits from these transaction costs, Alice and the railroad suffer. Since Bob now keeps $1000 of the gains created by the transaction, Alice and the railroad have only $500 in gains left to split between them.

Because Bob is better off with the $2000 transaction costs than without them, he prefers when they are present. Moreover, if they are not present, Bob will have a strong incentive to manufacture them, and would happily pay up to $500 to do so.105

The phenomenon of individuals intentionally making themselves unavailable will be familiar to most people who have been tasked with rounding up volunteers to serve on an office committee, or who have tried to get roommates to help clean up the common areas of an apartment.106 This dynamic also applies to fundraising for public goods, such as parks or fireworks displays;107 indeed, the difficulty of collecting voluntary contributions for public goods is one of the chief economic justifications for empowering governments to levy and collect taxes.108

It is worth noting that this same dynamic can also affect negotiations between governments. For example, countries often enter into multilateral treaties as a way to take coordinated action toward a common goal. For instance, environmental treaties, in which signatories commit to taking coordinated action to reduce pollution, often fit this mold.109 These treaties generally require

105 On the other hand, Alice and the railroad would be willing to pay up to $500 to eliminate Bob’s transaction costs.106 All of these problems are essentially free-rider problems. See MANCUR OLSON, THE LOGIC OF COLLECTIVE ACTION (1965).107 See id.108 See VARIAN, supra note Error: Reference source not found, at 605-26. 109 See, e.g., UNITED NATIONS, FRAMEWORK CONVENTION ON CLIMATE CHANGE, KYOTO PROTOCOL (1997), available at unfccc.int/kyoto_protocol/ items/2830.php. (reducing emissions); UNITED NATIONS, MONTREAL PROTOCOL ON SUBSTANCES THAT DEPLETE THE OZONE LAYER (1997), available at ozone.unep.org/new_site/en/ (phasing out production of ozone-depleting substances); UNITED NATIONS, PROTOCOL ON FURTHER REDUCTION OF SULPHUR EMISSIONS (1994), available at unece.org/env/lrtap/fsulf_h1.html (reducing emissions to combat acid rain).

26

taking actions that are economically costly.110 From the perspective of an individual country, it may be advantageous to refrain from joining the treaty; doing so may enable that country to avoid incurring costs, even as it reaps the benefits from the other signatories’ actions.111 Many other international compacts also exhibit this dynamic, such as those establishing trade sanctions,112

sovereign debt bailouts,113 contributions to international peacekeeping forces,114 or banking transparency.115

This dynamic can also arise within a country that has a federalist form of government. Consider a beneficial infrastructure project that will be located in a particular city. That city is contained within a county, which is contained within a state, which is contained with the nation. Each level of government has an interest in the project being completed. However, each has a limited amount of funds at its disposal. Thus, each governmental entity would prefer that the others fund the project. Accordingly, each government may endeavor to keep itself away from the negotiating table in order to force the other governments to commit the necessary funds.

The dynamic explored throughout this Subpart is, in many ways, the flipside of the dynamic underlying our analysis in Part III.A. Here, Bob is a dispensable party—his participation is not needed to reach the efficient outcome—just like Zelda was there. The difference between Bob and Zelda is that Zelda preferred an inefficient outcome; thus, she would not contribute cash payments to the Coasean bargain, and the other parties were happy to exclude her. Bob, in contrast, prefers the efficient outcome. Accordingly, he is theoretically willing to make cash payments to

110 See FRANK, supra note Error: Reference source not found, at 177 (“Firms don’t pollute because they derive pleasure from doing so. They do it because removing pollution costs money.”). 111 See Dixit & Olson, supra note Error: Reference source not found (illustrating this dynamic).112 Cf. Howard LaFranchi, What Do Iran Sanctions Cost You?, CHRISTIAN SCIENCE MONITOR, Apr. 5, 2012 (discussing the cost to U.S. citizens of imposing sanctions on Iran).113 See Jack Ewing, Germany Resists Europe’s Pleas to Spend More, N.Y. TIMES, Jan. 8, 2012. 114 See David Bosco, The Price of Peace, FOREIGN POLICY, May 29, 2013.115 See ORGANISATION FOR ECONOMIC COOPERATION & DEVELOPMENT, CONVENTION ON MUTUAL ASSISTANCE IN TAX MATTERS (2011), available at oecd.org/tax/exchange-of-tax-information/.

27

bring it about. But, if he can avoid coming to the table, he can avoid spending his own money and free-ride off of Alice’s contributions. Consequently, Bob wishes to exclude himself.

Combining the results of Parts III.A and III.B, we see that when there are dispensable parties, individuals will generally want to create transaction costs. If a dispensable party favors an inefficient result, the other parties will wish to create transaction costs for her.116 If a dispensable party favors an efficient result, then she will wish to create transaction costs for herself.117

The combined effect of these two dynamics is that parties will frequently have incentives to create transaction costs. As we demonstrate below, this has important consequences.

C. Voluntary Transaction Costs and Inefficiency

The possibility that parties may voluntarily create transaction costs imposes significant additional limitations on the circumstances in which the Coase Theorem’s predictions apply.118

Consequently, even when there are no structural transaction costs, voluntary transaction costs may prevent the parties from contracting their way to an efficient result. In these circumstances, legal rules may determine whether the parties reach an efficient outcome.

In our previous examples illustrating voluntary transaction costs, the parties generally reached an efficient result, regardless of the legal rule.119 In Part III.A, no matter the legal rule, excluding the dispensable parties from negotiations did not endanger the efficient result.120 Nor did the legal rule affect efficiency in Part III.B; Alice and the railroad negotiated an efficient result even

116 By the same token, the dispensable party will be willing to pay to eliminate transaction costs that keep her away from the table. 117 Similarly, the other parties to the Coasean bargain will be willing to pay to eliminate transaction costs that keep the dispensable party away from the table. 118 Shades of this point have been made in the literature. See Calabresi & Melamed, supra note Error: Reference source not found (stating that zero-transaction-cost assumption must imply full knowledge of reservation prices to address freeloading); Dixit & Olson, supra note Error: Reference source notfound (considering the free-rider dynamic in a specific context). 119 This is not quite accurate. In the farmer/conservationist example explored in Part II.A, if Zelda creates transaction costs that prevent the other parties from coming to the table, the parties do not achieve an efficient result. 120 See note Error: Reference source not found, supra, and accompanying text.

28

when Bob opted out of negotiations.121 In both instances, the key predictions of the Coase Theorem persisted. Legal rules, and the addition of new transaction costs, had only distributional effects.122

However, this is not always the case. There are instances in which creating transaction costs produces net benefits for the party creating them, but shrinks the size of the economic pie overall. Voluntary transaction costs can create inefficiencies just like any other transaction cost can.

We first illustrate this point with a scenario in which parties have incentives to impose transaction costs on themselves. We once again assume that railroad tracks run next to the wheat fields of two farmers, Alice and Bob. The tracks are parallel to each other, so that one track (the “near track”) is closer to the farmers’ fields and the other track (the “far track”) is farther away. The railroad can run trains, which emit sparks, on either or both tracks. If the trains run, the farmers must plant back from the tracks to prevent errant sparks from igniting their fields. The farmers must plant further back from the tracks, and incur greater opportunity costs from forgone wheat, if a train runs on the near track than if a train runs on the far track.

Assume that, if the railroad runs trains on both tracks, Bob and Alice must each forgo planting $1000 of wheat. If trains only run on the far track, Alice and Bob can plant most of the way to the tracks, letting each of them earn $650 more from growing wheat than they could if trains ran on both tracks. Further assume that running trains on both tracks gives the railroad $800 in benefits, and that running trains on only one track gives the railroad $400 in benefits.123 Table 1, below, summarizes these results.124

121 See note Error: Reference source not found, supra, and accompanying text. 122 See notes Error: Reference source not found and Error: Reference source notfound, supra, and accompanying text.123 Equivalently, the railroad gets $400 for each track on which it runs trains. 124 We assume that the railroad does not choose to run a train on only the near track; i.e., it is a good neighbor. But cf. Daniel B. Kelly, Strategic Spillovers, 111 COL. L. REV. 1641 (2011) (arguing that parties sometimes intentionally take actions that produce negative externalities in order to extract payments from the negatively affected individual in exchange for abating the harmful activity).

29

Table 1

Value to Alice

Value to Bob

Value to Railroad

No Trains Run $1000 $1000 $0

Trains Run on Far Track Only

$650 $650 $400

Trains Run on Both Tracks

$0 $0 $800

As in our previous example, the efficient result is for the railroad not to run any trains; doing so creates $2000 of value for the parties to divide amongst themselves.125 This is more than both the $1700 of value created by only running trains on the far track and the $800 of value created from running trains on both tracks.126

But, if Bob has the ability to impose transaction costs on himself, the result that the parties achieve will depend on the legal rule.

Consider first what happens if the farmers are legally entitled to not have sparks on their land. In this circumstance, there is no agreement to be struck: Alice and Bob would demand that the railroad pay them a combined total of at least $700 before they would allow the railroad to run trains on the far track, but the railroad would not be willing to pay more than $400 for that right.127 Similarly, Alice and Bob will only allow the railroad to run trains on both tracks if they receive a payment of at least $2000. That amount is far more than the $800 that the railroad might be willing to pay them for that right. Accordingly, the parties will achieve an efficient result. All of the surplus will accrue to Bob and Alice, and Bob will not choose to create transaction costs.128

125 This can be seen by adding the values in the top row of Table 1. Equivalently, this is the sum of the $1000 that Alice and Bob each receive from planting additional wheat and the $0 value that the railroad receives from not running any trains. 126 See Table 1, supra. 127 See Table 1, supra.

30

Next, consider what happens if the law gives the railroad the right to run trains on the far track only. It is in Alice and Bob’s joint best interests to induce the railroad not to run any trains; doing so will enable them to earn $700 more in profits.129 Not running the train on the far track costs the railroad $400. Thus, there is the potential for the farmers and the railroad to make an agreement that leaves each party better off.

But this agreement is only possible if both Alice and Bob participate in negotiations. The railroad reaps a $400 benefit from running the train on the far track, so it will not agree to cease doing so unless it receives a cash payment of at least $400.130 Each individual farmer reaps only $350 in benefits from the train not running; accordingly, neither individual farmer will be willing to pay the railroad enough to induce the railroad to stop running the train.

Thus, if Bob creates transaction costs, Alice and the railroad will not strike any further agreements. The railroad will run the train on the far track, and Bob will get $650.

