Printer Friendly

Contracting for cooperation in recovery.

INTRODUCTION

1. THE PRESENT LIMITATIONS OF CONTRACT LAW
   A. Underenforcement and Contract Duties To Cooperate in Recovery
   B. Compensatory Remedies
     1. The Catch-22
     2. Other Losses: Avoidable Harms, Delayed Recovery, and Litigation
     Costs
     3. Liquidated Damages for Obstructive Breach

II. EXTRACOMPENSATORY CONTRACT REMEDIES
    A. The Right To Terminate
    B. Adverse Inferences
    C. Penalties and Punitive Damages: A New Argument

III. CONTRACTING FOR FRAUD LIABILITY
     A. An Early Foray: Rereading Seaman's Direct Buying Services
        v. Standard Oil
     B. Contracting for Fraud Liability
     C. Promissory Representations of No Intent To Obstruct
     D. The Limitations of Liability in Fraud

IV. THE UTILITY OF CONTRACTUAL DUTIES NOT TO OBSTRUCT
    A. The Costs of Penalties and Punitive Damages
    B. Multipliers

CONCLUSION


INTRODUCTION

In 1984, the city of Richmond contracted with McDevitt Street Bovis for the construction of a new baseball stadium. The stadium was built, but a decade later the city discovered deterioration in the concrete tubes supporting the cantilevered roof, caused by McDevitt's breach of its contractual duty to fill the tubes with grout. The city sued, claiming both breach of contract and fraud, the latter based on false certificates of completion and other documents McDevitt had submitted. In Richmond Metropolitan Authority v. McDevitt Street Bovis, Inc., the Virginia Supreme Court held that because the contract required McDevitt to provide the certificates and other documents, misrepresentations in them might give the city a right to damages for breach, but could not give rise to liability in fraud. (1)

McDevitt is a good example of how courts police the border between contract and tort, protecting contract against, among other things, incursion by the punitive damages available in tort. But consider the holding's effect on Virginia contractors' decisions whether to submit accurate certificates of completion. Restricting recovery to the compensatory measures available in contract means that a Virginia builder who has breached its construction duties incurs little or no additional liability when it files a false certificate. The false certificate harms the purchaser first and foremost if it prevents her from discovering and recovering for nonconforming work. (2) To prove that harm--and even that the certificate was false--the purchaser must first show that the work was nonconforming, i.e., that but for the false certificate, she would have recovered for breach. But if the purchaser can show nonconforming work, then she already has a winning claim for breach of the underlying construction duty, which will compensate her for her losses. That is, if the purchaser can show that the false certificate harmed her, it did not. The Virginia Supreme Court's decision to limit recovery for false certificates of completion to the compensatory damages available in contract renders legal liability for breach of the certification requirement irrelevant.

This strange situation is not limited to certificates of completion in the building industry. The above argument, or one like it, applies to any contract term whose purpose is to make it easier for the promisee to discover and prove breach. Common examples include royalty reports, record-keeping requirements, certificates of compliance, and auditing rights. Such terms belong to a hitherto underappreciated genus of contract terms: duties to undertake acts that promote the other side's recovery for breach. I will use "obstruction of recovery" or "obstruction" to refer both to promisor actions that aim to avoid legal liability and to the failure to act in ways that would assist in the recovery of damages due where there is a duty to do so. I will use "cooperation in recovery," or simply "cooperation," to designate nonobstruction. (3) Where one side is particularly worried that she might be unable to recover for any breach, the other side might offer to undertake a contractual duty to cooperate in recovery--to agree, for instance, not to hide nonperformance, to keep complete records, or to provide information about performance.

But here's the problem: to recover compensatory damages for the breach of a duty to cooperate, a plaintiff must be able to demonstrate harm. The primary harm of an obstructive breach is that it prevents the promisee from recovering for breach of the underlying, or first-order, duty. Showing that harm requires independent proof of first-order breach--that the plaintiff was entitled to the damages she did not recover. But if the plaintiff can prove first-order breach, she can recover on that basis--which means that the obstructive breach has not harmed her. The upshot is a catch-22: a plaintiff can show obstructive harm only if she has not suffered it.

The way to break out of the circle is to attach extracompensatory remedies--remedies that are not tied to the plaintiff's verifiable losses--to obstructive breach. While the received wisdom is that optimal remedies for breach are always at or near the expectation measure, this cannot be so when it comes to duties to cooperate in recovery. (4) If such contract terms are to make a practical difference, they must be backed by punitive damages, penalties or other remedies neither conditioned on nor limited to compensation for harm done.

At present, courts will not enforce a penalty or punitive damages clause, which brings the analysis back to the potentially positive role of fraud liability. There is a longstanding debate within the courts and legal scholarship about whether parties should be able to contract out of liability for their fraudulent misrepresentations. (5) What has not been noticed is that many agreements are structured to opt into such liability using terms that require one party to represent that it is not in breach (by submitting, for example, a certificate of completion). When false, such representations not only breach the contract, but can satisfy the elements of fraud and support a claim for punitive damages. The contractual duty to share information about performance can be secured by the extracompensatory remedies available in fraud.

The existence of such contract terms suggests that parties want effective duties to cooperate in recovery. And their utility provides additional support for a thesis Ian Ayres and I have developed elsewhere: that the extracompensatory remedies available in tort can have "well-defined function[s] within the apparatus of the law of contracts." (6) Yet, as exemplified in McDevitt, many courts are uncomfortable with this incursion of extracompensatory remedies into the world of contracts and have found ways to exclude fraud liability for acts that are also breaches. The most recent trend in this direction involves an expansive reading of the economic loss rule to bar liability for fraud in the performance. This Article argues that such rulings are mistaken and that courts should recognize the positive role fraud liability can play in contracts. But the fraud solution is second best. A better solution would be an exception to the rules against penalties and punitive damages when those remedies are attached to contractual duties to cooperate in recovery.

Robert Scott and George Triantis have recently observed that "contracts scholars [have focused] principally on the substantive terms and not on the ability of the parties to regulate the procedural course of their future enforcement." (7) This is confirmed by the scholarly neglect of duties to cooperate in recovery. Such duties are attempts to regulate contract enforcement. This Article therefore fills in the picture of how parties contract for the case of breach. (8)

The Article's conclusions also bear on an old dispute about Holmes's famous dictum that "[t]he duty to keep a contract at common law means a prediction that you must pay damages if you do not keep it,--and nothing else." (9) This "Holmesian heresy" is sometimes read to mean that, as far as the law is concerned, the promisor does not have a duty to perform, but an option to perform or pay damages. (10) Holmes himself rejected the alternative-promise reading, explaining that "the statement that the effect of a contract is the assumption of the risk of a future event does not mean that there is a second subsidiary promise to assume that risk, but that the assumption follows as a consequence directly enforced by the law, without the promisor's cooperation." (11) This Article examines what happens when the parties do make a "second subsidiary promise" to pay damages--or at least to cooperate in their recovery. Its surprising result is that mandatory rules limiting contract damages to compensatory measures mean that even when there is a second subsidiary promise, legal liability for its breach does not make a practical difference.

Part I of the Article describes why some parties, including promisors who know they might breach, should want to contract for legally enforceable duties not to obstruct recovery, and explains why currently available contract remedies--which are limited to compensation for actual losses--are often insufficient to enforce such duties. The core reason is the catch-22 described above: the promisee who can prove obstructive harm has not suffered it. In some cases, the effects of the catch-22 can be reduced through recovery for other harms of obstructive breach, such as delayed compensation and higher litigation costs, or by liquidating damages. But these alternative grounds of recovery do not sufficiently protect against obstruction in all contexts.

Part II considers three forms that contractually specified extracompensatory remedies for obstructive breach might take: the right to terminate the contract, an adverse inference with respect to first-order breach, and penalties or punitive damages. The right to terminate works only where the promisor attaches significant value to the continued existence of the contract, and where the promisee is not likely already to have the right to terminate based on a showing of first-order breach. Nor is the adverse inference solution effective where the promisee has separate proof of first-order breach. I conclude that only penalties and punitive damages provide generally effective remedies for obstructive breach. The law of most states presently prevents parties from contracting for either.

Part III describes how, with no contractual solution available, duties to cooperate might be secured through the law of torts. The most important judicial step in this direction was California's failed experiment with punitive damages for postbreach obstruction in Seaman's Direct Buying Service v. Standard Oil Co. of California. (12) A more promising approach can be found in parties' attempts to contract for representations of performance backed by fraud liability, though many courts have closed off this path by adopting expansive readings of the economic loss rule or similar doctrines. Finally, the parties can secure limited benefits with representations of intent to cooperate in recovery, again backed by the law of fraud.

Part IV completes the analysis by arguing that not only are extracompensatory remedies necessary for effective no-obstruct duties, but the benefits of cooperative duties are worth the costs of those remedies. Many familiar costs of penalties and punitive damages, such as deterring efficient breach, do not apply when they are attached to obstructive breach. And no-obstruct clauses backed by penalties or punitive damages are often a cheaper alternative than a more familiar solution to underenforcement: damage multipliers.

Before jumping into the analysis, a few words about method. My argument employs a broadly instrumentalist perspective, in that it assumes that among contract law's primary aims is facilitating transactions that would otherwise be prevented by lack of trust. (13) In order to capture practical reasoning in the absence of trust, I employ nontechnical economic analysis of the self-interested calculations of individuals entering into and acting within legally binding agreements. I assume both that each party engages in such cost-benefit analysis to decide whether to enter into and how to perform under the contract, and that each tries to anticipate and influence the cost-benefit analysis of his or her counterpart. In undertaking this analysis, I generally assume that the only curbs on self-interest are legal ones--that is, the prospect of legal liability for bad acts. This is of course a theoretical fiction: extralegal norms, sanctions, and other consequences also enter into the mix. Nor is every decision in a contractual relationship based on such heartless cost-benefit analysis. But if contract law has a function in structuring the ongoing relationship between the parties, it is to substitute where nonlegal mechanisms for coordination, cooperation, or collaboration do not suffice. Ignoring extralegal norms, incentives, and reasons is an analytic device for exploring how legal mechanisms might substitute for them.

I. THE PRESENT LIMITATIONS OF CONTRACT LAW

A. Underenforcement and Contract Duties To Cooperate in Recovery

I begin with a more precise description of the category of contract terms under discussion. These are terms that attempt to address underenforcement--the fact that many harmful breaches do not give rise to a remedy or equivalent settlement. The causes of underenforcement are manifold and familiar. Litigation costs are in some cases so high that it is not worth the bother to sue. When the parties are in a valuable ongoing relationship, the nonbreaching party may choose to forgo a lawsuit to secure future deals. Defendants are sometimes judgment proof. And most important for my purposes, when the nonbreaching party does sue, insufficient evidence, limited or unbalanced litigation resources, and court or jury error can conspire to prevent proof of breach.

The threat of underenforcement can be an obstacle to working out mutually acceptable terms. Parties enter into a contract when they believe its benefits outweigh its costs. A contract's benefits include the opportunity to recover damages in the case of breach. The value of that benefit depends on the probability of enforcement, or the likelihood that the promisee will recover when she has a meritorious breach of contract suit. When in the course of negotiations party B balks because she is worried about underenforcement, and party A wants to make the deal happen, A must find a way to sweeten the pot. A often has a number of options to choose from and will presumably pick the one that provides B the greatest additional benefit at the least cost to himself. Thus if the transaction is a simple sale of goods, seller A might offer concessions in the price or quantity term, additional assurances that he will deliver conforming goods on time, or a higher damage measure. If buyer B is particularly worried about underenforcement, A's cheapest option might be an additional term designed to promote B's ability to recover should A breach.

Several categories of contract provisions fit this description. Some contracts put the payment of damages out of the control of the breaching promisor and into the hands of the disappointed promisee or a third party. Examples include security deposits, installment payments, funds held in escrow, and performance bonds. (14) The parties can also contract into alternative adjudicative procedures that they believe will increase the likelihood of recovery. Choice of forum and choice of law clauses can be used to opt into more plaintiff-friendly or less costly legal regimes. Similarly, mediation and arbitration provisions can increase the probability of enforcement through specialization (the use of expert adjudicators) and cost savings (lower litigation costs). This second category also includes procedural tinkering like specialized discovery procedures or nonstandard burdens of proof. (15)

This Article examines a third category: contractual duties to cooperate in, or not obstruct, the promisee's recovery of damages due her. Obstruction comes in a variety of flavors. Most obvious is outright dissimulation. Promisors attempt to hide breach, they try to prevent or delay lawsuits with false assurances that performance will happen, they lie in court. But not all obstruction involves misrepresentation. A promisor might destroy records or other evidence of his nonperformance, transfer or manipulate assets to make himself judgment proof, or raise frivolous defenses and employ other delaying tactics in the course of litigation. Finally, my broad definition of "obstruct" is meant also to capture sins of omission where there is a duty to assist in recovery. These can include a promisor's decision not to inform the promisee of breach, his failure to keep records that would show nonperformance, his refusal to voluntarily pay damages obviously due, or his noncooperation in judicial proceedings. Each of these tactics decreases the likelihood that a meritorious breach-of-contract suit will succeed--both directly, by affecting legal processes, and indirectly, by increasing the costs of litigation.

Some obstructive behavior is subject to mandatory legal penalties. If a defendant gives false testimony about the existence of a contract, its content, or his performance, he commits perjury and risks criminal or civil sanctions. A promisor who, after nonperformance but before a lawsuit, lies about his intent to cure might be held liable for fraud. (16) Litigation tactics that fall short of misrepresentation can violate provisions like Rule 11, which permits a court to sanction pleadings that are "presented for any improper purpose, such as to harass or to cause unnecessary delay or needless increase in the cost of litigation." (17) Finally, discovery rules, which are again backed by deterrence sanctions, impose affirmative duties to provide information the other side can use to show nonperformance.

The point of each of these mechanisms is to prevent defendants from obstructing plaintiffs' recovery of damages due. None, however, is specific to the contracts setting. Nor are most contractual in a more technical sense: they are not duties that the parties have contracted for, or could contract out of. Rather, they are mandatory or quasi-mandatory rules that apply in all cases, no matter what any contract between the parties says. (18)

The question of this Article is what the parties can do when they want more. In circumstances where these mandatory and quasi-mandatory no-obstruct rules provide insufficient protection, how might the parties supplement those rules with contractual duties to cooperate in recovery? (19) In answering that question, I will focus on a subclass of no-obstruct duties: duties to create, preserve, or share information about performance and breach. Examples include terms that require the promisor to create or preserve records, to inform the promisee of noncompliance, or not to hide breach. (20) Such duties make it easier to verify whether performance has happened and thereby increase the probability that the promisee will be able to prove any breach in court. In addition, such terms can reduce litigation costs (by reducing the cost of information to the promisee) and make performance more observable (by telling the promisee about it). (21) The thesis of this Article is that the law should allow parties to contract for effective no-obstruct duties of this sort, and that in order to do so, it must allow them to attach extracompensatory remedies to the breach of those duties.

Damage multipliers are another familiar solution to underenforcement. Where a contractual duty to cooperate adds another layer of regulation to increase the probability of performance, a multiplier aims to increase the penalty to the point at which the promisor gets the right risk-adjusted incentives (compensation multiplied by the reciprocal of the probability of enforcement). I will argue that effective duties to cooperate require an exception to the rules against penalties and punitive damages. But if we are going to make an exception to those rules, perhaps it should instead be to give parties the option of multiplying damages, a solution to underenforcement that is easier to adopt and to administer. For the moment I simply flag this objection, leaving its answer for Part IV, after I have developed a more detailed picture of what effective duties to cooperate would look like. But the basic answer can be stated now: damage multipliers are more costly than they first appear. In many cases, contracting for specific duties not to obstruct is the more efficient means of realizing the benefits of contractual liability.

B. Compensatory Remedies

The limitations of compensatory damages when it comes to deterring obstructive behavior are nicely illustrated by a memo that came to light in Alabama's recent suit against Exxon for nonpayment of oil and gas royalties. (22) In the early 1980s, Alabama entered into twenty-two oil and gas leases with Exxon on terms that were unusually favorable to the state, in that they prohibited Exxon from deducting many of the usual costs from its royalty payments. (23) As with most royalty contracts, the leases also required that Exxon provide royalty reports, documenting the basis for its payments. (24) In 1999, Alabama sued Exxon for underpayment of royalties. The jury found that in 1993 Exxon formulated a plan to take prohibited cost deductions. (25) In a memo to a senior vice president, executives argued that Alabama's "inexperienced regulatory staff and processes," which were already engaged in a complicated audit of Shell Oil, were unlikely to detect the proposed underpayments. (26) Moreover, they argued, Exxon's exposure if caught would be only the underpayment plus twelve percent annual interest. (27) Exxon apparently concluded that falsely reporting royalties was "a no-lose proposition." (28) As the trial court explained:
   The downside associated with any detection of its underpayment was
   nonexistent from Exxon's perspective: it was certain that the very
   most it would have to pay the State would be what it already owed,
   the amount of its underpayment.... [E]ven if the State sued and
   won, Exxon knew that its return on the monies it withheld would in
   any event substantially exceed the 12% simple interest penalty it
   might ultimately have to pay. Thus, Exxon's scheme was tantamount,
   it thought, to a cost-free, risk-free option. (29)


The familiar point here is this: where enforcement is imperfect, compensatory damages do not sufficiently deter breach, especially opportunistic breach. But there is a deeper lesson as well. Exxon breached not only its duty to pay royalties, but also the duty to provide accurate accountings. That reporting duty was presumably put in place to guard against, and ensure recovery for, any breach of the underlying duty to pay. While compensatory damages provided too little deterrent against underpayment, they provided no deterrent against false royalty reports. Alabama could show that the royalty reports were false only if it could also demonstrate underpayment. But if Alabama could show underpayment, it could recover for the first-order breach on that basis, and the false reports caused it no actual harm. In fact, Exxon's defense team attempted to rely on just this logic, arguing that "although Exxon withheld tens of millions of dollars in royalties owed under the leases, the State suffered zero detriment [as a result of the false reports, because] the harm from its underpayments is now 'fully compensated through the contract damages.'" (30) So long as recovery for breach of the reporting duty required a proof of actual harm, Exxon would be off the hook.

The trial court escaped the dilemma by finding a substantial basis not only for compensatory breach of contract damages, but also for a punitive award based on Exxon's fraud. (31) (That decision is currently on appeal, and is awaiting a ruling by the Alabama Supreme Court. (32)) Part III discusses how this fraud solution can work. But before getting there, or even to contractual solutions, we need a more general understanding of the limitations of compensatory damages when it comes to securing effective duties to cooperate in recovery. That understanding has three components. The first is the catch-22 Exxon relied on: the proof structure of compensatory damages for obstructive breach is such that a plaintiff can prove actual harm only if she has not suffered it. The catch-22, however, only applies to the most obvious obstructive harm: the promisee's inability to recover for first-order breach. It is also necessary to show that the problem is not solved by recovery for other losses, which include increased litigation costs, delayed recovery, and inability to mitigate. Finally, there is the possibility that liquidated damages are a solution. I conclude that so long as liquidated damages are subject to the rule against penalties, they do not solve the problem.

1. The Catch-22

A more detailed account of the catch-22 begins with a distinction between two types of obstructive behavior. Recall that in my artificial use of the term, obstruction includes both affirmative acts (such as misrepresentations about performance or the destruction of records) and failures to act (not informing the promisee that a breach has occurred or failing to keep records, where there is a duty to do one or the other). Cutting across this distinction between active and passive obstruction is a difference between how obstruction is proven in court. I will call obstruction "independent" if it can be demonstrated without first showing that there was a first-order breach. Other obstruction is "dependent," which is to say capable of proof only by first showing breach of the underlying duty.

An example will illustrate the difference and be useful in the analysis that follows. Suppose Developer is deciding whether to hire Contractor to install high-speed network wiring in a residential building she is constructing. Developer is concerned about underenforcement. In particular, she is worried that nonconforming work may be both unobservable, because Developer does not have the necessary expertise, and unverifiable, because the wiring will be built into the walls and it will not be worth tearing the building apart to prove breach. Contractor, aware of Developer's concerns, suggests adding a recordkeeping clause, which will require Contractor to keep detailed records of materials and work performed.

There are two relevant ways Contractor might breach the recordkeeping clause. First, he could fail to keep any of the required records. That breach would, in my terminology, be a form of independent obstruction, for Developer can show that Contractor did not keep the records whether or not she can show faulty installation. Other examples of independent obstruction include the refusal to allow audits, nonprovision of required performance reports, and destruction of evidence. Alternatively, Contractor might breach the recordkeeping clause by falsifying the required records to cover up the faulty installation. In order to demonstrate that obstructive breach, Developer would have to show that the records were false, that is, that the installation was in fact noncompliant. Falsifying records is generally a form of dependent obstruction, for its proof depends on a showing of first-order breach. Among the other varieties of dependent obstruction are not reporting a first-order breach, hiding it, and providing false performance reports.

The distinction between independent and dependent obstruction marks a difference between types of obstructive behavior, not between no-obstruct duties. Thus breach of the recordkeeping clause is either independent or dependent, depending on the manner of noncompliance. The breach of some duties to cooperate, however, is necessarily dependent. Suppose Contractor agrees instead to a reporting clause, which requires that he immediately disclose any nonconforming work. That duty has noncompliant installation as its condition precedent. To prove obstructive breach--failure to report--Developer must first prove satisfaction of the condition precedent, which is a first-order breach.