On the other hand, if Bob chooses to participate in the negotiations, he and Alice can jointly induce the railroad not to run any trains, creating additional surplus. Assuming, as previously,131

that the parties divide this surplus evenly among themselves, Alice and Bob will each pay the railroad $250 not to run the train.132 In this scenario, Bob pays out $250 to the railroad, but gets to plant an additional $350 worth of wheat, leaving him with $750 of value.133 As this is more than the $650 of value that he receives when he does not participate in negotiations, Bob will prefer to participate.

128 More precisely, if Bob can create transaction costs at no cost to himself, he will be indifferent between creating them and not. In either event, there is no deal to be struck even in the absence of transaction costs, so the effect is the same. 129 See Table 1, supra.130 See Table 1, supra.131 See text accompanying notes Error: Reference source not found and Error:Reference source not found, supra. 132 Each of Alice and Bob pays $250 for a $350 benefit. The railroad loses a $400 benefit but receives $500 in payments. 133 See Table 1, supra.

31

Thus, when the law gives the railroad the right to run trains on the far track only, Bob will choose not to create transaction costs and will instead choose to join Alice in negotiating with the railroad. In fact, Bob might now be willing to pay up to $300 to eliminate transaction costs that keep him away from the negotiating table.134 This legal rule will result in the railroad not running the train, the efficient result.

Finally, suppose that the law permits the railroad to run trains on both tracks. Under this rule, the parties will never reach an agreement under which the railroad does not run any trains.135

Striking a Coasean bargain in which the farmers pay the railroad not to run any trains creates $1200 of value.136 As discussed above, this agreement is only possible if all three parties come to the table.137 Since we assume that the parties evenly divide up the surplus that their bargaining creates, Bob and Alice will each net $400,138 as will the railroad.139

Alternatively, if Bob faces high transaction costs that prevent him from participating in the agreement, Alice will be left to negotiate with the railroad on her own. In that case, Alice will pay the railroad not to run trains on the near track, as that lets Alice plant an additional $650 of wheat but only costs the railroad $400.140 But Alice will not be willing to pay the railroad not to run any trains: Such an agreement would allow Alice to plant $350 more wheat than if trains only run on the far track. However, she must pay the railroad an additional $400 not to run any trains. Thus, if Bob is removed from the negotiating table, Alice and the

134 There is potentially $300 of surplus to be gained from having the railroad run no trains instead of running trains on the far track only. See Table 1, supra. As shown below, this surplus will not be realized unless Bob participates in the agreement. Thus, Bob might potentially be willing to pay up to $300 to be involved in the agreement. 135 More specifically, the parties will never reach such an agreement if Bob is rational and Alice and the railroad play subgame-perfect strategies. See Companion Paper, supra note Error: Reference source not found, at § 2.1.136 Without a bargain, the railroad will run trains on both tracks, creating $800 of value. With a bargain, no trains will run, producing $2000 of value, an increase of $1200. See Table 1, supra. 137 See text accompanying notes Error: Reference source not found-Error:Reference source not found, supra. 138 This corresponds to each paying the railroad $600 not to run the train. 139 The railroad will receive $1200 and will forgo the $800 benefit it would receive from running its train, leaving it $400 better off as well.140 See Table 1, supra.

32

railroad will reach an agreement under which the railroad will run trains on the far track only.

Accordingly, if Bob faces prohibitive transaction costs that prevent him from participating in an agreement, he will be able to grow $650 worth of wheat. This is less than the $1000 worth of wheat that he is able to grow when he participates in the agreement. However, when he does not participate in the agreement, Bob does not have to make any payments to the railroad. This means that his $650 in additional wheat is all profit, making Bob $250 better off than he is when he teams up with Alice to negotiate with the railroad. Bob would therefore be willing to pay up to $250 to create transaction costs—even though his leaving the table costs Alice and the railroad a combined total of $550 and leads to an inefficient outcome.141

Table 2 summarizes these results. When voluntary transaction costs are possible, the legal rule affects Bob’s choice and the efficiency of the final outcome. Thus, when there are voluntary transaction costs, the choice of legal rule becomes crucial; the Coase Theorem and its underlying logic simply do not apply.

141 As noted above, if Bob comes to the table, each party gets $400 of the gains created by the Coasean bargain. Thus, Alice and the railroad receive $800 surplus between them. If Bob leaves the table, Alice and the railroad will split $250 in surplus between them. Accordingly, Bob’s leaving the table costs Alice and the railroad a combined total of $550, more than twice the amount that Bob gains by opting out of the agreement.

33

Table 2

Railroad’s Legal Right

Voluntary Transaction

Costs?

UltimateResult

Efficient Result?

Can Run Trains on

Both Tracks

Bob Will Pay to Create

Train Runs on Far Track

No

Can Run Trains on Far Track Only

Bob Will Pay to Remove

No Trains Run Yes

Cannot Run Trains on

Either Track

Bob Is Indifferent142 No Trains Run Yes

We emphasize that this result—that, when the law gives the railroad the right to run trains on both tracks, it is not in each party’s best interests to come to the table and strike an efficient Coasean bargain—holds no matter how the parties divide up the gains from their agreement.

To see this, suppose that both of the farmers and the railroad each has the ability to create transaction costs that keep themselves away from the negotiating table. In any scenario in which all three parties come to the table and reach an efficient agreement, the farmers must compensate the railroad for the value it loses by not running any trains. That means the railroad must receive cash payments from the farmers totaling at least $800.143

This means that at least one of the farmers must make a cash payment of $400 or more;144 for simplicity, assume that that farmer is Bob.145 The parties’ Coasean bargain enables Bob to plant an additional $1000 of wheat, and requires him to make a payment of

142 See discussion in footnote Error: Reference source not found, supra.143 See Table 1, supra.144 If both farmers were making cash payments of less than $400, then the sum of both farmers’ payments would be less than $800. A payment of that size is not large enough to induce the railroad not to run any trains. 145 This assumption produces no loss of generality; Bob and Alice are identically situated, so the analysis is the same if Alice pays at least $400 instead of Bob.

34

at least $400. Thus, Bob gets, at most, $600 in net value from the parties’ Coasean bargain.146

Consider instead what happens if Bob does not come to the negotiating table. Without Bob, Alice will still want to negotiate with the railroad: running trains on the near track costs Alice $650 in forgone wheat, but only gives the railroad $400 in benefits. It is therefore in both Alice’s and the railroad’s interests to make an agreement under which Alice pays the railroad between $400 and $650 not to run any trains on the near track. Because the near track runs next to both Alice’s farm and Bob’s farm, this agreement will also allow Bob to plant $650 of additional wheat. This is less than the $1000 in wheat that Bob can plant when no trains run. However, this $650 is all profit for Bob; he does not need to make any cash payments to secure it. Thus, Bob is better off creating transaction costs that keep him away from the table.

Accordingly, it is never in all parties’ best interests to come to the negotiating table. But, when the law gives the railroad the right to run trains on both tracks, the parties can only achieve an efficient result when they all participate in negotiations. Thus, when parties have the ability to create large transaction costs for themselves, the parties will never achieve an efficient result under this legal regime.147

This Subpart’s analysis has been built around scenarios in which individuals want to keep themselves away from the negotiating table. Similar dynamics apply to scenarios in which individuals want to keep other parties away from the negotiating table.

For example, consider again the farmer/conservationist scenario described in Part III.A, supra. Once again, we assume that running the train creates benefits of $500 for the railroad and $200 for conservationist Zelda, and creates costs of $1000 for farmer Alice. Further suppose that there is a hill on the edge of Zelda’s property that is prone to occasional, small mudslides. When these mudslides happen, mud lands on farmer Alice’s property and destroys $10 worth of wheat. If she chooses, Zelda

146 Bob gets $1000 of extra wheat and must pay at least $400; thus, he nets at most $600. 147 This assumes that the parties have an injunctive remedy. See Part IV.B, infra.

35

can stop these mudslides, thereby ending the destruction of Alice’s wheat, at a cost of $5.148

If Zelda is excluded from negotiations, she will not stop the mudslides, which is inefficient. However, the potential value lost to this inefficiency ($5) is smaller than the cash payment that Zelda receives from the other parties if she participates in the Coasean bargain (≈$300).149 Accordingly, it is still in Alice’s and the railroad’s interests to exclude Zelda from negotiations, even though doing so results in inefficiency.

Both the two-farmer and the farmer/conservationist examples described above share an important feature. In each instance, there are three possible outcomes: The legal default outcome and two alternatives. The legal default is the least efficient outcome;150 both alternatives—the “efficient outcome”151

and the “intermediate outcome”152—are more efficient than the legal default.153 Moreover, the efficiency gains from moving from the legal default to the intermediate outcome are larger than the incremental gains from moving from the intermediate outcome to the efficient outcome.154 Thus, creating transaction costs will often benefit a party if doing so helps her to secure a larger share of the gains produced by moving to the intermediate outcome—even if

148 For example, Zelda could build a wall to reinforce the hill or catch the mud before it lands on Alice’s property. 149 Fixing the mudslide problem saves $10 of wheat at a cost of $5, producing $5 in net gains. Zelda receives a $300 payment when the parties split the gains from their agreement; see note Error: Reference source not found, supra, and accompanying text. 150 In the two-farmer example, the legal default is trains running on both tracks. In the farmer/conservationist example, the default is running the train and not fixing the mudslide-prone hill. 151 In the two-farmer example, the efficient outcome is no trains running. In the farmer/conservationist example, the efficient outcome is no trains running and fixing the mudslide-prone hill. 152 In the two-farmer example, the intermediate outcome is trains running on the far track. In the farmer/conservationist example, the efficient outcome is no trains running and not fixing the mudslide-prone hill. 153 See Table 1 and footnote Error: Reference source not found and accompanying text, supra.154 In the two-farmer example, moving from the legal default to the intermediate outcome produces $900 of value; moving from the intermediate outcome to the efficient outcome produces $300 of value. See Table 1, supra. In the farmer/conservationist example, moving from the legal default to the intermediate outcome produces $300 of value; moving from the intermediate outcome to the efficient outcome produces $5 of value.

36

those transaction costs prevents an efficient Coasean bargain.155 In such circumstances, the parties achieve most of the potential social gains (the intermediate outcome) but fall short of the full social gains available (the efficient outcome).156

For our purposes, the key point is that these circumstances are likely to arise with great frequency: Parties often have multiple ways in which they can accomplish a particular goal, and they generally select the most cost-effective options first.157 Thus, the dynamic described above—in which the efficiency gained by moving to an intermediate outcome exceeds the incremental efficiency gained by moving from an intermediate outcome to the efficient outcome—is likely to occur frequently. In other words, the first steps toward the efficient result will often produce the largest net benefits.158 Accordingly, parties may often wish to create voluntary transaction costs.