With these distinctions in hand, I now turn to the functioning of unliquidated compensatory damages for the most salient harm of obstructive breach: the promisee's inability to recover for first-order nonperformance. (The next Section will consider the effects of compensation for other harms.)

Consider the situation where Contractor has falsified required records, which is a dependent obstructive breach (proving the records false presupposes showing noncompliant installation). Either Developer has separate sufficient evidence of the first-order breach or she does not. If Developer has the evidence, she can prove that Contractor breached his duty to install. She can therefore recover for the first-order breach, and the falsified records have not caused her the relevant loss. If Developer cannot prove first-order breach, then (assuming Contractor breached the duty to install) Developer has been harmed by Contractor's failure to keep accurate records. Such records would have demonstrated the noncompliant installation and permitted recovery for the first-order breach. But without proof of noncompliant installation, Developer cannot prove that the records are false--that is, she cannot prove obstructive breach. Either way, breach of the duty to cooperate in recovery costs Contractor nothing.

The recordkeeping clause requires Contractor to take affirmative steps to assist Developer's recovery for the first-order breach. But the result is the same for the dependent breach of no-obstruct duties that prohibit bad behavior. Thus suppose the parties adopt instead a no-conceal clause, requiring that Contractor not attempt to hide noncompliant installation. The duty not to conceal applies only if Contractor is in first-order breach, so breach of the no-conceal clause is necessarily dependent. If Contractor is deciding whether or not to cover up his faulty work, the threat of compensatory damages will not affect his decision. Should his attempt to hide the first-order breach succeed, he will not be held liable for the obstructive breach; should it fail, Developer then has independent proof of the noncompliant installation, and the salient harm of obstructive breach did not occur.

The general and inherent limitation of compensatory damages for dependent obstructive breach is this: if the promisee cannot demonstrate first-order breach, she cannot prove the dependent obstructive breach and recovers nothing for the obstructive breach. If the promisee can demonstrate first-order breach, she can recover for the first-order breach on the basis of that demonstration alone. Because the obstructive breach has not caused her the salient harm, there is no loss to compensate, so again she recovers nothing for the obstructive breach. So long as the remedy is limited to compensatory damages, the duty to cooperate makes no practical difference.

The same result holds for independent obstruction, though here the problem is not proof of the obstructive breach, but proof of proximate harm. Suppose Contractor breaches the recordkeeping clause not by falsifying records, but by failing to keep them. In that case, it may be easy for Developer to prove obstructive breach, in part because proof of no records does not depend on a showing of noncompliant installation. But to recover compensatory damages, Developer must also be able to show that the obstructive breach harmed her. On the present assumptions, this means showing that Contractor's failure to keep records prevented her recovery for a first-order breach. Again there are two possibilities. If, despite Contractor's failure to keep records, Developer discovers and can prove nonconforming installation, Developer can recover for the first-order breach. In that case, Developer has not suffered the salient obstructive harm, and, so long as damages are limited to compensatory measures, there is no recovery. If Developer cannot prove nonconforming installation, then she cannot prove that but for the failure to keep records, she would have recovered for a first-order breach. Again there is no recovery for the obstructive breach, now because she cannot show that the obstruction caused her the relevant loss.

This second version of the catch-22 is even more general. To show the most salient harm of an obstructive breach, dependent or independent, the promisee must show that, but for the obstruction, she would have recovered for the first-order breach. That is, she must prove that she has a meritorious claim for breach of the underlying duty. But this is precisely what she cannot do if the obstructive breach was successful. The result is a win-win situation for the promisor. If his obstructive breach succeeds (that is, the promisee cannot prove first-order breach), the promisee cannot prove proximate harm and so recovers nothing. If the obstructive breach is unsuccessful, the promisee can prove first-order breach and so suffers no compensable loss. Once again, there is no possible world in which compensatory damages for the obstructive breach make a practical difference.

2. Other Losses: Avoidable Harms, Delayed Recovery, and Litigation Costs

The catch-22 I have identified applies to unliquidated compensatory damages for the most salient and potentially costly harm of obstructive breach: the promisee's inability to recover for first-order breach. But obstruction not only decreases the chance of first-order recovery, it also increases its cost. The promisee will have to spend more to discover and prove first-order breach, and recovery may be delayed, costing the promisee the earlier use of the funds. And obstruction can make first-order breach not only more difficult to verify (by a legal factfinder), but also to observe (by the promisee), preventing her from taking early actions to avoid loss. Increased litigation costs, delayed recovery, and the inability to mitigate do not redound to the benefit of the promisor. Forcing him to internalize those secondary costs of his obstructive breach might therefore tip the scales against the benefit he expects to receive from avoiding liability, thereby deterring obstructive breach. If compensation for the most salient harm of obstructive breach, inability to recover, makes no practical difference, compensation for its secondary harms might.

Parties need not add no-obstruct terms to their contract to recover for the secondary harms of obstruction. Fee-shifting and prejudgment interest clauses make the promisor found to have breached liable for all of the promisee's litigation and delay costs. And the mitigation rule ("a party cannot recover damages for loss that he could have avoided by reasonable efforts" (33)) forces the promisor to internalize the costs of withholding information about breach: the longer the promisee remains in the dark, the longer she will be unable to take reasonable efforts to avoid losses, and the greater the promisor's liability.

My analysis of the parties' interest in securing incentives against obstruction of recovery casts new light on these familiar mechanisms. Proponents of fee shifting, for instance, often emphasize its value as a deterrent to frivolous lawsuits. (34) Less commonly noticed is the flip-side: fee shifting also deters obstruction, since the more the defendant obstructs recovery, the greater the plaintiff's litigation costs, which the defendant will have to pay if he loses. Similar arguments apply to prejudgment interest awards and the mitigation rule. In addition to their other benefits, these mechanisms encourage cooperation in recovery.

In fact, such mechanisms have an important advantage over more narrowly tailored no-obstruct terms. Contractually specified duties not to obstruct function only if the parties can identify, at the time of formation, specific behaviors likely to affect the probability of recovery. Fee shifting, prejudgment interest, and the mitigation rule, on the other hand, capture all obstructive behavior, whether identified in advance or not.

There are therefore two relevant questions. First, do compensatory damages for the breach of tailored no-obstruct terms add anything to these existing mechanisms? Second, can compensation for the secondary harms of obstruction--whether in the form of compensatory damages for obstructive breach or pursuant to more familiar means--provide the promisee sufficient protection against obstructive behavior?

The one place where tailored no-obstruct duties provide a clear advantage over other modes of recovery is the extra litigation expenses caused by independent obstructive breach. (35) Fee-shifting clauses typically provide for the recovery of attorneys' fees and other costs only if the promisee can show first-order breach. But a promisee might prove an independent obstructive breach even if she cannot prove the first-order breach. If she can recover litigation costs solely because of the obstructive breach, the no-obstruct term provides greater protection than a fee-shifting clause. Thus suppose Contractor has failed to keep any of the records required by the recordkeeping clause. That obstructive breach will increase Developer's recovery costs--Developer has to invest more to confirm Contractor's performance, or to discover and prove noncompliant installation. While a fee-shifting clause allows Developer to recover these costs only if she proves first-order breach, compensatory damages for the obstructive breach should mean Developer recovers them whether or not she can show noncompliant installation. (36) Because proof of no records is independent of and more likely than proof of the underlying breach, litigation cost recovery for the obstructive breach provides greater assurances against that form of obstruction than a simple fee-shifting provision would.

Compensation for the other secondary harms of obstructive breach--delay in recovery and inability to mitigate--is in all cases more effectively accomplished using prejudgment interest clauses or the mitigation rule, regardless of whether the obstruction in question is dependent or independent. Compensation for delayed recovery presupposes proof that recovery was warranted, which is to say, a showing of first-order breach. Thus delayed-recovery damages for obstructive breach provide no greater protection than a simple prejudgment interest clause, which compensates for all delays upon proof of first-order breach. Recovery for losses that could have been avoided but for the obstructive breach similarly presupposes proof that those losses are attributable to first-order breach. If the promisee can show first-order breach, the mitigation rule already permits recovery for those losses, since ex hypothesi they were not avoidable. A similar argument applies to litigation costs resulting from dependent obstructive breach.

This answers the first question: compensatory damages for the secondary harms of obstructive breach provide at best only slightly more protection (litigation costs for independent obstructive breach) against obstruction than more familiar and broadly effective mechanisms like fee shifting, prejudgment interest, and the mitigation rule. The remaining question is this: How much protection does liability for these secondary harms, whether based on specific no-obstruct terms or on the generic mechanisms, provide? Is liability a sufficient deterrent against obstruction to protect the promisee against obstruction's more salient harm--the inability to recover for first-order breach?

Common sense, and examples like Exxon's relative indifference to the twelve percent interest charge on unpaid royalties, (37) suggest that the answer is often "no." Forcing the promisor to internalize the costs to the promisee of delayed recovery, more complex litigation, and the inability to mitigate is not enough. In many cases, these costs of obstruction will be outweighed by the potential payoff: avoiding first-order liability. This commonsense judgment is based on three facts. First, because the promisor generally pays for the secondary harms of obstruction only if the promisee can show first-order breach (the only exception being litigation costs for independent obstructive breach), he discounts those costs by the chance of no first-order liability, an eventuality that obstruction makes all the more likely. As a result, the promisor internalizes a diminished portion of secondary costs of his obstructive tactics. Second, should the promisor be held liable, the added cost of his obstructive behavior will not be all of the promisee's secondary costs, but only those attributable to the obstruction. For instance, litigation-cost recovery does not force the promisor to pay for all of the promisee's attorney fees, but only for fees caused by the promisor's obstructive behavior. These figures are likely to be orders of magnitude less than the damages for the underlying breach that the promisor might hope to avoid. Finally, the potential upside of obstruction--the increased chance that the plaintiff will not recover and the defendant will not have to pay anything--can be very large. This is so when the remedy for the underlying breach is costly or there is a high probability that the obstruction will succeed. The latter case is particularly relevant, since effective obstruction both lowers the likelihood that the promisor will have to pay for its secondary harms (the first point above) and means a big gain in terms of avoiding first-order liability. In a great many cases, therefore, compensatory damages for the secondary harms of obstructive breach provide insufficient protection against obstruction's primary harm, no recovery for first-order breach.

3. Liquidated Damages for Obstructive Breach

The conclusion so far: unliquidated compensatory damages for obstructive breach will, in many cases, do little or nothing to deter the promisor who would otherwise obstruct recovery. The primary reason for this is the catch-22 embedded in the proof of harmful obstructive breach. To prove the most significant harm of the breach (that the obstruction prevented recovery), and sometimes even to prove that there was a breach (in the case of dependent obstruction), the promisee must be able to show a first-order breach; but if she can do that, she can recover on the basis of that showing alone, and the obstructive breach did not cause her that harm. While recovery for secondary harms (increased litigation costs, delayed recovery, inability to mitigate) provides some measure of deterrence, we can predict that in many cases it will not be enough to assure a promisee who is worried that the promisor will not cooperate in recovery.

Liquidated damages relieve the plaintiff of the burden of proving harm. Given that the catch-22 turns in part on the promisee's ability to prove actual loss, it is natural to ask whether a liquidated damages clause cannot resolve it. Closer examination will show that, like recovery for secondary harms, liquidated damages can provide some additional protection. But contractually specified damage amounts are limited by the compensation principle, in the form of the rule against penalties. As a result, liquidated damages cannot do the whole job. (Section III.C shows that penalties can be effective against obstruction.)

To analyze how compensatory liquidated damages for obstructive breach might work, let me return to Contractor's agreement with Developer to install network wiring. Assume now that their contract liquidates damages for first-order breach at $10,000, which represents the parties' reasonable estimate of the replacement cost for defective wiring. Suppose further that the contract includes the recordkeeping clause, and that Developer and Contractor want to liquidate damages for its breach as well.

The rule against penalties requires that damages be liquidated "at an amount that is reasonable in the light of the anticipated or actual loss," and states that a "term fixing unreasonably large liquidated damages is unenforceable ... as a penalty." (38) So, in seeking out an enforceable liquidated damage amount, the parties must predict what breach of the recordkeeping clause is likely to cost Developer. Putting aside the secondary harms discussed in the previous section (which often pale in comparison to the inability to recover for first-order breach), breach of the recordkeeping clause will harm Developer only if two conditions are met: Contractor has breached the agreement to install, and the missing or false records prevent Developer from proving first-order breach. If both conditions are satisfied, Contractor's obstructive breach costs Developer $10,000--the amount she would have recovered in liquidated damages had she been able to show noncompliant installation. If either condition is not satisfied--if Contractor has performed the installation correctly, or if Developer can prove noncompliant installation despite the missing or falsified records--the obstructive breach costs Developer nothing.

So, at what amount should the parties liquidate damages? If an obstructive breach causes Developer a significant harm, it will be in the amount of $10,000--the first-order liquidated damages amount. But it is hard to imagine a situation in which a court would award $10,000 for the faulty records. If Developer can show improper installation despite the obstructive breach, she can already recover $10,000 in first-order damages, and the $10,000 in liquidated damages for obstructive breach would mean double recovery. If Developer cannot show improper installation, the court is likely to balk at so large an award where there is proof of nothing more than faulty recordkeeping.

To satisfy the compensation principle, Developer and Contractor will therefore choose some lower amount, say $1,000, and stipulate that it represents the cost to Developer of the decreased chance of proving first-order breach--in mathematical terms, the expected percentage-point decrease in probability of enforcement, multiplied by the value of a first-order damage award. Now suppose Developer can prove breach of the recordkeeping clause, say by showing that Contractor failed to keep any records, though she cannot prove noncompliant installation. A court might still enforce the $1,000 in liquidated damages, reasoning that the $1,000 amount is a reasonable risk-adjusted forecast of damages and allowing for the possibility that there is an unverified first-order breach. The $1,000 liquidated damages amount will protect Developer against obstructive breach, however, only if it correctly captures the effect of improper recordkeeping. If failure to keep the required records decreases the probability of enforcement by ten percentage points or less, the $1,000 liquidated damages amount is enough. (39) But if the obstruction is more effective, liquidated damages will have to be increased. And the greater the liquidated damage amount, the less likely the court is to enforce it where the promisee cannot show any actual loss. If no records will result in, for example, a fifty percentage-point decrease in the likelihood of enforcement, liquidated damages will have to be at least $5,000 (40)--an amount that begins to look like a penalty for what, as far as the court knows, may be no more than a bookkeeping error.

The underlying problem is that courts find it difficult to assess the reasonableness of damages where the very existence of a loss (not just its amount) depends on nonverifiable facts--in this case, whether or not there was a first-order breach. As a result, liquidated damages must be reduced to take account of the possibility, in the court's mind, of first-order performance. But the promisor deciding whether to obstruct recovery is likely to know whether there has been or will be a first-order breach. And this means that the reduced liquidated damage amount will not be enough to convince him to cooperate in recovery.

This is not to say that liquidated damages are not an improvement in cases where the promisee cannot show first-order breach. Where unliquidated compensatory damages result in no recovery (the catch-22 of proving damages), liquidated damages can impose some cost on and provide some insurance against obstructive breach. Nonetheless, the penalty rule will tend to push enforceable damage amounts below what is needed to fully deter or insure against obstruction.

The last hypothetical, however, describes a situation where the promisee cannot prove first-order breach but can show obstructive breach, which is possible only where the obstruction is independent (such as failure to keep records). What about where the promisee can demonstrate both?

Many courts follow the Second Restatement rule that "[i]f ... it is clear that no loss at all has occurred, a provision fixing a substantial sum as damages is unenforceable." (41) The reasoning is this: if it is certain that the promisee has suffered no loss, liquidated damages fail their essential purpose, for the amount of actual loss is not difficult to prove. (42) Now suppose Developer can prove that Contractor failed to keep the records required by the recordkeeping clause, but she can also show noncompliant installation. If the court follows the Second Restatement, Contractor has a good argument against any liquidated damages award for his obstructive breach. Proof of noncompliant installation entitles Developer to the $10,000 first-order liquidated damages amount, from which it follows that the obstructive breach did not significantly harm her.

This last dynamic is fatal when it comes to dependent obstructive breach--where proof of obstruction requires first showing a first-order breach. If an obstructive breach is dependent, the plaintiff can show obstructive breach only when the defendant can show that that breach did not cause the plaintiff a significant loss. In jurisdictions that follow the Second Restatement rule, liquidated damages for dependent obstructive breach should never be enforceable. Any liquidated damage award for obstruction--be it $1,000, $100, or $10--is a pure penalty.

This may even be the case in jurisdictions that do not follow the no-loss, no-liquidation rule, though the small number of decisions makes it difficult to be certain. Williston suggests that whether a court will award liquidated damages where there is proof of no loss depends on the extent to which the jurisdiction considers actual losses (as opposed to anticipated losses) in evaluating the reasonableness of the liquidated damages amount. (43) But in the case of dependent obstructive breach (and recall that the breach of some no-obstruct terms is always dependent), we get the same result whether reasonableness is evaluated ex ante or ex post. Unlike other sorts of breach, we can say in advance that if a plaintiff can prove dependent obstructive breach, the defendant will be able to show no significant loss. Even from the ex ante perspective, liquidated damages for dependent obstruction will be available only where the amount of loss is certain--namely, where it is demonstrably zero. Liquidated damages for dependent obstructive breach violate both the rule against penalties and the rule that damages may only be liquidated when actual loss is difficult to prove.

The above observations about when courts are likely to enforce liquidated damages attached to duties to cooperate are based not on decisions considering such terms, but on the rules that govern liquidated damages generally. It might well be that, if presented with the above arguments, a court might relax the rules against penalties and proof of actual loss so as to allow the recovery of higher liquidated damages for obstructive breach. Such an outcome, however, would also mean a retreat from the compensation principle, confirming the broader thesis of this Part.

The above analysis has shown how the compensation principle systematically pushes damage amounts below the level necessary fully to deter or insure against obstructive breach. The core cause is the catch-22 that inheres in proof of actual loss. Obstruction harms the promisee first and foremost by decreasing the probability of first-order recovery. But a plaintiff who can prove that harm has not suffered it: to demonstrate the harm, the plaintiff must show that but for the defendant's obstruction, she would have recovered for first-order breach; if she can prove that, she can recover on that basis alone, and the obstructive breach did not cause her the harm. The proof structure that gives rise to the catch-22 infects other modes of compensation as well. Recovery for the secondary harms of obstructive breach--additional litigation costs, delayed recovery, and otherwise avoidable losses--provide some measure of deterrence. But in many cases it will not be enough to dissuade a promisor who would otherwise choose to obstruct recovery, in part because recovery for secondary harms often depends on proof of first-order breach--proof that obstruction makes all the less likely. Liquidated damages, too, can partly avoid the catch-22. But the rule against penalties significantly diminishes their utility. Where there is no proof of first-order breach, liquidated damages for obstruction can appear unduly harsh, and where there is such proof, liquidated damages create double recovery and violate the no-loss, no-liquidation rule.

It would be wrong to draw too strong a conclusion. Recovery for secondary harms and liquidating damages can provide some protection against obstruction. And for some parties that is enough. This could be so because the cost of liability for secondary harms or the enforceable liquidated damages amount is greater than the promisor's expected gains from obstructing recovery (where, for example, obstruction has a low probability of success). It could also be so where there are other, nonlegal reasons not to obstruct. In relational contracts the threat of no future dealings might alone be enough to ensure cooperation. And in industries that are highly organized or composed of repeat players, reputational or other nonlegal sanctions can deter obstructive behavior. But where there is a significant threat of obstruction and these extralegal protections are not available, the parties are likely to want more than contract law presently provides them.

II. EXTRACOMPENSATORY CONTRACT REMEDIES

The analysis in Part I has shown that, in many cases, effective duties not to obstruct require extracompensatory remedies. A contract remedy is extracompensatory if its application is not tied to the magnitude of harm caused. The extralegal sanctions mentioned above--no future dealings and loss of reputation--are extracompensatory in just this sense. (The reputational costs of obstructing recovery, for instance, might be much greater than the promisee's losses.) This Part describes three sorts of extracompensatory legal remedies the parties might apply to obstructive breach. (44) The first is termination of the contract, which is effective only when the promisor has not yet realized the contract's entire value, and then only when the promisee does not already have the right to terminate. The second is a contractually specified adverse inference, where the parties stipulate that proof of obstruction shall tip the evidentiary scales for a showing of first-order breach. Adverse inferences can be effective against independent obstructive breach, but for obvious reasons will make no difference where proof of obstruction presupposes proof of first-order breach. The third category of legal sanctions is penalties and punitive damages, which, I will argue, are the only generally effective contract remedies against obstructive breach.

A. The Right To Terminate

In some cases, it can be enough if an obstructive breach gives rise to the right to terminate, or rescind, the agreement. (45) The termination remedy is extracompensatory, since it does not vary with or depend on the magnitude of the obstructive harm. Where the promisor values the promisee's continued performance more than the potential benefit of obstructing recovery, termination can be effective. This is confirmed by evidence from the recording industry. Recording contracts are usually drafted by and often favor record labels. And while they typically require royalty accounting statements and give the artist a right to audit (duties to cooperate), they often stipulate that a breach of those duties shall not be material. (46) If one believes artists' claims that labels systematically underreport royalties and obstruct audits, the explanation for the latter provision is obvious. (47) A material breach gives the nonbreaching party the right to terminate, (48) and a termination remedy would deter labels from breaching these duties to cooperate--good from the perspective of enforcement, but bad from the point of view of the drafter and dominant party in the transaction.