For example, voluntary transaction costs play a critical role in the behavior of creditors when a borrower suffers a financial decline. This topic has been the subject of considerable recent attention from scholars, practitioners, and journalists.159

155 See notes Error: Reference source not found and Error: Reference source notfound, supra, and accompanying text. 156 This dynamic is distinct from the problem of the empty core. See, e.g., Yakar Kannai, "The Core and Balancedness", in 1 HANDBOOK OF GAME THEORY WITH ECONOMIC APPLICATIONS, 355–395 (Robert J. Aumann & Sergiu Hart, eds., 1992); Olga N. Bondareva, Some Applications of Linear Programming Methods to the Theory of Cooperative Games, 10 PROBLEMY KYBERNETIKI 119 (1963); Lloyd S. Shapley, On Balanced Sets and Cores, 14 NAVAL RESEARCH LOGISTICS Q. 453 (1967); see also Varouj A. Aivazian & Jeffrey L. Callen, The Coase Theorem and the Empty Core, 24 J.L. & ECON. 175 (1981); Ronald H. Coase, The Coase Theorem and the Empty Core: A Comment, 24 J.L. & ECON. 183 (1981); Varouj A. Aivazian & Jeffrey L. Callen, The Core, Transaction Costs, and the Coase Theorem, CONST’L POL. ECON. (2003).157 This principle is sometimes called the low-hanging-fruit principle. The analogy is to fruit-pickers harvesting the most easily accessible fruit first before going after fruit that is harder to access. See ROBERT H. FRANK & BEN S. BERNANKE, PRINCIPLES OF MICROECONOMICS 46-47 (2d ed. 2004).158 Formally, this will not always be the case. There are circumstances in which this dynamic may not hold, such as when the returns from an activity are convex. However, it is generally believed (and assumed) that such scenarios are relatively rare. See id.159 See, e.g., Henry T.C. Hu & Bernard Black, Debt, Equity and Hybrid Decoupling: Governance and Systemic Risk Implications, 14 EUR. FIN. MGMT 663 (2008); Theresa Einhorn, 28th Annual Review of Developments in Business Financing: Current Trends in the Corporate Debt Market and Credit Derivatives, 2010 AM. BAR ASSOC. SEC. BUS. L. 1, available at haynesboone.com/files/Publication/; Daniel Gross, The Scary Rise of the

37

Historically, when homeowners or profitable businesses have found themselves overextended and unable to pay their debts, they have negotiated agreements with their creditors.160 These agreements generally provide for the debt to be repaid over a longer period of time than originally contemplated,161 and they often reduce the total amount of debt outstanding.162 For their part, creditors generally agree to receive less than they are legally entitled to receive because, by doing so, they ultimately receive more money than they would if they strictly enforced the original terms of their loans: In the case of the homeowner, she would likely abandon the house, leaving the creditor with an asset she would have to liquidate, which could be a difficult, costly, and lengthy process.163 With respect to the business, if the creditors stand on their rights they will force a profitable company out of business, ultimately resulting in the creditors receiving less money than if they had agreed to a reduction or a restructuring of debt.

These scenarios are classic Coasean bargains: The creditors have a legal right to certain payments at certain times. The creditors and the borrower come to the table and work out an agreement. Pursuant to that agreement, the creditors relinquish some of their legal rights. This produces significant gains for the

“Empty Creditor”, MONEYBOX, SLATE.COM, Apr. 21, 2009; Dante Altieri Marinucci, Empty-Creditor Syndrome and Vivisepulture: Preventing Credit-Default-Swap Holders from Pushing Companies into Premature Graves by Refusing to Negotiate Restructurings, Comment, 62 CASE W.L. REV. 1285, (2012); András Danas, Do Empty Creditors Matter? Evidence from Distressed Exchange Offers, wpweb2.tepper.cmu.edu/wfa/wfasecure/upload/ 2012_PA_952361_202294_910787.pdf; SUMIT AGARWAL ET AL., FED. RES. BANK OF CHICAGO, THE ROLE OF SECURITIZATION IN MORTGAGE RENEGOTIATION, Working Paper 2011-02, available at chicagofed.org/ digital_assets/publications/working_papers/2011/wp2011_02.pdf; Tomasz Piskorski et al., Securitization and Distressed Loan Renegotiation: Evidence from the Subprime Mortgage Crisis, 97 J. FIN. ECON. 369 (2010); Eric A. Posner & Luigi Zingales, A Loan Modification Approach to the Housing Crisis, 11 AM. L. & ECON. REV. 575 (2009); Christopher Mayer, Housing, Subprime Mortgages, and Securitization: How Did We Go Wrong and What Can We Learn So This Doesn’t Happen Again?, Testimony in front of the U.S. Financial Crisis Inquiry Commission (2010); Samuel Kruger, The Effect of Mortgage Securitization on Foreclosure and Modification, available at scholar.harvard.edu/skruger/ions/effect-mortgage-securitization-foreclosure-and-modification (last visited Feb. 24, 2014). 160 See BETHANY MCLEAN & JOE NOCERA, ALL THE DEVILS ARE HERE: THE HIDDEN HISTORY OF THE FINANCIAL CRISIS 308 (2010)161 Id.162 Id.163 Jacoby & Janger, supra note Error: Reference source not found, at 894.

38

borrower that exceed the losses to the creditors.164 The agreement grows the size of the economic pie, ultimately allowing the creditors to receive more money than they would have received if they insisted on enforcing the letter of their agreements.165

That said, it is not necessary for every creditor to relinquish some of its rights; it is only necessary that the buyer’s debt be restructured in a way that allows it to stay in business and keep earning profits. An individual creditor receives the most money when she refrains from restructuring the debt that she holds while the other creditors restructure theirs. Thus, if a creditor can prevent herself from coming to the table, she can avoid having to make concessions, benefiting herself at the expense of the debtor and the other creditors. In recent years, creditors’ increasing use of two particular investment techniques—both of which have the effect of keeping creditors away from the negotiating table when the borrower wants to restructure its debt—has drawn a considerable amount of attention.

First, creditors have increasingly purchased large amounts of credit default swaps on the debt that they are owed. 166 Credit default swaps resemble insurance policies; the holder of a credit default swap is entitled to receive a payment from a third party if the borrower in question does not meet all of its obligations under a particular debt instrument.167

These credit default swaps do not change the value of arranging a loan modification for the distressed debtor; they

164 See id. at 892-93.165 See id.166 In several recent instances, creditors have acquired credit default swaps that were as large as the debt they held or even larger. See, e.g., Einhorn, supra note Error: Reference source not found (discussing similar situations involving General Motors, YRC Worldwide, and Gannett Co.); Martin Hutchinson, Ban Credit Default Swaps? These Corporate Bankruptcies Show We Should, MONEY MORNING, Apr. 23, 2009, moneymorning.com/2009/04/23/ban-credit-default-swaps/ (discussing similar situations at AbitibiBowater Inc. and General Growth Properties Inc.); Adam Reiser, An Economic Analysis and Legal Framework for Credit Default Swap Regulation, 14 N.C. BANKING INST. 101, 120-21 (2009), available at law.unc.edu/documents/journals/ncbank/balancesheet/ (discussing a similar situation involving Tower Automotive); Loren Steffy, Some Bondholders Bank on Bankruptcy, HOUSTON CHRON., Jul. 21, 2009, chron.com/business/steffy/ article/Some-bondholders-bank-on-bankruptcy-1622890.php (discussing similar situations involving Six Flags). 167 See FRANK PARTNOY, INFECTIOUS GREED 372-73 (2009).

39

merely change which parties reap gains and bears costs from a modification.168 However, they do increase the number of parties who have a stake in the debtor’s solvency; by entering into a credit default swap, the creditor essentially distributes part of her interest in the debt to her counterparty.169 This means that more parties must be brought to the negotiating table to resolve the issues pertaining to that creditor’s debt.170 In other words, by entering into credit default swaps, the creditor can raise the transaction costs of modifying the debt that she holds.171 This can make it hard for her to come to the table, leaving the other creditors to make larger sacrifices.172

Second, creditors have increasingly divided their interests among a larger number of parties, making it more difficult and unwieldy for them to participate in a deal. The most visible example of this in recent years is the securitization of residential mortgages.173 Historically, these mortgages were held by the bank that made the loan.174 If a homeowner became unable to pay, she

168 Every payment that the creditor receives from her counterparty is an equal benefit to the creditor and cost to the counterparty, and vice versa. 169 Taken to its extreme, a creditor can divest herself of any economic interest in a debtor. Such creditors are sometimes referred to as empty creditors, because they retain legal rights to payments from the debtor but have emptied out their economic interest. See, e.g., Hu & Black, supra note Error: Reference sourcenot found, at 709; Gross, supra note Error: Reference source not found; see also Jordan M. Barry et al., On Derivatives Markets and Social Welfare: A Theory of Empty Voting and Hidden Ownership, 99 VA. L. REV. 1103 (2013) (discussing decoupling strategies more generally).170 For example, suppose that a creditor has a claim to $1 million in payments from the borrower, then enters into credit default swaps on $100,000 of that debt with nine counterparties. Before the creditor entered into those transactions, assembling all of the people with an interest in that $1,000,000 payment being made simply meant bringing the creditor to the table. After those transactions, accomplishing the same result required bringing in the creditor and all nine counterparties. 171 This is especially true when she enters into credit default swaps with multiple different counterparties. 172 To be sure, there are other reasons why creditors might wish to purchase credit default swaps on the debt they owe. For example, there are regulatory and risk-management benefits. See PARTNOY, supra note Error: Reference sourcenot found, at 374-77; IVO WELCH, CORPORATE FINANCE 129-30 (2d ed. 2011). Credit default swaps can also reduce the frequency of socially inefficient strategic defaults by borrowers. See Patrick Bolton & Martin Oehmke, Credit Default Swaps and the Empty Creditor Problem, 24 REV. FIN. STUD. 2617 (2011). 173 See MCLEAN & NOCERA, supra note Error: Reference source not found, at 308.174 Id.

40

could go to that bank directly and try to negotiate a modification.175

Securitized mortgages, in contrast, are packaged together in large quantities, made into bonds or other securities, and sold off to various investors in pieces.176 Often, these various securities have differing cash flow rights.177 Both the number of holders and their varied interests substantially complicate restructuring negotiations.178 This encourages individuals to restructure their other debts instead of their mortgages, which can redound to the benefit of the mortgage lender.179

Staying away from debt renegotiations is often good for individual lenders, but it is quite costly for society more generally.180 When it is unclear whether a business will be able to continue operations, employees—especially those with the best outside options—start to leave.181 Customers may abandon a company if they are worried about its ability to support a product over the long-term.182 Suppliers demand higher prices and greater contractual protections.183 All of these effects can produce a downward spiral that can sink an otherwise viable business.184

Thus, it is better to renegotiate a company’s outstanding debts sooner rather than later.185 Yet in a number of large and high-

175 Id.176 See GERALD P. DWYER, FED. RES. BANK OF ATLANTA, THE FINANCIAL SYSTEM AFTER THE CRISIS: STRUCTURED FINANCE AND CREDIT RATING AGENCIES (2010), available at frb.atlanta.org. 177 Many mortgage-backed securities were broken into tranches. Payments on the mortgages underlying the securities went to the holders of the most senior tranches first. Only once those investors were paid in full would the holders of the next-most-senior tranch receive any money, and so on and so forth. Id.178 See Piskorski et al., supra note Error: Reference source not found, at 370-73. 179 For example, approximately 37 million Americans currently have outstanding student loan debt, and over 50 million U.S. households have outstanding credit card debt. See AM. STUDENT ASSISTANCE, STUDENT LOAN DEBT STATISTICS, available at asa.org/policy/resources/stats/, last visited Feb. 24, 2014; Tim Chen, American Household Credit Card Debt Statistics: 2013, NERDWALLET.COM, last visited Feb. 24, 2014. 180 See In re Gen. Motors Corp., 407 B.R, 463, 493 (Bankr. S.D.N.Y. 2009) (“This is hardly the first time that this Court has seen creditors risk doomsday consequences to increase their incremental recoveries . . . .”).181 See Donald C. Hambrick & Richard A. D’Aveni, Top Team Deterioration As Part of the Downward Spiral of Large Corporate Bankruptcies, 38 MGT. SCI. 1445 (1992).182 Jacoby & Janger, supra note Error: Reference source not found, at 894.183 Id.184 Hambrick & D’Aveni, supra note Error: Reference source not found.185 See, e.g., Gross, supra note Error: Reference source not found.

41

profile examples—including AIG,186 Chrysler,187 General Motors,188 Six Flags,189 and others190—the distribution of the economic interest in a sizable bloc of outstanding debt significantly complicated and lengthened the negotiating process and cast the outcome of that process into doubt.191

Similarly, renegotiating mortgages in the wake of an unexpected economic downturn can create significant value for both borrowers and lenders.192 Foreclosures can also harm other homeowners in the same neighborhood by lowering housing values.193 Moreover, declining housing values can further fuel the economic downturn.194 Studies have found that securitized loans 186 See, e.g., id.; Marinucci, supra note Error: Reference source not found, at 1301-02. 187 See Neil King Jr. & Jeffrey McCracken, Chrysler Chapter 11 Is Imminent, WALL ST. J., Apr. 30, 2009, online.wasj.com/article/ SB124102375931669205.html. 188 Einhorn, supra note Error: Reference source not found, at 19. 189 See Credit Default Swaps and Bankruptcy: No Empty Threat, PAYMENT-TIMES BLOG, payment-times.com/wordpress/2009/06/credit-default-swaps-and-bankruptcy-no-empty-threat/ (last visited Aug. 18, 2013); Gross, supra note Error: Reference source not found; Michael S. Rosenwald, Plagued by Death, Six Flags Faces Its Own Wild Ride, WASH. POST, Apr. 13, 2009, washingtonpost.com/wp-dyn/content/article/2009/04/12/AR2009041202152.html; Steffy, supra note Error: Reference source not found.190 See, e.g., Einhorn, supra note Error: Reference source not found (discussing YRC Worldwide and Gannett); Henny Sender, Hedge-Fund Lending to Distressed Firms Makes for Gray Rules and Rough Play, WALL ST. J., Jul. 18, 2005, online.wsj.com/article/1,,SB112164567837987920,00.html (discussing the case of Tower Automotive); Danas, supra note Error: Reference source notfound (“Anecdotal evidence suggests that corporations such as AbitibiBotwater, General Motors, Harrah's Entertainment, McClatchy, and Unisys had difficulties in reducing their debt because of CDS [credit default swap] holders.”). 191 See Danas, supra note Error: Reference source not found (finding that the creditors of companies with credit default swaps on their debt are more than 50% less likely to participate in voluntary restructurings). But see Mascia Bedendo et al., In- and Out-of-Court Debt Restructuring in the Presence of Credit Default Swaps, www2.warwick.ac.uk/fac/soc/wbs/subjects/finance/fof2012/programme/empty_cds.pdf (not finding any effect of credit default swaps on voluntary restructurings). 192 See Patrick Bolton & Howard Rosenthal, Political Intervention in Debt Contracts, 110 J. POL. ECON. 1103 (2002); Randall S. Kroszner, Is It Better to Forgive Than to Receive? An Empirical Analysis of the Impact of Debt Repudiation (2003); Tomasz Piskorski & Alexei Tchistyi, Stochastic House Appreciation and Optimal Mortgage Lending (2008); see also Piskorski et al., supra note Error: Reference source not found (describing the evidence from these studies as “compelling”). 193 Danas, supra note Error: Reference source not found; Posner & Zingales, supra note Error: Reference source not found. 194 See Danas, supra note Error: Reference source not found.

42

are significantly more likely to be foreclosed upon than similar loans held by a single lender,195 and some have argued that the difficulties of renegotiating securitized loans significantly exacerbated the 2007-2008 financial crisis and the great recession.196

Thus voluntary transaction costs, and the efficiency problems that they can create, are of great real-world importance. We now turn to some of the implications that voluntary transaction costs have for the selection and design of legal rules.

IV. IMPLICATIONS FOR CHOOSING LEGAL RULES

As Part III demonstrated, when parties have the ability to create transaction costs, the Coase Theorem generally will not apply.197 Thus, we cannot necessarily expect that parties will reach an efficient result on their own.198 Accordingly, different legal rules can produce very different outcomes. Having established the importance of legal rules when voluntary transaction costs are possible, we now explore the effects of different legal rules on the parties’ negotiations in greater detail. We then consider the effects

195 See AGARWAL ET AL., supra note Error: Reference source not found (finding that bank-held loans are 26-36% more likely to be renegotiated than comparable securitized mortgages); Piskorski et al., supra note Error: Reference source notfound, at 374 (finding that seriously delinquent securitized loans are between 13% and 32% more likely to be foreclosed upon than seriously delinquent loans held by banks); Kruger, supra note Error: Reference source not found, at 1 (finding that securitization increases the probability of foreclosure by 35% and decreases the probability of loan modification by 69%); see also Posner & Zingales, supra note Error: Reference source not found (arguing that unnecessary foreclosures in the wake of the financial crisis could destroy hundreds of billions of dollars of value); Mayer, supra note Error: Referencesource not found. But see MANUEL ADELINO ET AL., WHY DON’T LENDERS RENEGOTIATE MORE HOME MORTGAGES? REDEFAULTS, SELF-CURES, AND SECURITIZATION, FED. RES. BANK OF BOSTON PUBLIC POLICY DISCUSSION PAPER NO. 09-4 (arguing that securitization does not reduce renegotiations); MANUEL ADELINO ET AL., RENEGOTIATING HOME MORTGAGES: EVIDENCE FROM THE SUBPRIME CRISIS, FED. RES. BANK OF BOSTON PUBLIC POLICY DISCUSSION PAPER NO. 09-4 (similar); FOOTE ET AL., REDUCING FORECLOSURES: NO EASY ANSWERS, NBER MACROECONOMICS ANNUAL 24, 89 (2009) (similar). 196 See, e.g., Edward L. Glaeser, Debating the Securitization of Mortgages, ECONOMIX, NYTIMES.COM, July 27, 2010 (providing an overview of the debate on this issue); Kruger, supra note Error: Reference source not found, at 1-2 (finding that securitization led to 500,000 additional foreclosures during the financial crisis). 197 See Part III.C, supra.198 See Coase, supra note Error: Reference source not found.

43

of the remedies that the law provides when those rules are violated.199

A. Structuring Legal Rights

We begin by considering who legal rules should protect, and to what degree.200 The answers to both questions can significantly affect whether the parties achieve an efficient result.

1. Legal Rules That Always Produce Efficiency

As a preliminary matter, when policymakers know that a particular course of action for each party will produce an efficient result, they can easily guarantee an efficient outcome: They can enact a legal rule that requires each party to take that efficiency-creating course of action. We term such a rule a “fully specified efficient rule.”201 A fully specified efficient rule always guarantees that the parties will achieve an efficient outcome.202

To understand the logic behind this result, suppose that a party—let’s call her Alice—is considering deviating from her legally mandated action. Suppose that the new action Alice is contemplating produces an inefficient outcome.203 Because the new outcome is inefficient, Alice’s action must make someone worse off.204 Each of the parties made worse off by Alice’s new behavior—the “damaged parties”—has the legal right to insist that Alice fulfill her legal obligations. To keep each of the damaged parties from standing on his rights and blocking her new behavior, Alice must pay each damaged party enough to fully compensate

199 We explore these issues through examples like that in Part III.B, supra, in which actors have incentives to keep themselves away from the table. Similar examples can be constructed for instances in which parties want to keep other individuals away from the table. 200 In doing so, we supplement extensive literature on this topic. See, e.g., Krier & Schwab, supra note Error: Reference source not found, at 447-49 (discussing the literature as of 1997). 201 Cf. A. MITCHELL POLINSKY, AN INTRODUCTION TO LAW AND ECONOMICS 31-33 (4th ed. 2011) (discussing fully specified contracts).202 A fully specified efficient rule also has other benefits, such as avoiding any transaction costs that would otherwise be incurred in bringing about an efficient result. See Coase, supra note Error: Reference source not found, at 19.203 If the new action does not produce an inefficient result, then the parties still achieve an efficient outcome and there is no issue. 204 If no one is worse off, then the combined utility of the parties must have grown or stayed the same, in which case the new outcome is not inefficient.

44

him for the losses that her new behavior inflicts upon him. However, because Alice’s legally mandated action produces an efficient outcome, moving to an inefficient outcome shrinks the economic pie. Thus, the aggregate gains to Alice and any parties who benefit from her new action will be smaller than the damaged parties’ aggregate losses. Accordingly, the beneficiaries of Alice’s new action will not be willing to pay the damaged parties enough to fully compensate them, and at least one damaged party will always choose to stop Alice from changing her behavior.

Note also that, when there is a fully specified efficient rule, it is not worthwhile to keep the “opposition”205 away from the table. If actor Alice wishes to change her own behavior, she must secure the consent of the opposition in order to do so, and this can only be done through a negotiated agreement. Conversely, if the opposition is considering a change in its behavior, it must secure Alice’s consent. Alice will only give her consent if the opposition fully compensates her for any losses she would suffer. Thus, she has no interest in keeping the opposition away from the negotiating table.