Richard Craswell provides a general description of how the right to terminate can provide a form of property-rule protection against certain contractual wrongs. His primary concern is how the law can prevent duress, fraud and other acts that interfere with consent, but the basic point applies more broadly.
   One way to [deter bad acts] is to not enforce any obligation
   whatsoever on behalf of any X who has failed to obtain Y's proper
   consent. For example, if X forces Y to purchase her goods at
   gunpoint, the court can allow Y to rescind the contract entirely.
   This outcome may deter X from using duress because it denies X any
   profit from her dealings with Y, unless she properly obtains Y's
   consent. (49)


The same holds true for obstructive breach. Where the promisor values the continued existence of the contract more than he expects to gain from obstructing recovery, and where the obstructive breach is likely to be detected and demonstrated and will give the promisee the right to terminate, the promisor will choose to cooperate in recovery. That is, a termination remedy will be most effective where the obstructive breach is easily observed and verified (such as refusal to permit an audit) and the promisor expects to continue extracting value from the contract (as in long-term recording agreements).

These conditions are not always satisfied. Most obviously, if the transaction is a one-time deal and the promisee's first-order performance is before the promisor's, the termination remedy is no disincentive against obstructive breach. More generally, the effectiveness of the termination remedy will depend on the contract's remaining unrealized value to the promisor, the likelihood that his obstructive breach will be discovered and proven, the cost of the first-order liability he wishes to avoid, and so on. Termination is, in Craswell's terms, the "minimum sanction"--in some cases a "penalty heavier than mere nonenforcement is needed." (50)

There is, however, another aspect of the termination remedy that is specific to duties to cooperate. Where the promisee has separate proof of a material first-order breach, she might already have a right to terminate. A contract can be cancelled only once, so in such cases a termination right for obstruction is redundant. In some cases it is much easier to show obstructive breach than it is to show first-order breach--proof of a record company's refusal to permit audits as compared to proof of underpayment. But this is not always so, and the more likely separate proof of first-order material breach, the more diluted the effect of the right to terminate for obstruction. And because proof of dependent obstruction always requires separate proof of first-order breach, the termination remedy for obstructive breach is worth nothing. The catch-22 replicates itself here, now in the form of redundant remedies.

B. Adverse Inferences

If the point of obstruction is to prevent proof of first-order breach, one obvious remedy for obstructive breach is an adverse inference. Thus the parties might decide in advance that proof of obstructive breach shall have a certain evidentiary value with respect to first-order breach by agreeing to a conditional presumption, admission, or other burden-shifting device. (51) Such a remedy is extracompensatory, since it does not depend on proof of actual harm. But rather than a monetary penalty, the deterrent is evidentiary.

At present there is little law on when or to what extent parties can contract for atypical burdens of proof or standards of evidence that will govern the contract's enforcement. Probably the most informed guess is that of Scott and Triantis, who report: "While we have not found direct authority, we believe that courts would enforce reasonable contractual burden of proof provisions. And, we have found ample evidence that many contracts in fact contain such provisions." (52) All of the contracts Scott and Triantis describe, however, unconditionally modify the otherwise applicable burden of proof or standard of evidence. None condition the change on a party's bad behavior, such as the defendant's attempt to obstruct recovery. (53)

While there is little law in the area, courts would likely be sympathetic to such contract terms. The party-specified remedy--a change to the way breach is proven--is closely related to the wrong--obstruction of recovery. Moreover, one can find analogs elsewhere in the law. For example, the rules governing the destruction or spoliation of evidence provide for a presumption that the lost evidence was unfavorable to the party at fault. (54) Equitable principles of estoppel might be applied to reach a similar result, the argument being that the defendant should not be permitted to benefit from obstructive behavior.

The threat of an adverse inference will fully deter obstructive breach only when the inference increases the probability that the promisor will pay damages for the underlying breach more than the obstruction reduces that probability. The former depends on two factors: first, the probability that the promisee will be able to show an obstructive breach in circumstances where she would not otherwise (absent the adverse inference) be able to demonstrate first-order breach; second, the strength of the inference--whether the obstruction is to be conclusive evidence of a first-order breach, to shift the burden of proof, or simply to be placed on the scales with everything else. An adverse inference is most likely to work where it is easier to detect obstructive breach than it is first-order breach, and the inference is a strong one.

As an example, recall the hypothetical recordkeeping clause. Contractor's failure to keep the required records is easy to detect--much easier than noncompliant installation. Assume that the recordkeeping clause also specifies a strong adverse inference as its remedy: improper records shift the burden of proof, so Contractor must show compliant installation. Finally, recall my assumption that it is very difficult to prove breach or performance after the wiring is installed. Faced with the prospect of an adverse inference, Contractor may well choose to keep the required records. The benefit of hiding breach is not worth the increased chance that he will be found liable as a result of the adverse inference.

The reader might object that, faced with the prospect of an adverse inference, Contractor would never breach in so obvious a way as failing to keep records. If he wants to avoid liability for noncompliant installation, he will falsify the records, which is much more difficult to detect. This objection is correct and shows why adverse inferences are at best a partial solution. But the difference between not keeping records and keeping false records is not only a difference in the degree of observability or verifiability of each.

An adverse inference works only if there are possible states of the world in which (a) the promisee cannot otherwise prove the first-order breach, (b) she can prove obstructive breach, and (c) the resulting adverse inference tips the balance with respect to proof of the first-order breach. Conditions (a) and (b) jointly entail that the obstructive breach is an independent one--that it is possible to prove the obstruction without first proving first-order breach. The adverse inference is ineffective against falsification of records not because of the low probability of detection or proof, but because that breach of the recordkeeping clause is a form of dependent obstruction. If Developer can show false records, she has already proven first-order breach, and the adverse inference makes no practical difference. The upshot is similar to the limitation of the termination remedy: while the threat of an adverse inference can deter some forms of independent obstruction--those that are easy to detect and verify, and where the adverse inference is likely to result in a finding of a first-order breach--it is ineffective against dependent obstruction.

C. Penalties and Punitive Damages: A New Argument

The only generally effective contract remedy for obstructive breach is a penalty or punitive damages clause. To be able to contract for meaningful duties against the full range of obstructive behavior, parties must be able to attach either extracompensatory liquidated damages (penalties) or court- or jury-assessed extracompensatory money payments (punitive damages) to the breach.

How large must a penalty or punitive damages be fully to deter obstructive breach? The answer is familiar: the lower limit is the promisor's expected gain from breach multiplied by the reciprocal of the probability of enforcement. (55) The promisor's expected gain from obstruction is the reduced chance that he will have to pay first-order damages. More specifically, and assuming that the promisor has breached his first-order duties, it is the percentage-point reduction in the probability of first-order enforcement multiplied by what first-order liability would cost the promisor. Thus if falsifying records decreases the probability of $10,000 in liability from eighty percent to thirty percent, the fifty percentage-point reduction in the probability of first-order enforcement is worth $5,000 to the promisor. (56) Deterrence requires the sanction be greater than that gain multiplied by the reciprocal of the probability that the sanction will be applied. To continue the example, if we assume there is a twenty percent chance that the promisee will discover and be able to prove false records, the standard punitive damage formula recommends a sanction greater than $25,000. (57)

Penalties and punitive damages avoid both prongs of the catch-22. There is no double-recovery problem in cases where the plaintiff has independent proof of first-order breach, for penalties and punitive damages are not meant to compensate. They therefore work equally well against independent and dependent obstructive breach (despite the fact that proof of the latter presupposes proof of first-order breach). Where the plaintiff cannot prove first-order breach, and therefore cannot prove actual harm, the penalty or punitive damages should still be awarded, as the remedy is not meant to compensate actual loss. (58)

At present, however, "[p]unitive damages are not recoverable for a breach of contract," (59) and "[a] term fixing unreasonably large liquidated damages is unenforceable on grounds of public policy as a penalty." (60) The above analysis provides a strong instrumentalist argument for making an exception to these rules for contractual duties not to obstruct. (61) There are good reasons to think that, if given the option, most parties would not choose to attach penalties or punitive damages to first-order breach. Among other things, extracompensatory damages can interfere with efficient breach, reducing the overall value of the transaction. (62) When it comes to obstructive breach, however, extracompensatory damages can be the only effective remedy--there is no more efficient compensatory or subcompensatory alternative. If the parties want duties to cooperate in recovery, they probably have a good reason to want to attach a penalty or punitive damages to the breach of those duties.

I am not the first to recommend revising the existing mandatory rules against penalties and punitive damages. In fact, it is perhaps the majority opinion among efficiency theorists. But the above argument is new. To begin with, it does not depend on a commitment either to freedom of contract (the parties should be allowed to contract for whatever terms they wish) or to choice as the metric of value (if the parties choose it, it must be value-creating). (63) Nor does the argument rely on empirical assumptions about the capacity of the courts to distinguish penalties from reasonable estimates of harm. (64) Finally, the argument does not correspond to any of the transaction structures (all involving informational asymmetries or verification problems) in which extracompensatory damages have been shown to create efficient investment incentives for first-order duties. (65) Rather than identifying a category of transactions where extracompensatory measures are more efficient, the argument identifies a category of terms for which legal liability does not work unless the parties can choose penalties or punitive damages. There is also a practical difference. The basic insight behind the economic argument dates back at least to 1977. (66) Yet in the thirty years since, it has had little or no influence on lawmakers. The mandatory rules against penalties and punitive damages appear as vibrant today as ever. (67) The explanation might be that courts and legislatures pay too little attention to efficiency. Or perhaps there are other, noneconomic reasons for keeping the rules. Either way, it should be easier to convince courts to take a smaller step. Contractual duties to create, preserve, or share information so as to promote recovery for first-order breach constitute a well-defined, limited class of terms, where the reasons for allowing extracompensatory remedies are easy to grasp and where penalties or punitive damages do not threaten more general intuitions about the proper limits of contractual liability. Presented with the right arguments, courts might be willing to enforce a penalty or punitive damages clause for obstructive breach.

III. CONTRACTING FOR FRAUD LIABILITY

In the absence of effective contract remedies, legal actors have had to look elsewhere to secure cooperation in recovery. The first two Parts of this Article have provided an armchair analysis of why parties might want no-obstruct terms (when they are the cheapest way of addressing underenforcement) and how such terms can be effective (when backed by extracompensatory remedies that avoid the catch-22 of proving actual harm). This Part takes an empirical turn and describes how courts and parties have attempted to secure such duties in the face of the rules against penalties and punitive damages. In the absence of a contractual solution, they have turned to the extracompensatory remedies available in tort.