While the fully specified efficient rule is a useful theoretical construct, it is unlikely that policymakers could enact many such rules in practice. To implement a fully specified efficient rule, policymakers must be able to determine the efficient outcome in advance, as well as a behavior by each party that brings about that outcome.206 Acquiring and processing the necessary information to make these determinations will often prove prohibitively difficult.

In certain circumstances, there are more practicable legal rules that also promise efficiency. One subset in particular is worth highlighting. Sometimes, an activity benefits only one individual but harms many other parties.207 We term this a “one-

205 See footnote Error: Reference source not found, supra, and accompanying text. 206 Moreover, the underlying economic facts can change over time, making this task even more difficult.207 Equivalently, an activity can harm only one individual but benefit many others. For example, consider a factory that produces pollution that harms many surrounding homeowners. If the factory operates, that harms many shareholders but benefits only the factory (ignoring the issue of public shareholders, employees, and so forth). If the behavior is framed as the right to cease the factory’s production, that harms only the factory and benefits many

45

against-many” scenario. The classic one-against-many scenario involves a factory that produces pollution that harms many homeowners.208 Both of our recurring examples are one-against-many scenarios: In the farmer/conservationist example, not running the train benefited only the farmer, and hurt both the railroad and the conservationist.209 Similarly, in the two-farmer example, running the train benefitted only the railroad, and hurt both farmers.210

When there is a one-against-many scenario, a legal rule that enables each member of the many to stop the one from engaging in activity (a “many-privileging rule”) ensures that none of the parties wishes to create transaction costs. By discouraging voluntary transaction costs, many-privileging rules can encourage efficiency. To appreciate why this is so, it is helpful to examine this dynamic through the lens of our recurring two-farmer example.

Assume that the law imposes a many-privileging rule—that is, it gives farmers the right to keep sparks off of their lands. First, consider the railroad’s incentives. It wishes to run trains on the railroad track. Every individual farmer has the right to stop the railroad from doing so. The railroad must therefore reach an agreement with every farmer if it wishes to run trains. That means the railroad will not want to keep any farmers away from the table. Nor does the railroad wish to keep itself away from the table. It is the only party who is willing to pay the farmers to allow trains to run; if it does not arrange a deal with the farmers, there will be no deal.211 Therefore, the railroad does not wish to create transaction costs.

Next, consider the farmers’ incentives. Each farmer has the power to prevent the trains from running if she so chooses. Each farmer can refuse any deal that is not in her interest; accordingly, if the trains run, each farmer must have struck a deal that made her better off than she would have been if no trains ran. That means homeowners. But both scenarios are exactly identical. See FRANK, supra note ERROR: REFERENCE SOURCE NOT FOUND, at 90-95.208 For a modern spin, see Beggs, supra note Error: Reference source not found (discussing a wind farm whose turbines produce noise that bothers nearby households).209 See notes Error: Reference source not found-Error: Reference source notfound, supra, and accompanying text. 210 See Table 1, supra. 211 In other words, there is no one that the railroad can hope to free-ride off of.

46

that each farmer has no reason to keep herself away from the table.212 Moreover, each farmer can only strike a deal with the railroad if the railroad and all the other farmers come to the table as well.213 Consequently, neither farmer wishes to create transaction costs that keep the other farmer away from the table, either. This analysis applies with the same force if there are two farmers, three farmers, or a hundred farmers.214

In a one-against-many scenario, imposing a many-privileging legal rule guarantees that no party will want to create transaction costs. It is worth noting that we established this result without making any assumptions about which outcome was efficient; a many-privileging legal rule encourages the parties to reach an efficient result regardless of which outcome is efficient.

That said, we must note that many-privileging legal rules also have a serious drawback. A many-privileging legal rule can create the perfect conditions for a hold-out problem, in which each member of the many attempts to hold out on giving his consent in order to extract as much of the gains from the efficient bargain as possible.215 This sort of strategic behavior—often considered a transaction cost216—can prevent the parties from reaching an efficient Coasean bargain. This suggests a trade-off between voluntary and structural transaction costs in one-against-many scenarios.

2. The Value of a More Efficient Starting Point

The Coase Theorem recognizes that legal entitlements are not final judgments. They are merely default rules that the parties can alter through private agreement if they so choose. Legal entitlements are not the last word, but instead the beginning of the

212 But see notes Error: Reference source not foundError: Reference source notfound, infra, and accompanying text. 213 It is clear why the railroad must come to the table for the farmer to strike a deal with it. Moreover, the railroad only wants to strike a deal if doing so allows the railroad to run the train. This can only happen if the railroad can strike a deal with every farmer, which is only possible if all farmers come to the table. 214 Cf. William Yardley, Turbines Too Loud? Here, Take $5,000, N.Y. TIMES, July 31, 2010 (discussing how a wind farm is paying local residents $5,000 to waive their right to complain about noise produced by the wind turbines).215 We note that this is not a problem if protected parties are only entitled to a liability remedy instead of an injunctive remedy. See Part IV.B.2, infra. 216 See, e.g., Fennell, supra note Error: Reference source not found, at 1486-87 (discussing the classification issue).

47

conversation.217 When voluntary transaction costs are possible, the starting point for negotiations can be extremely important. The smaller the gap between the parties’ starting point and the efficient outcome, the better the parties’ prospects are of traversing that gap.

To see the intuition behind this result, recall our prior discussion of when parties chose to create transaction costs, even though doing so produced inefficient outcomes.218 In the examples we examined, the first steps toward the efficient result produced the largest net benefits. This is likely to be a common scenario; the worse one’s starting point, the easier it is to make large improvements.219 At the same time, larger bargains are likely to entail larger transfer payments, which increases the incentives for a potential payer to create transaction costs to keep herself away from the table. Even if those transaction costs prevent an efficient bargain, the other parties who come to the table are still likely to strike some deal—and that deal will produce large benefits without a cash payment from the absent party.

It is helpful to illustrate these points by modifying our familiar two-farmer example. Once again, assume that there are two wheat farmers, Alice and Bob, whose farms abut parallel railroad tracks, and who must leave ground fallow next to the tracks to prevent sparks from passing trains from lighting their fields on fire. In this example, however, we assume that there are four railroad tracks, which we will number 1 to 4, with track 4 being the closest to the farmers’ fields and track 1 being the farthest. There are two railroads—the Reading Railroad and the Short Line Railroad220—that send trains down these tracks.221 Like Bob, the Short Line Railroad has the option to create transaction costs that will prevent it from being party to any transaction between the railroads and the farmers.

217 See, e.g., Jerry Kang, Information Privacy in Cyberspace Transactions, 50 STAN. L. REV. 1193, 1250-51 (1998) (“[E]ven if transaction costs are not large enough to transform default rules into immutable ones, the default rule still matters because ‘it determines who will bargain and at what cost.’”); Parts IV.A-B, infra. 218 See Part III.C, supra. 219 This is an application of the low-hanging-fruit principle. See FRANK & BERNANKE, supra note Error: Reference source not found, at 46-47.220 We do not consider the incentives of the B&O and Pennsylvania Railroads. 221 Presumably, they are not sending trains down the tracks at the same time.

48

Both railroads have the same access to all four tracks.222

Each can run trains on all four tracks, the furthest three tracks, the furthest two tracks, the furthest track, or run no trains at all.223

Running trains on more tracks increases the benefits to the railroads, but requires the farmers to forgo planting more wheat. For simplicity, we number these various options based on the closest track to the farmers’ fields on which trains run—which also corresponds to the number of tracks on which trains run.224

Table 3, below, shows the benefits that running trains on various tracks accords to the farmers and the railroads, and the net surplus that each scenario creates.

TABLE 3225

Closest Track on Which

Trains Run

Value to Farmers

Value to Railroads

Total Value

4 $0 $3800 $3800

3 $1800 $3200 $5000

2 $2700 $2700 $5400

1 $3200 $1800 $5000

No Trains Run $3800 $0 $3800

This table groups the consequences of running trains on varied numbers of tracks for both farmers as a group and for both railroads as a group. This approach makes sense for two reasons: First, note that because the same railroad tracks abut both farms, 222 For example, this scenario could arise if the government owned the tracks and made them available to all railroads, or if a private third party owned the tracks and leased them out for use by others. 223 Again, we assume that the railroads are good neighbors. See note Error:Reference source not found, supra, and sources cited therein. 224 For example, recall that track 4 is the closest to the farmers’ fields. If trains run on track 4, then they are also running on tracks 1, 2, and 3. Thus, there are also trains running on 4 tracks. 225 Note that the symmetry inherent in Table 3 is not necessary to any of the results of this Subpart. Structuring the table this way merely simplifies the exposition.

49

any locomotive going by one farm at a particular distance from the field will go by the other farm at the same distance. Thus, each farmer will have the same amount of sparks thrown onto his or her land. The railroads’ situation is equivalent; if Alice and Bob plant far enough back from the train tracks that the Reading Railroad can safely run trains on a particular track, the Short Line Railroad can also safely run trains on that same track.

Second, we assume that Alice and Bob have similar fields, so that they receive the same benefit from planting wheat closer to the track, and that the two railroads have similar cost structures and opportunities, so that they receive the same benefits from running trains on particular tracks. We therefore assume that the value to each individual farmer in each scenario in Table 3 is one half of the total shown for both farmers, and the value that each railroad receives from running a given number of trains is one half of the total shown for both railroads.

Examining the rightmost column of Table 3 makes clear that the efficient outcome is to run trains on tracks 1 and 2 only; doing so creates more value than any other option. Running trains on one additional or one fewer track are the next most efficient options. The least efficient option is running trains on all four tracks or not running any trains at all. Note also that as the outcome becomes farther from the efficient outcome, the loss of value relative to the efficient outcome becomes much larger.

We now demonstrate that the parties will not reach the efficient outcome if the legal default rule enables railroads to run trains on all four tracks, or forbids them from running any trains at all.

We first consider what happens if the law accords the railroads the right to run trains on all four tracks.226 The question then becomes whether the farmers will strike a bargain with the railroads in which the farmers pay the railroads to run trains on fewer tracks and, if so, how many fewer tracks.

There are two cases to consider. In the first scenario, Bob chooses not to create transaction costs that keep him out of the

226 This is equivalent to a legal rule that grants railroads an absolute right to emit sparks from their locomotives.

50

negotiations. Combined, Alice and Bob should strike a deal with the railroads to run trains on tracks 1 and 2 only;227 this maximizes the joint value that all the parties receive.228 The railroads would be willing do so in exchange for any amount that exceeds $1100, and Alice and Bob stand to gain $2700 combined.

Assuming, as previously, that the parties split the gains created evenly among themselves,229 each farmer should contribute $950 to defray the railroads’ costs of installation. This would leave each railroad $400 better off than it was when it ran trains on all four tracks.230 Meanwhile, each farmer would earn an additional $1350 from growing additional wheat. After paying $950 to the railroads, each farmer would be $400 better off as well.

In the second scenario, Bob chooses to create transaction costs that prevent him from negotiating with the railroads. This means that Alice is the only one negotiating with the railroads. Alice gets $900 of value from having no trains run on track 4, and running trains on track 4 is only worth $600 to the railroads. Thus, Alice and the railroads will reach a contract under which the railroads do not run trains on track 4. If trains stop running on track 3, Alice benefits by $450. However, running trains on track 3 is worth $500 to the railroads. Alice therefore will not be willing to pay the railroads enough to induce them to stop running trains on track 3. Accordingly, if Bob stays away from the negotiating table, Alice will pay the railroads not to run trains on track 4.

This result is less efficient than the first scenario. Running trains on tracks 1, 2, and 3 creates $5000 of value, while only running trains on tracks 1 and 2 creates $5400 of value.231

However, the value is divided very differently in the two scenarios. In the first scenario, Alice and the two railroads have a $300 increase in value to divvy up between themselves. Bob, on the other hand, captures the remaining $900 of value himself: The

227 This gives Alice and Bob combined benefits of $2700, but only costs the railroads $1100. 228 See Table 3, supra.229 See text accompanying notes Error: Reference source not found and Error:Reference source not found, supra. 230 Each railroad gets $950 cash; simultaneously, running trains on 2 tracks instead of all 4 costs each railroad $550 of value. Thus, the agreement gives each railroad $400 in net benefits. 231 See Table 3, supra.

51

railroads not running trains on track 4 allows Bob to grow $900 more wheat, and he does not need to pay the railroads anything. This means that Bob does much better under this scenario than under the previous one.

Because Bob makes much more profit in scenario two than scenario one, he will choose to create transaction costs in order to bring that scenario into effect. Accordingly, a legal rule that gives railroads the legal right to run trains on all four tracks will produce scenario two above and result in trains running on tracks 1, 2, and 3 instead of only on tracks 1 and 2—an inefficient result.

We now turn to the opposite extreme—a legal rule that prevents trains from running on any tracks.232 This corresponds to the bottom row of Table 3, in which no trains run. This produces $3800 of value for the farmers, but none for the railroads.

Similarly to the previous example, the railroads may be able to pay the farmers in exchange for the farmers allowing trains to run on certain tracks. The Short Line Railroad, like Bob, has the ability to create transaction costs that prevent it from being party to any agreement. Thus, as before, there are two cases to consider.

In the first case, the Short Line Railroad does not create transaction costs and both railroads negotiate with both farmers. With all parties at the table, the parties should be expected to reach an efficient result. As noted previously, that corresponds to trains running on tracks 1 and 2 only.

Compared to the situation in which no trains run, running trains on tracks 1 and 2 creates an additional $1600 of value. Assuming that the parties again divide that surplus among themselves evenly,233 this corresponds to the railroads paying each farmer $950 to allow trains to run on tracks 1 and 2 instead of insisting that no trains run. This will require the farmers to plant farther back from the train track, which will cost each farmer $550 in forgone wheat. However, after the payment from the railroads is accounted for, each farmer is $400 better off. Similarly, each railroad will earn $1350 from running trains on tracks 1 and 2.

232 This is equivalent to a legal rule that grants the farmers an absolute right to not have sparks cast onto their land. 233 See text accompanying notes Error: Reference source not found and Error:Reference source not found, supra.

52

After subtracting out the $950 each railroad must pay the farmers, each railroad is $400 richer.

Alternatively, the Short Line Railroad could instead choose to create transaction costs that prevent it from being party to any negotiations with Alice and Bob. If that happens, the Reading Railroad will still be negotiating with the farmers. Reading Railroad will be willing to pay the farmers enough to allow it to run trains on track 1. Reading would earn $900 from running trains on that track, and the farmers would only lose $600 worth of wheat production. This shift creates $300 of surplus for the two farmers and Reading to allocate among themselves.

Allowing trains to run on track 2 as well would cost the farmers an additional $500 in lost wheat production, but would only benefit Reading Railroad by $450. Accordingly, if the Short Line Railroad does not come to the negotiating table, the Reading Railroad will not be willing to pay the farmers enough to allow it to run trains on track 2.

Overall, this scenario produces a less efficient result than when Short Line participates in the negotiations. By not striking an agreement that would let the railroads run trains on track 2, the parties are leaving $400 of potential gains from trade on the table.

Nonetheless, Short Line Railroad prefers this second scenario. Compared to when no trains run, the first scenario gives Short Line $400 in additional value. The second scenario, in contrast, makes Short Line $900 better off, because it does not need to make any payments to the farmers. Thus, Short Line prefers the second scenario. Accordingly, a legal rule that forbids railroads from running trains on any of the tracks will lead to trains running on track 1 only. This is an inefficient outcome.

In this example, initial legal default rules that are too far from the efficient result in either direction produce inefficient outcomes. However, as long as the legal default rule is at least an intermediate outcome—that is, the legal default rule is close enough to the efficient result—the parties reach an efficient outcome. The logic underlying this effect is the same as in Table 2234: The parties that can create transaction costs must participate

234 See Table 2, supra.

53

in the agreement to produce an efficient outcome. Alice alone will never be willing to pay the railroads enough to stop running trains on track 3.235 Thus, if the law allows the railroads to run trains on tracks 1, 2, and 3, Bob can only improve on that outcome by teaming up with Alice to buy off the railroads and increase the size of the pie. The same logic applies to the Short Line Railroad if the law only permits the railroads to run trains on track 4. Table 4, below, summarizes these results.

235 See note Error: Reference source not found, supra, and accompanying text.

54

TABLE 4

Closest Track on Which

Law Allows Trains to Run

Parties Receiving Payments

Voluntary Transaction

Costs?

Closest Track on Which

Trains Run

Efficient Result?

4 RailroadsBob Will

Pay to Create

3 No

3 RailroadsBob Will

Pay to Remove

2 Yes

2 Neither Indifferent 2 Yes

1 FarmersShort Line Will Pay to

Remove2 Yes

No Trains

RunFarmers

Short Line Will Pay to

Create1 No

The analysis above naturally raises the question of when and how the legal default is likely to approximate the efficient outcome. One such way is to define legal rights by implicitly or explicitly weighing costs and benefits. For example, there are a number of instances in which the common law gives parties the right to engage in a particular activity so long as their behavior is reasonable.236 These “reasonable use” rules roughly correspond to courts weighing the costs and benefits of the proposed activity and ruling accordingly.237

236 Ideally, such rules should consider all costs and benefits, including those that accrue to parties outside of the litigation. It is unclear to what extent courts incorporate such factors. 237 This is, of course, an oversimplification, but it is a useful high-level view. See, e.g., JOHN W. JOHNSON, UNITED STATES WATER LAW: AN INTRODUCTION 90 (2008) (describing how courts decide whether an action is reasonable by

55

Many of the areas governed by reasonable use rules are areas for which voluntary transaction costs are relevant.238

Nuisance law, for example, protects landowners against substantial and unreasonable interference with their use and enjoyment of their land.239 Landowners have typically brought nuisance cases against neighbors engaging in activities that produce significant amounts of noise, dust, smoke, odors, or vibrations.240 These types of behaviors are likely to affect multiple neighbors of the offending party. However, in many cases they will not affect so many other landowners as to render a negotiated transaction prohibitively difficult.241 Indeed, nuisance cases provide the fact patterns for some of the most famous examples illustrating the Coase Theorem.242

To take another example from property law, landowners in most states have the right to modify their land to reasonably alter or divert the natural flow of water across their properties.243 Such diversions will often affect multiple identifiable neighbors, who could plausibly negotiate an agreement. Similarly, in states with riparian water law systems, landowners whose properties adjoin waterways may take water for all reasonable uses, so long as they

weighing the value of the action against the damages caused to others’ property from the action in question, as compared to the damages that would result from alternative actions).238 To be clear, we do not argue that state legislatures or common law courts imposed these rules based on some intuitive understanding of voluntary transaction costs. These rules can be derived based on other principles, such as conceptions of justice or natural law. See, e.g., Swett v. Cutts, 50 N.H. 439, 444 (1870) (stating that the right of a riparian owner is “a natural right . . . to partake in the enjoyment of the common bounty of Providence, as in the cases of light and air”). They can also be derived based on other efficiency arguments; since all real-world transactions involve transaction costs, an efficient legal rule that obviates the need for further negotiations among the parties will enhance efficiency. See Coase, supra note Error: Reference source not found, at 19.239 See, e.g., San Diego Gas & Electric Co. v. Superior Court, 920 P.2d 669, 696 (Cal. 1996); JOSEPH SINGER, PROPERTY LAW: RULES, POLICIES, AND PRACTICES 307 (3d ed. 2002).240 In re Chicago Flood Litigation, 680 N.E.2d 265, 278 (Ill. 1997); SINGER, supra note Error: Reference source not found, at 306 (also listing attracting insects among the actions giving rise to causes of action for nuisance). 241 Such scenarios often constitute public nuisances instead of private ones. See RESTATEMENT (SECOND) OF TORTS § 821B(1). 242 See sources cited in notes Error: Reference source not found-Error: Referencesource not found, supra.243 See JOHNSON, supra note Error: Reference source not found, at 90-91 (describing reasonable use rules and their prevalence, and noting that states that purport to follow other rules effectively follow reasonable use rules in practice).

56

do not unreasonably interfere with others’ uses of water from the waterway.244 The possibility of voluntary transaction costs provides additional support for these reasonable use rules and suggests that such rules could prove beneficial in additional circumstances.245

A drawback of reasonable use rules is that they require courts and parties to have more information than more bright-line rules do. This can make bright-line rules more attractive in certain instances.

For example, if a railroad has an absolute right to emit sparks, the railroad knows what its legal rights are and can act accordingly. On the other hand, if the railroad has the right to emit sparks only if it does not unreasonably interfere with the farmer’s right to use its property, it may not be clear to the railroad what its legal rights are. At the very least, the railroad will have to estimate the likely effects of its spark emission activity on the farmer.246

Similarly, courts will have to weigh the costs and benefits to the railroads and farmers against each other. This may prove challenging, and the court may err.