The most striking and potentially powerful approach lies in attempts to piggyback on the punitive damages available in fraud by contracting for representations about performance. I have already mentioned two examples: the Richmond stadium construction contract, which required the builder to certify that work was done to specification, and the Alabama oil and gas royalty contracts, which required Exxon to submit royalty reports. In both cases, compensatory damages for the obstructive breach (false certificates of compliance or false royalty statements) provided insufficient protection against obstructive breach as a result of the now familiar catch-22. And in each, the plaintiff invoked the law of fraud as an additional source of legal liability, albeit with varying degrees of success.

The potential value of fraud liability in these transactions shows that contract duties to represent performance are not just a place where fraud and contract happen to overlap. Fraud liability for the breach of such terms can play a positive role within the contracting relationship. This parallels a claim Ian Ayres and I have made about the utility of fraud liability for precontractual misrepresentations of intent. (68) But there is also something new here. While there has been a fair amount of academic discussion as to whether, when, and how parties should be permitted to contract out of liability for fraud, (69) my analysis indicates that some parties are trying to contract into fraud liability. Parties appear to be using the law of fraud to get the extracompensatory remedies for obstruction that contract law denies them. The existence of these terms is evidence that effective duties to cooperate are not only useful in theory, but sought-after in practice. But this salutary use of fraud faces its own doctrinal obstacles, most recently in the form of an expansive reading of the economic loss rule. Perhaps held captive by the theory of efficient breach, courts have overlooked this positive role fraud liability can play in contracting relationships.

Fraud liability for misrepresentations about performance is not the only way the punitive damages available in tort can be used to enforce contractual duties to cooperate. (70) Before examining the fraud option, I discuss California's brief experiment with punitive tort damages for obstruction of recovery, first announced in 1984 in Seaman's Direct Buying Service, Inc. v. Standard Oil Co. (71) While Seaman's is most often criticized as an example of the generic tort of bad faith breach, attention to the holding shows that it rightly singled out obstructive behavior for penalization. The real problem with Seaman's is not that it imported tort liability into the land of contract. The problem is its vague yet mandatory rule that punitive damages should be available for all postbreach obstruction, rather than a rule permitting parties to contract into or out of such remedies and also specify the relevant duty.

I then describe how fraud liability for misrepresentations about performance can correct for this defect, though many courts frown on this intrusion of fraud into contractual relationships. Yet a third technique for preventing obstruction uses the law of promissory fraud. Representations of intent not to obstruct, backed by the law of fraud, can provide a promisee some assurance that the promisor will cooperate in recovery. In the last Section of this Part, I argue that fraud liability, while potentially effective in securing some duties to cooperate in recovery, is a second-best solution.

A. An Early Foray: Rereading Seaman's Direct Buying Services v. Standard Oil

The analysis of Parts I and II sheds new light on California's brief experiment with the tort of postbreach bad faith behavior. In 1984, the California Supreme Court held in Seaman's Direct Buying Service v. Standard Oil Co. that "a party to a contract may incur tort remedies ["including punitive damages"] when, in addition to breaching the contract, it seeks to shield itself from liability by denying, in bad faith and without probable cause, that the contract exists." (72) That is, certain postbreach denials of the contract were grounds for punitive damages in tort. Seaman's sowed a fair amount of confusion in the lower courts, which arrived at a variety of standards for determining when a defendant's behavior was so wrongful as to constitute the tort. (73) Eleven years later, the California Supreme Court, citing inter alia the greater efficiency of compensatory measures, expressly overruled Seaman's and abandoned the experiment. (74)

Seaman's is often criticized as an example of the tortification of bad faith breach in general. This expansive reading is partly attributable to the case's procedural history. The plaintiff had asked the court to recognize the broader tort of bad faith breach, extending the rule from the insurance context to all commercial claims. And the majority left open the possibility of recognizing such a tort in a later case, with Chief Justice Bird arguing in a partial concurrence that "[a] breach of contract may also constitute a tortious breach of the covenant of good faith and fair dealing in a situation where the possibility that the contract will be breached is not accepted or reasonably expected by the parties." (75) Commentators therefore read Seaman's as representing more than it said--as putting California on the road to a general tort of bad faith breach. (76) That reading was amplified by the fact that contort was in the air in the mid-1980s, and by the perceived need to repel assaults on freedom of contract. (77) Judge Kozinski expressed the common reading and wisdom as follows:
   Nowhere but in the Cloud Cuckooland of modern tort theory could a
   case like this have been concocted. One large corporation is
   complaining that another obstinately refused to acknowledge they
   had a contract. For this shocking misconduct it is demanding
   millions of dollars in punitive damages. I suppose we will next be
   seeing lawsuits seeking punitive damages for maliciously refusing
   to return telephone calls or adopting a condescending tone in
   interoffice memos. Not every slight, nor even every wrong, ought to
   have a tort remedy. The intrusion of courts into every aspect of
   life, and particularly into every type of business relationship,
   generates serious costs and uncertainties, trivializes the law, and
   denies individuals and businesses the autonomy of adjusting mutual
   rights and responsibilities through voluntary contractual
   agreement. (78)


But whether or not Seaman's was part of a larger trend, what it said was relatively narrow: tort damages are appropriate where "a contracting party seek[s] to avoid all liability on a meritorious contract claim by adopting a 'stonewall' position ('see you in court') without probable cause and with no belief in the existence of a defense." (79) Stonewalling of this sort is, in my terms, obstruction of recovery. Read narrowly, the Seaman's rule does not impose a Cloud Cuckooland generic duty of good manners, but is supported by many of the above arguments for permitting the parties to adopt extracompensatory remedies for obstructive breach. Thus court-assessed compensatory damages alone neither insure against nor deter the sort of stonewalling at issue in Seaman's. Punitive damages or other extracompensatory remedies are necessary. (80) Moreover, where there is a measure of distrust and uncertainty at the outset of a deal, the parties themselves might rationally prefer Seaman's protection. A promisee who knows that the promisor faces a significant deterrent against unreasonably denying the existence of the contract knows she is more likely to recover in the case of nonperformance, and will therefore be more willing to enter into the transaction. Finally, Kozinski is simply wrong when he writes that Seaman's "denies individuals and businesses the autonomy of adjusting mutual rights and responsibilities through voluntary contractual agreement." (81) Parts I and II have shown that the rules against penalties and punitive damages already deny parties the ability to contract for meaningful duties to cooperate in recovery. Seaman's simply flipped the rule. Where before there was never meaningful legal liability for postbreach obstruction, Seaman's held that such behavior should always be subject to punitive liability in tort.

This does not mean that California should have kept the Seaman's rule, which suffered from serious defects. The experience of lower courts shows that Seaman's failed to define the wrong with sufficient precision. (82) The predictable result of the duty's vague definition was that the threat of punitive damages deterred defendants from raising legitimate defenses and caused plaintiffs to overlitigate minor claims in the hopes of big payoffs. (83) The California Supreme Court might have tried to fix this problem by clarifying the definition of "stonewalling" and by limiting it to highly verifiable behavior, so as to exclude both discretion and false positives. But replacing the "in bad faith and without probable cause" standard with a narrower test would have robbed the Seaman's rule of much of its utility. Because postbreach obstruction comes in endless varieties, generic narrow rules will invite circumvention and have limited effect. (84)

These difficulties result from a deeper problem with the Seaman's approach: the court's attempt to address obstruction of recovery by creating a new tort rather than by modifying the law of contracts. Because tort law provides one-size-fits-all solutions, the Seaman's court had to rely on its own understanding of "accepted notions of business ethics," (85) ignoring the industry- and context-dependence of such norms. More generally, why should courts attempt to fashion generic no-obstruct rules when the parties, if given the tools, can generate rules to fit their particular needs and circumstances? (86) Rather than the one-size-fits-all rules of tort law, the better approach lies within the contractual framework. The parties to a transaction know best what duties will increase the probability of enforcement. If they want to prohibit stonewalling or other forms of obstruction, let them specify both the relevant duty and the remedy for its breach.

B. Contracting for Fraud Liability

There is one species of tort liability and extracompensatory remedy that contracting parties might easily opt into: liability for fraud. A speaker exposes himself to fraud liability when he makes material representations on which the listener is likely reasonably to rely. If the parties can contract for such representations, they potentially can secure for themselves extracompensatory remedies for obstructive breach by way of the law of fraud. The idea is to tailor informational duties not to obstruct so that their nonperformance qualifies not only as breach, but also as deceit. The most obvious example is a requirement that the promisor say whether he has breached. A duty to inform about performance, if backed by the law of fraud, gives the promisor in first-order breach three options: honestly inform the promisee of nonperformance; decline to make the representation, breaching the no-obstruct duty and effectively communicating to the promisee that there may be a first-order breach; or falsely represent that he has performed, exposing himself to punitive damages for fraud. Timely information about performance is secured by the extracompensatory remedies available for misrepresentation.

In fact, piggybacking on the law of fraud has a potential advantage over liability in contract. While contracts are governed by principles of strict liability, liability for fraud requires proof of scienter--that the misrepresentation was intentional or reckless. (87) This protects promisors against large punitive damage awards for small mistakes in complying with their no-obstruct duties. The scienter requirement also provides some measure of assurance against false allegations of obstructive breach, since the promisee not only has to show nonperformance, but also intent or recklessness. (88)

While it is difficult to say from the outside exactly why contracts have the terms they do, there is evidence that parties attempt to use this strategy. Consider Alabama's gas and oil leases with Exxon, which required that Exxon represent that its royalty payments complied with the contract. The jury determined that Exxon both underpaid and intentionally misrepresented the basis for the payments. The trial court held that the first breach warranted compensatory damages, while the second supported punitive damages in fraud. (89) In Holmesian terms, Exxon had the right to breach and pay damages, but not to lie about it in contractually required representations.

This sort of fraud claim is perhaps most common in royalty disputes, but is not limited to them. (90) The city of Richmond's fraud claim against the contractor in McDevitt was based on false certificates of completion and other required documents. (91) In Life Insurance Co. v. Murray Investment Co., the purchaser of real estate notes claimed fraud based on contractually required quarterly reports that misrepresented progress on a construction project. (92) And in Robinson Helicopter Co. v. Dana Corp., a helicopter maker sued a supplier for fraud based on required certificates of compliance that falsely represented that parts met contractual specifications. (93) Robinson Helicopter in particular repays close study: while the facts and holding demonstrate the potential utility of fraud liability for obstructive breach, the opinions in the case illustrate the doctrinal hurdles this salutary use of tort liability faces.