However, we note that a legal default rule that does not exactly match the efficient outcome may still provide significant benefits.247 Returning to our two-farmer, two-railroad example above, a perfectly conceived and implemented reasonable use rule would result in trains running on tracks 1 and 2 only.248 But if courts overweigh the interests of the farmers or the railroads, this might not happen, and the actual legal rule might permit trains to

244 Id. at 36-37.245 To be clear, we do not argue that state legislatures or common law courts imposed these rules based on some intuitive understanding of voluntary transaction costs. These rules can be derived based on other principles, such as conceptions of justice or natural law. See, e.g., Swett v. Cutts, 50 N.H. 439, 444 (1870) (stating that the right of a riparian owner is “a natural right . . . to partake in the enjoyment of the common bounty of Providence, as in the cases of light and air”). They can also be derived based on other efficiency arguments; since all real-world transactions involve transaction costs, an efficient legal rule that obviates the need for further negotiations among the parties will enhance efficiency. See Coase, supra note Error: Reference source not found, at 19.246 Unclear rights may also increase the frequency of litigation, which is costly. See POSNER, supra note Error: Reference source not found, at 791. 247 This point also helps ameliorate the concern raised in note Error: Referencesource not found, supra. 248 See Table 3, supra.

57

run on tracks 1, 2, and 3, or only on track 1. In either instance, the legal rule will be close to efficient, and the parties will be able to close the remaining gap.249 On the other hand, assigning an absolute right to either the farmers or the railroads does not produce an efficient result.250

This same logic applies to statutes that create bright-line rules. Legislatures are often better equipped than courts to make large-scale societal judgments regarding the costs and benefits of particular activities.251 A legal rule that draws a bright line—even an arbitrary one—will have the same benefits as a reasonable use rule, so long as that bright line is close to the efficient outcome.252

B. The Importance of Legal Remedies

In all of the examples we have analyzed so far, we have assumed that legally protected parties can prevent any behavior that violates the law if they so choose. However, the remedies that the law provides to protected parties have strong effects on both whether the parties will strike a deal and the type of deal that the parties will strike.253 We first discuss the differing ways in which remedies affect the parties’ decisions when transaction costs are low and when they are high. We then consider the relative merits of injunctive remedies and liability remedies, the two most common types of remedies that U.S. laws provide.

249 See Table 4, supra. 250 See Table 4, supra.251 POSNER, supra note Error: Reference source not found, at 715-16.252 A more efficient legal default rule is also preferable if no bargaining takes place. This is particularly important if parties are able to create transaction costs that prevent all bargaining. See Part III.A, supra. 253 In doing so, we supplement extensive literature on this topic. See, e.g., Brooks, supra note Error: Reference source not found; Louis Kaplow & Steven Shavell, Property Rules Versus Liability Rules: An Economic Analysis, 109 HARV. L. REV. 713 (1996); A. Mitchell Polinsky, Resolving Nuisance Disputes: The Simple Economics of Injunctive and Damage Remedies, 32 STAN. L. REV. 1075, 1088-92 (1980); Ayres & Talley, supra note Error: Reference source notfound; Louis Kaplow & Steven Shavell, Do Liability Rules Facilitate Bargaining? A Reply to Ayres and Talley, 105 YALE L.J. 221 (1995); Ian Ayres & Eric Talley, Distinguishing Between Consensual and Nonconsensual Advantages of Liability Rules, 105 YALE L.J. 235 (1995).

58

1. Remedies in Low- and High- Transaction Cost Environments

Remedies are important in both low- and high-transaction cost environments, but they have different effects in each circumstance.254 When there are large transaction costs, so that Coasean bargaining is impossible, legal remedies determine what consequences the parties will reap from various competing courses of action. This affects what decisions the parties will make, and can either prevent or facilitate the achievement of an efficient result. In contrast, when there are no transaction costs, the parties can always strike a Coasean bargain that produces efficiency, regardless of what remedies the law provides. However, the legal remedies available still shape the agreements that the parties will make.255

These points are best illustrated through example. Consider another farmer and railroad scenario. Assume that running the train is worth $1000 to the railroad, and that the farmer will lose $500 if he plants further back from the track. Since the farmer’s loss from planting back from the track is smaller than the $1000 gain the railroad gets from running the train, it is efficient for the train to run. Assume that farmers have the legal right to not have any sparks cast onto their fields.

First, suppose that transaction costs are high, so that the parties cannot negotiate. Even though farmers have the legal right to not have any sparks cast onto their fields, that does not necessarily mean that the railroad will not run the train; the answer to that question depends on the consequences that the railroad will suffer if it violates the law.256 If running the train will incur a penalty that exceeds the $1000 benefit to the railroad of running

254 See also Giuseppe Dari-Mattiaci, Endogenous Transaction Costs (exploring the extent to which different legal remedies create transaction costs). 255 This dynamic is often referred to as bargaining in the shadow of the law. See, e.g., Robert H. Mnookin & Lewis Kornhauser, Bargaining in the Shadow of the Law: The Case of Divorce, 88 YALE L.J. 950, 968 (1979).256 The cost of violating the law may extend beyond the legally mandated punishment. There may be bad publicity, or political costs, associated with such violations as well, and these costs should be included in the railroad’s calculus as well. For simplicity, we assume throughout this paper that the only relevant consequences are the legally mandated punishments.

59

the train,257 a rational railroad will not run the train.258 Conversely, if the punishment for running the train is a fine of $500, a rational railroad will run its train and pay the fine; from the railroad’s perspective, it would not make sense to forgo $1000 from running its train solely to avoid paying a $500 fine. Thus, when transaction costs are high, the legal remedy can determine the actions the parties will take, which can prevent or enable an efficient result.

On the other hand, if there are no transaction costs, then regardless of what remedy the law provides, the railroad and farmer can strike a deal in which the railroad pays the farmer to plant further back from the track. The size of the payment to the farmer could be anywhere between $500 and $1000. In general, it is difficult to predict precisely where in this range the parties will strike their agreement. However, legal rules can channel the parties towards particular outcomes.

For example, many statutes entitle parties to fixed damage awards for violations of particular rights.259 Suppose that the farmer’s jurisdiction has a statute that entitles the farmer to an $800 payment from any railroad that throws sparks onto her property, and that this is the farmer’s only remedy. In such a situation, the railroad will be unwilling to pay the farmer more than $800.260 Similarly, the farmer should not accept a payment of less than $800.261 In such circumstances, the parties seem likely to transact at a price of approximately $800.262 Thus, regardless of

257 Or, equivalently, the $1000 opportunity cost of not running the train. 258 We assume there are no intermediate options. 259 See, e.g., 17 U.S.C. § 504(c) (2012) (setting statutory minimum damages of $750, and maximum damages of $30,000, for copyright infringement); 15 U.S.C. § 1117(c) (2012) (setting statutory minimum damages of $1000, and maximum damages of $200,000, for using counterfeit trademarks); 18 U.S.C. § 2520(c)(2)(B) (setting statutory damages for intercepting electronic communications of $100 per day, up to $10,000 total).260 Suppose the railroad agreed to pay the farmer $900 to allow it to run its train. If the railroad did not pay the farmer and simply ran the train, it would only owe $800 in damages. This would leave the railroad better off than under its deal with the farmer. Thus, the railroad should never agree to such a deal. 261 Suppose the farmer and the railroad do not reach an agreement. The railroad will be better off if it runs its train and pays the $800 penalty than if it does not run the train and loses $1000 of value. Thus, the railroad is likely to run its train and pay the farmer the $800 penalty. Knowing this, the farmer will be reluctant to part with her legal right for less than $800. 262 In reality, there are always costs to enforcing legal rights. Since striking a deal allows the parties to avoid incurring these costs (in exchange for incurring the costs of deal-making), the set of possible outcomes would generally extend

60

the size of transaction costs, legal remedies remain important, though their effects vary in each instance.

2. Injunctive Remedies vs. Liability Remedies

U.S. laws provide considerable variation in the legal remedies that they accord to parties who suffer various wrongs, but injunctive remedies and liability remedies are the most common.263

A party entitled to an injunctive remedy can secure a court order commanding the other party to cease its unlawful behavior.264

A party who violates such an order may be held in contempt of court, which can be punished through heavy fine or imprisonment.265 In all of the examples we analyzed in previous subparts, we assumed that legally protected parties could prevent any behavior that violates the legal rule if they so choose. This is equivalent to assuming that aggrieved parties are entitled to an injunctive remedy.

Liability remedies, in contrast, only entitle the damaged party to recover monetary damages from the party violating the law.266 In most instances, the wronged party is entitled to recover the amount of damages caused by the violator’s wrongful actions—such rules are termed compensatory liability rules, as they exactly compensate the wronged party for the harm they suffered—but this is not a necessary feature of liability rules and there are many counterexamples.267

to some range around and including an $800 payment. For example, suppose that enforcing a legal right by lawsuit requires each party to incur $50 of legal costs. If the railroad illegally runs its train and the farmer files suit, the total cost to the railroad will be $850 and the farmer will net $750 cash. Assuming that negotiating a deal is costless, the parties could plausibly strike a deal at any price between $750 and $850 in this instance. See POSNER, supra note Error:Reference source not found, at 791. 263 See Calabresi & Melamed, supra note Error: Reference source not found, at 1092. 264 Id.265 Id.266 Id.267 Perhaps the most well-known is the damages rule for antitrust violations, which allows wronged parties to recover three times the damages that they suffer, but there are many examples. See, e.g., 15 U.S.C. § 1117(b) (2012) (trademark violations involving counterfeit marks); 35 U.S.C. § 284 (2012) (authorizing courts to award treble damages in patent infringement cases).

61

Liability remedies facilitate an efficient result in more circumstances than injunctive remedies do. Given a legal rule, any result that can be achieved via injunctive remedies can also be achieved via liability remedies. The key intuition is that a liability remedy that provides for a large enough damages payment will always prevent a party from acting in a way that triggers her liability. In other words, sufficiently large liability remedies produce the same effects as injunctive remedies. The reverse is not true; thus, there are certain instances in which liability remedies can produce efficient results that injunctive remedies cannot.268

However, neither injunctive nor common liability remedies guarantee that the parties will reach an efficient result.

To see how liability remedies can promote efficiency when injunctive remedies do not, we revisit the two-farmer, two-railroad-track example that we considered in Part III.C. The railroad can run trains on both tracks, the track farther from the farmers’ property, or neither track.269 The closer that the railroad runs trains to the farmers’ property, the farther back the farmers must plant their wheat from the track.

The payoffs for the parties under different circumstances are shown in Table 5, below.