Robinson Helicopter began purchasing clutch mechanisms from Dana Corporation for use in its helicopters in 1984. (94) Robinson's purchase orders required that shipments be accompanied by a Certificate of Compliance, which stated that the clutches met contractual specifications. (95) Dana regularly included the certificates, though clutches shipped between July 1996 and October 1997 were noncompliant. High failure rates resulted in the discovery of Dana's first-order breach, and Robinson sued for breach of contract, breach of warranty, and negligent and intentional misrepresentation. The California Supreme Court upheld the jury award of $1.5 million in compensatory damages for the breach of contract and warranty, and an additional $6 million in punitive damages, the latter based on the jury's finding that in providing the certificates, Dana had made false representations of fact and had knowingly misrepresented or concealed material facts with the intent to defraud. (96)

From the perspective of this Article, this was the right outcome. The certificates of compliance apparently corrected for two worries Robinson had going into the transaction. The first was the difficulty of observing certain sorts of breach. At issue in the case was the metallurgical integrity of the clutches, which was presumably difficult or expensive to detect. Second, if such a breach was not discovered early, Robinson would likely suffer significant noncompensable damages. Systematic clutch failures would result not only in possibly catastrophic liability, but also regulatory noncompliance, loss of reputation, and other nonrecoverables. Robinson might have attempted to capture some of these first-order harms in a liquidated damages clause. But there was the risk that a court would view the high liquidated damages amount as a penalty and therefore unenforceable. The parties therefore constructed a contract that permitted Dana the usual option of breaching and paying unliquidated compensatory damages, but effectively required that it immediately inform Robinson of such breach--by not sending the clutches or by not including the Certificate of Compliance. This no-obstruct duty to share information about performance was structured such that any injurious breach would involve misrepresentation. In this way, the parties were able to attach extracompensatory damages, by way of liability for fraud, to the most salient form of obstruction.

One can imagine other ways to structure a contract so as to secure cooperation in recovery via fraud liability. Recall the hypothetical reporting clause. Unlike the Robinson-Dana contract, in which performance was episodic, the imagined contract between Developer and Contractor envisions a continuous period of performance. (97) Rather than arbitrarily chosen daily, weekly, or monthly performance reports ("I am not in breach"), Developer might prefer that Contractor be required immediately to report any breach ("I have breached"), where failure to report would expose Contractor to liability for fraudulent concealment. (98) The reporting clause is cheaper than the certificate of compliance requirement in Robinson Helicopter. Instead of regular reports of compliance, there is at most one communication, and that only in the case of breach. This means cost savings both in the number of reports and in their value: a contemporaneous report of breach conveys to the promisor just the information she needs at the moment she needs it most. As a doctrinal matter, however, recovery for fraud might be more difficult under the reporting clause. While the false certificates of compliance at issue in Robinson Helicopter involved garden-variety misrepresentations (Dana said it performed when it had not), breach of the reporting clause qualifies as fraud only under a theory of fraudulent concealment. It appears to be an open question whether, as a doctrinal matter, fraud by silence applies when the duty to speak is purely contractual. (99)

There is, however, a larger obstacle to the whole project of incorporating fraud liability with contractual duties to represent: the tendency of courts to exclude tort liability from contractual relationships. The strong pull in this direction is exemplified by the various opinions in Robinson Helicopter and their uniformly expansive reading of the economic loss rule. Given that the American Law Institute is undertaking a revision of the Restatement's treatment of that rule, this reading deserves scrutiny. (100)

The economic loss rule "precludes a recovery in tort where the sale of a defective product has resulted in no property damage or bodily injury, but only economic loss to the buyer of that product." (101) The doctrine first emerged in the mid-1960s as a means of drawing the line between, on the one hand, breach of warranty and, on the other, strict product liability and negligence. (102) The first articulations of the rule held that so long as a defect did not cause the purchaser physical injury or property damage, she could recover only in contract--only for breach of warranty. Early cases did not decide whether the doctrine extended to intentional torts.

In the past decade, several courts have held that the economic loss rule bars liability in fraud between contracting parties except for "acts considered to be independent from acts that breached the contract." (103) Courts define "independent" in different ways. Many, including the intermediate appellate court in Robinson Helicopter, pay particular attention to whether the misrepresentation occurred before or after formation--to the distinction "between fraud in the inducement of the contract, which is not barred by the economic loss rule, and fraudulent misrepresentations relating to the performance of the contract, which are barred." (104) Fraud in the inducement, these courts reason, cannot interfere with contractually chosen risk allocation. When fraud in the inducement occurs, the contract does not yet exist, and fraudulent inducement means that the promisee's agreement to the contractual allocation of risk is not truly voluntary. (105) Misrepresentations in the course of performance, on the other hand, are understood to be "interwoven with the breach of contract," (106) so liability for fraud in the performance should be governed exclusively by the contractual allocation of risk.

This expansive reading of the economic loss rule prevents the use of liability in fraud to secure information-sharing duties that promote recovery. The relevant misrepresentations in these cases are necessarily "interwoven" with the breach, for the contract requires the promisor to make them truthfully. For example, the intermediate court in Robinson Helicopter concluded that Robinson "has neither claimed nor established any fraudulent representation or concealment that was not intertwined with Dana's performance of its contract and warranty breaches." (107)

The California Supreme Court rejected that holding, stating without further explanation that "the economic loss rule does not bar Robinson's fraud and intentional misrepresentation claims because they were independent of Dana's breach of contract." (108) The court avoided a broad ruling on whether or when the economic loss rule applies to fraud in the performance by limiting its holding "to a defendant's affirmative misrepresentations on which a plaintiff relies and which expose a plaintiff to liability for personal damages independent of the plaintiff's economic loss." (109) While less than crystalline, this statement suggests that the California Supreme Court agrees that the economic loss rule bars recovery for most fraud in the performance, or fraud that is intertwined with the breach.

There is considerable irony in applying the economic loss rule to defeat fraud liability in cases like Robinson Helicopter. One of the most commonly cited reasons for excluding tort liability from contractual relationships is that it interferes with the parties' chosen allocation of risk. As the U.S. Supreme Court has explained:
   The commercial buyer and commercial seller can negotiate a
   contract--a warranty--that will set the terms of compensation for
   product failure. If the buyer obtains a warranty, be will receive
   compensation for the product's loss, whether the product explodes
   or just refuses to start. If the buyer does not obtain a warranty,
   he will likely receive a lower price in return. Given the
   availability of warranties, the courts should not ask tort law to
   perform a job that contract law might perform better. (110)


This rationale makes sense when the tort at issue is negligence or product liability, both of which typically provide compensatory damages only. But it makes no sense when the wrong is fraud. The parties in Robinson Helicopter could not have accomplished with contract liability what they got with liability in fraud--extracompensatory damages for obstructive breach. Applying the economic loss rule in this context does not protect the parties' chosen allocation of risk, but defeats it. (111)

Despite its defects, California's potentially expansive reading of the economic loss rule agrees with the approaches of several other jurisdictions, including Michigan, Florida, and Wisconsin. (112) Nor is the economic loss rule the only bar to fraud liability between contracting parties. (113) In McDevitt, the Virginia Supreme Court ruled that "[i] n determining whether a cause of action sounds in contract or tort, the source of the duty violated must be ascertained," where a noncontractual source means no fraud liability. (114) The Nebraska Supreme Court has required that a fraud plaintiff allege "separate and distinct factual occurrences that could stand alone as a separate cause of action sounding in tort." (115) And Ohio courts have held that
   [i]n addition to containing a duty independent of that created by
   contract, an action arising out of contract which is also based
   upon tortious conduct must include actual damages attributable to
   the wrongful acts of the alleged tortfeasor which are in addition
   to those attributable to the breach of the contract. (116)


Each of these rules prevents the parties from using the law of fraud to contract for effective duties to cooperate in recovery. All are misguided according to this Article's analysis.

C. Promissory Representations of No Intent To Obstruct

There is another way a promisor might use fraud liability to assure the promisee that he will not obstruct recovery: by making a legally backed representation of intent not to obstruct. While precontractual representations of intent do not provide the assurance of a postformation representation of performance, they are of theoretical interest and can make a difference at the margins.

The doctrine of promissory fraud interprets every contractual promise as representing that the promisor intends to perform and imposes liability for fraud where that representation is false--that is, where the promise is made without an intent to perform. Punitive damages for promissory fraud add value to contracting relationships. (117) Where damages for breach are systematically subcompensatory, a promisor who intends to perform wants a credible means of sharing that information. A representation of intent to perform, backed by the law of fraud, gives it to him.

By the same token, if standard breach of contract damages cannot protect against obstruction (not because they are subcompensatory, but because the proof structure of proximate harm renders them largely irrelevant), perhaps credible representations of intent not to obstruct can. A promisee who is worried about obstruction presumably believes first-order breach is a real possibility--even if the promisor presently intends to perform. To convince the promisee that she will be able to recover, the promisor might represent not only that he intends to perform, but also that he intends to cooperate in recovery (by adequate recordkeeping, by informing her of breach, etc.) should the first intent fail. Such a representation can, if credible, assure the promisee that she will recover for any first-order breach. The law lends credibility to such representations of intent not to obstruct by imposing fraud liability when they are false.

Because the doctrine of promissory fraud holds that every promise represents an intent to perform, a promisor's agreement to a no-obstruct term represents that he intends not to engage in that form of obstruction--and should bring with it the added benefit of promissory fraud protection. Thus the Dixie Chicks recently alleged that Sony Music not only breached its contractual duty to "permit them to conduct meaningful audits to ascertain the accuracy of its accountings," but entered the contract intending that obstructive breach. (118) It is unclear what evidence the Dixie Chicks had to back up that claim, which was dismissed early in the litigation, (119) but one can imagine cases where a plaintiff could make the showing and recover punitive damages for the misrepresentation of an intent to cooperate.

There are, therefore, two ways to represent an intent not to obstruct recovery backed by the law of fraud: expressly ("I intend to keep such-and-such records."), and implicitly, as part of a promise not to obstruct ("Party A shall keep such-and-such records."). Yet a third way would exist if courts adopted an interpretive rule that, absent evidence to the contrary, every first-order promise represents not only an intent to perform, but also an intent to cooperate in recovery. That is, the law might stipulate that among the default representations embedded in a legally binding promise is a representation that the promisor does not intend to obstruct recovery should he fail to perform. This default makes sense given the law's general interest in the effectiveness of legal remedies. A generic default, however, might have Seaman's-like bad consequences. Seaman's imposed a duty so amorphous that parties did not know what was required of them. At the very least, one would want the content of the default representation of intent to cooperate spelled out in detail, with a clear and predictable definition of what counts as obstruction, so promisors know just what they are representing their intentions to be.

D. The Limitations of Liability in Fraud

So we have two ways parties might use fraud liability to circumvent the rules against penalties and punitive damages and contract for effective duties not to obstruct recovery. The first is by including in the contract a duty to make a representation about performance; the second is by representing an intent not to obstruct. These fraud solutions are, however, second best.