TABLE 5

Value to Alice

Value to Bob

Value to Railroad

No Trains Run $1000 $1000 $0

Trains Run on Far Track Only

$650 $650 $400

Trains Run on Both Tracks

$0 $0 $800

268 For the remainder of this Subpart, we only consider liability rules that are compensatory.269 Again, we assume that the railroad is a good neighbor. See note Error:Reference source not found, supra, and sources cited therein.

62

Suppose that all three parties have the ability to create transaction costs that keep themselves away from the negotiating table. In addition, suppose that Alice has a hill on the edge of her property that is prone to mudslides. If she chooses, Alice can redirect these mudslides so that they spill onto the near track, rendering it impassible. Suppose that the law gives the railroad the right to run trains on both tracks, and that it is illegal for Alice to redirect the mudslides to block the near track.

If the railroad’s legal right is backed by an injunctive remedy, Alice will not be able to block the near track. If she attempts to do so, she is violating the law, and the railroad can secure an injunction that forces her to stop. This scenario is therefore exactly identical to the scenario analyzed in Part III.C: In both scenarios, the same parties have the same choices and the same payoffs. There, as here, the efficient result is for no trains to run.270 In both scenarios, both farmers and the railroad must come to the negotiating table for the parties to achieve an efficient result.271 Because this situation is identical to the scenario analyzed in Part III.C, it produces the same result; there is never an efficient outcome because it is never in the interests of both farmers to come to the negotiating table.

Now suppose instead that the law only provides the railroad with a compensatory liability remedy. In other words, if Alice blocks the near track, the railroad cannot unblock the track; it can only recover damages from Alice in an amount equal to the benefit that the railroad would have received if the near track had not been blocked. It is now possible for the parties to achieve an efficient result.272

Consider again a scenario in which all parties come to the table. Suppose that the farmers pay the railroad a total of $900 not to run the train, $600 of which comes from Alice and $300 of which comes from Bob. This produces an efficient result; the 270 This produces $2000 of value, which is more than the $1700 produced when trains run on the far track only and the $800 produced when trains run on both tracks. See Table 5, supra.271 In both scenarios, inducing the railroad not to run trains on the far track requires a payment of $400, but each farmer only benefits by $350 from that train not running. Thus, an efficient deal with the railroad is only possible if both farmers come to the table. See Part III.C, supra. 272 We note that it is not certain, however. See Companion Paper, supra note Error: Reference source not found, at 6.

63

question is whether all parties are acting in their interests by coming to the table.

The railroad is better off coming to the table. If the railroad does not come to the table, it will not be party to any agreement.273

Alice will then block the near track.274 The railroad will run trains on the far track and collect damages from Alice, which will compensate the railroad for the value it loses from not running trains on the near track. The end result is that the railroad will only receive $800 of value, which is $100 less than the $900 it receives pursuant to the parties’ agreement.275

Similarly, for a farmer to evaluate whether coming to the table makes her better off, she must know what will happen if she does not come to the table. Bob currently gets $700 of value.276 If he leaves the table, the best outcome that he can hope for is that Alice will stop the railroad from running trains on the near track.277

This will give Bob only $650 of net value.278 Thus, Bob is better off staying at the table.

But what of Alice? Currently, she receives $400 of value. One might think that, by leaving the table, she could force Bob to negotiate with the railroad alone. Bob would then pay the railroad not to run trains on the near track, giving Alice $650 of value.279

Indeed, that could happen—but it is not the only possibility.

Because of the liability remedy, Alice cannot truly leave negotiations. Even when Alice leaves the table, she still has the ability to change the railroad’s behavior: she can block the near track and pay the railroad for its losses. Bob could plausibly decide not to strike any deal with the railroad if he thinks that, when push comes to shove, Alice will block the near track. And, given that Bob does not negotiate with the railroad, it is in Alice’s

273 See Companion Paper, supra note Error: Reference source not found, at 4 n.5.274 Blocking the near track enables Alice to plant $650 in additional wheat. These gains exceed the $400 damages payment she must make to the railroad. 275 See Table 5, supra. 276 He grows $1000 worth of wheat and pays the railroad $300.277 See note Error: Reference source not found, supra, and accompanying text. Alice may accomplish this either by negotiating a deal with the railroad or by blocking the near track. 278 See Table 5, supra.279 Bob and Alice have the same payoffs; see note Error: Reference source notfound, supra, and accompanying text.

64

interest to block the near track, as this enables her to plant $650 of wheat at a cost of only $400.

That scenario nets Alice only $250, which is less than the $400 she receives when both Bob and Alice come to the negotiating table.280 So, if this scenario is likely—and this scenario is perfectly plausible, as it entails everyone acting in her best interest—then Alice is better off coming to the negotiating table. Thus, under a liability rule, an efficient result is possible, though generally not guaranteed.

A few additional points deserve mention. First, the assumption that only Alice can block the near track merely clarifies the analysis; it is not necessary to our result. The behaviors described above are also equilibrium behaviors when both Alice and Bob have the ability to block the near track.281 It is also worth noting that the real world is often asymmetric. Even when an action affects multiple parties in the same way, some of those parties may naturally be in a better position than others to block that action.

For example, consider a scenario in which a beekeeper’s bees pollinate a local orchard’s apple trees but also sting a local rancher’s cattle.282 Suppose that the rancher has the right to restrict the number of bees in the beekeeper’s hives, and that right is enforced by a liability remedy. In this scenario, the beekeeper, and not the farmer, is naturally in a position to violate the rancher’s right and pay him damages, even though both the farmer and the beekeeper benefit from the beekeeper having more bees.283 280 Under the proposed agreement, Alice pays $600 to the railroad and gets to grow an additional $1000 worth of wheat. 281 Those behaviors are not the only equilibrium behaviors. For example, Alice could pay $300 and Bob could pay $600. This would be an equilibrium so long as Bob thought that, if he refused to come to the table, Alice would not negotiate with the railroad, leaving him to block the track.282 This is essentially a combination of two famous Coasean examples. See Elodie Bertrand, What do Cattle and Bees Tell Us About the Coase Theorem?, Communication to the 13th Conference of the Charles Gide Association, Paris, May 27-29, 2010, available at colloquegide2010.univ-paris1.fr/IMG/pdf/ Bertrand.pdf.283 Similarly, consider the famous Coasean example involving a rancher who neighbors a farmer, and who wishes to allow his cattle to graze on the farmer’s lands. See id. The rancher’s cattle may wander across the neighboring farmer’s property to other farmers’ lands, but it may be much easier for the first farmer to build a fence that restrains the cattle, particularly if there are only a handful of bridges or other chokepoints through which the cattle can cross onto the first

65

Moreover, liability remedies themselves are fundamentally asymmetric. An individual who deviates from the behavior required by the legal rule is required to pay damages to those parties who are hurt by her actions. However, that same individual is not entitled to receive a payment from those parties who benefit from her actions.

Theoretically, liability remedies could be fully symmetric. Such remedies would give all parties incentives to exercise their decision-making rights in an efficient manner, regardless of the legal rule. Essentially, there would no longer be positive or negative externalities—actors would experience all of the positive or negative consequences of their actions, which would lead them to act in an efficient manner.284 Yet, despite the theoretical advantages of fully symmetric liability rules, such rules have never been seriously considered because of the administrative challenges they pose.

Similarly, courts must have more information to administer a liability remedy than an injunctive remedy.285 A liability remedy requires courts to calculate how much damage a protected party suffered from a violation of the law. In contrast, a court does not need to determine the amount of damages caused to apply injunctive relief; the court must merely decide whether a particular action violates the legal rule.286 Moreover, enforcing a liability remedy requires less information than enforcing a reasonable use rule,287 but is less likely to bring about an efficient result.288

farmer’s property. 284 POSNER, supra note Error: Reference source not found, at 3-4.285 Note also that enforcing a liability remedy requires less information than enforcing a reasonable use rule. The reasonable use rule requires a court to weigh the benefits of an activity against its costs, while a liability rule only requires a court to evaluate the negative consequences of an activity. 286 See, e.g., Calabresi & Melamed, supra note Error: Reference source notfound; David McGowan, Website Access: The Case for Consent, 35 LOY. U. CHI. L.J. 341, 342-43 (2003) (arguing that injunctive remedies are preferable when bargaining is possible because judicial cost-benefit analysis is imperfect).287 The reasonable use rule requires a court to weigh the benefits of an activity against its costs, while a liability rule only requires a court to evaluate the negative consequences of an activity. 288 See Part III.A.2, supra.

66

C. Distributional Consequences

Throughout this Article, we have emphasized how various legal rules and legal remedies affect the likelihood that the parties achieve an efficient result. However, this is only half of the story. As Coase recognized, the choices among legal rules and remedies have distributional consequences. This is true irrespective of whether these choices have efficiency consequences. Legally protected parties may be able to secure cash payments from other parties that they otherwise would not receive.289 Liability rules can restrict the prices at which parties are willing to transact.290 This can be a boon or a bane to individual parties, depending on the specifics of a given situation.

Because the distributional consequences of these choices are so context-specific, they are not conducive to analysis in the abstract. We merely note that, depending on the specific circumstances of a particular situation and one’s policy preferences, one might prefer one legal rule or remedy over another based on its distributional effects, even though that rule or remedy is less likely to produce an efficient result.

V. CONCLUSION

In essence, the Coase Theorem makes an appeal to reason: If only everyone could get together at one big table and talk things out, they would always reach efficient results. Coase devised his famous theorem to highlight the difficulties involved in bringing everyone together and negotiating an agreement—that is, the structural transaction costs that parties face. Our analysis demonstrates that there are additional difficulties inherent in Coase’s hypothetical: In many instances, there will be parties who will not want to let everyone come to the table, and those parties will erect obstacles to prevent that from happening.

Consequently, the design of legal rules is very important, even when structural transaction costs are low. There is a trade-off between the likelihood that a legal rule or remedy will produce efficient results and the information that courts or legislators must have to implement it. Fully specified efficient rules guarantee 289 See, e.g., notes Error: Reference source not found-Error: Reference sourcenot found, supra, and accompanying text. 290 See Part III.B.1, supra.

67

efficiency, but require the strongest assumptions about the information available to policymakers.291 Defining legal rights in ways that track the efficient outcome is likely to encourage efficiency, but requires hard factual determinations. Similarly, liability remedies are more likely to promote efficiency than injunctive remedies, but are more difficult for courts to enforce.

Regardless of the legal regime, it is crucial to recognize that transaction costs are not merely a feature of the parties’ environment. The parties themselves can—and will—create them on their own. Thus, Coase’s point that transaction costs are ubiquitous and important was even more far-reaching than has generally been recognized—in short, Coase was more right than perhaps even he knew.

291 See Part IV.A, supra.

68