This is most obvious with respect to liability for misrepresentations of intent. While credible representations of intent to cooperate in recovery can provide some additional assurance at the outset of the transaction, such representations offer less protection than an effective contractual duty. For one thing, people change their minds. The promisor who now intends to cooperate in recovery may have a change of heart after breach, when he is faced with a choice between, say, falsifying records and paying a large damage award. There are also familiar problems of evidence. While it is sometimes possible to prove the promisor's bad initial intent--the smoking-gun internal memo outlining a plan to falsify records, repeated instances of similar behavior--the promisor's intent at the time of promising is typically much more difficult to prove than a mutually agreed upon contractual duty not to obstruct.

Opting into fraud liability with contractual duties to represent performance, in the form of certificates of compliance, performance records, or other required reports, provides a more practically powerful tool. But here too there are limits. Most obviously, fraud requires misrepresentation. Fraud liability is therefore of no use against obstruction that does not involve a lie, like not keeping records or even destroying records. Moreover, if contractual duties to speak cannot support claims of fraudulent concealment, parties will not be able to craft the duties to share information in the most effective or efficient way.

Fraud also has more elements than breach of contract. While I have argued that this is a good thing with respect to scienter, the reliance element can create an unwanted hurdle. In Hunt Petroleum Corp. v. Alabama, the Alabama Supreme Court found insufficient proof of reliance where there was evidence that the state intended to audit royalty statements, and it could not show that it changed its position based on the false reports. (120) In fact, one can imagine the reliance requirement giving rise to a catch-22 of its own. If a royalty recipient can discover underpayments only by exercising its auditing rights, and if exercising those rights entails not relying on the royalty reports, then false reports can never support a claim of fraud.

Yet another downside concerns the parties' ability to control their exposure to legal liability. When courts permit fraud liability between contracting parties, they often refuse to enforce the parties' attempts to limit their exposure. Thus the Delaware Court of Chancery recently held in ABRY Partners V, L.P. v. F & W Acquisition L.L.C. that "when a seller intentionally misrepresents a fact embodied in a contract-that is, when a seller lies-public policy will not permit a contractual provision to limit the remedy of the buyer to a capped damage claim." (121) Whatever the merits of this rule, familiar worries about plaintiff-friendly juries and insufficient judicial oversight mean that many parties will shy away from liability in fraud as a means of securing no-obstruct duties.

Finally, there may well be good arguments for a default rule of no liability for fraud in the performance. There are many representations about performance for which the parties want only contract remedies, the most obvious example being warranties. Nor is every contractually required representation a representation about first-order performance. Statements of an appraiser about the value of a property, for example, are not about the appraiser's performance, but elements of it. Lastly, there is a legitimate worry that fraud liability for express, contractually required representations will evolve into a more general cause of action for bad faith breach. Thus when the trial court in Alabama v. Exxon upheld the jury's finding of fraud, it pointed not only to explicit falsehoods in the royalty reports, but also to the implicit representation in each royalty payment that the payment complied with the agreement. (122) Such reasoning brings us to the place where semantic and legal analyses diverge. It may well be reasonable to interpret some contract behavior as implicitly representing conformity with the contract. But it does not follow that those implicit representations should support a claim of fraud. Where the parties have not separately agreed to make representations clearly designed to promote enforcement, the arguments for permitting the extracompensatory remedies of fraud lose much of their force. In some cases, it may be that the parties want duties not to obstruct and would prefer the extra protection of fraud liability. But majoritarian and information-forcing considerations, not to mention worries about a slippery slope, might recommend a default of no fraud liability, forcing parties who want that extra protection to say so in their contract.

The analysis of fraud liability has arrived at a curious place. A no fraud liability default, which requires parties who want such liability expressly to agree to it, looks more like a rule of contract than one of tort. Practically speaking, it simply permits the parties to attach a certain extracompensatory remedy (punitive damages) to a certain category of contract terms (duties to represent). But now it is difficult to see why we should not also permit parties to specify the extracompensatory remedies they want (adverse inferences, penalties, damage caps) and to attach them to the terms that will be most effective (no-obstruct duties that may not involve representations). The better solution is the contractual one.

IV. THE UTILITY OF CONTRACTUAL DUTIES NOT TO OBSTRUCT

Parts I and II demonstrated that parties can contract for effective no-obstruct duties only if they can also attach penalties or punitive damages to their breach. Part III showed that duties to represent performance backed by the law of fraud can play the same role and suggested that the existence of such terms is evidence that parties want effective contractual duties not to obstruct. I have not yet discussed the costs of penalties and punitive damages, or weighed them against the potential benefits of effective duties to cooperate. This Part returns to the level of theory and fills in that analysis. I first show that many of the costs commonly associated with penalties and punitive damages--such as preventing efficient breach, creating bad investment incentives, and incentivizing opportunistic litigation- are much less likely to apply when those remedies are attached to no-obstruct terms. Fulfilling a promise from Part I, I then argue that extracompensatory remedies for obstructive breach will often be the preferred alternative to a more familiar solution to underenforcement: first-order damage multipliers.

A. The Costs of Penalties and Punitive Damages

There is broad agreement in the economic literature that, in most contexts, parties prefer compensatory or subcompensatory damage measures for breach of contract, with some version of expectation damages often identified as optimal. (123) The arguments for why this is so have evolved over the years to the point where they are fairly subtle. For present purposes, a generic blueprint will suffice. (124)

Most familiarly, supercompensatory damages (penalties or punitive damages) threaten efficient breach. The bilateral monopoly every contract creates makes renegotiation difficult, meaning that a promisor faced with a penalty or punitive damages is likely to perform in situations where there is more value in breach. (125) Supercompensatory damages also create the wrong incentives earlier in the transaction. The promisor, fearing the large payout should he breach, will invest an inefficient amount to ensure his performance, while the promisee, who now prefers breach to performance, will take too few precautions against nonperformance. In fact, supercompensatory damages create a moral hazard for the promisee, giving her a reason to try to induce and even falsely allege breach. (126) Finally, supercompensatory damages are unlikely to result in risk-shifting benefits. The greater the liability a promisor faces for breach, the more he will charge the promisee up front. But most promisees are either risk-neutral or risk-averse, meaning they will not want to pay more now to gamble on the chance of a big payoff for nonperformance later. (127) Taken together, these bad incentives decrease the aggregate gains of trade from the transaction. That is, they decrease the size of the pie the parties have to divide between them, a result no one prefers. (128) If the above considerations apply equally when supercompensatory damages are attached to no-obstruct terms, then even if the law were to permit parties to choose penalties or punitive damages, they would be unlikely to do so.

But those considerations do not apply equally. Duties to cooperate share four general features that distinguish them from first-order terms. First, compliance is relatively cheap. The affirmative no-obstruct duties I have described entail only minor administrative costs--keeping records, making reports, and the like. And prohibitions on bad acts, like falsifying records or hiding breach, cost less to perform than to breach, since it takes effort to engage in that obstructive behavior. (129) Second, the benefits of cooperative performance (increased chance of recovery) are transaction specific. Combined with the fact that performance is cheap, this makes it unlikely that circumstances will arise in which the resources used to perform a no-obstruct duty will create more value invested elsewhere. That is, the opportunity costs of performing are minimal or nonexistent. Third, we can predict that many promisors going into the transaction will expect to comply with no-obstruct duties. This is not only because of low compliance and opportunity costs, but also because those duties are special examples of the general virtues of honesty and good faith--behaviors many promisors are inclined toward anyway. Finally, the promisee's reliance on no-obstruct duties is of a different sort from her first-order reliance. She relies not on sharing in extralegal gains of trade, but on the vindication of a purely legal and distributive right--her right to be made whole in the case of first-order breach.

These four characteristic features of contract duties to cooperate entail that many inefficiencies commonly associated with penalties and punitive damages do not apply when those remedies are attached to no-obstruct terms. Perhaps the most important difference is that compliance with the no-obstruct duties is generally cheap or free, with respect to both out-of-pocket costs and opportunity costs. As a result, obstructive breach is far less likely to result in a more efficient reallocation of resources. That is, it is unlikely that the promisee will choose to obstruct because the costs of cooperation have gone up or because those resources could generate more value elsewhere. Obstructive breach is rarely efficient.

Willful obstructive breach is more likely opportunistic--an attempt to recapture a piece of the contractual pie that was bargained away, rather than reallocating resources to create more value. (130) Opportunism not only creates no added value, but the threat of opportunistic breach sucks resources from the transaction, as "potential opportunists and victims expend resources perpetrating and protecting against" it. (131) Discouraging opportunistic breaches with penalties or punitive damages is therefore not merely efficiency neutral, but represents a net gain in value. In Richard Posner's oft-quoted words: "If a promisor breaks his promise merely to take advantage of the vulnerability of the promisee ... we might as well throw the book at [him]." (132)

But not all breaches are intentional. Even where compliance is cheap, a promisor facing huge liability for obstructive breach might invest too much to guarantee that no mistakes are made. A penalty or punitive damages clause might cause him to triple-check records, overmonitor employees, or run every certificate of compliance by his lawyer--wasted effort as far as everyone is concerned. (133)

The discussion of fraud liability provides the answer: we can predict that contracts that attach penalties or punitive damages to no-obstruct terms will limit those remedies to intentional or reckless obstructive breaches. That is, parties will contract for an advantage of fraud liability: scienter. Requiring proof of scienter insulates the promisor from liability for mere negligence or reasonable mistake. (134) More to the point, by limiting penalties and punitive damages to intentional and reckless obstructive breaches, the parties can give each other a reason to take only reasonable precautions against obstruction.

Contracting for scienter also addresses two other costs commonly associated with extracompensatory remedies: promisee efforts to prevent, rather than promote, promisor performance and false allegations of breach. One answer to the first worry is that it is difficult for the promisee to interfere with the performance of no-obstruct duties. For instance, neither recordkeeping nor truthful reporting requires promisee cooperation--each is entirely in the power of the promisor. As for the second, the parties can choose to penalize only forms of obstruction where false positives are unlikely, diminishing the danger of false claims for breach. Contracting for scienter adds another layer of protection against both dangers. If it is difficult to induce obstructive breach in general, it is doubly difficult to induce an intentional or reckless obstructive breach. And adding a scienter requirement makes it even less likely that the promisee will win a nonmeritorious suit for obstructive breach.

The remaining two reasons that penalties and punitive damages are usually inefficient can be dealt with in short order. If the promisor does not expect to obstruct recovery and is not worried about false allegations of obstructive breach, the extra liability costs him very little, meaning there are fewer costs to pass along. The result is likely to be positive for a risk-averse promisee. While she might pay slightly more for a penalty or punitive damages clause, the extra assurance of first-order recovery is worth the price. (135) Nor need we worry so much about promisee overreliance on no obstructive breach. While first-order reliance often involves transaction-specific investment in the value-creating activity, when it comes to no-obstruct terms, the promisee is relying on the vindication of her legal right to recover for any first-order breach. It is hard to imagine what would count as overreliance on a legally guaranteed remedy.

There is yet another cost to consider, one unique to no-obstruct duties. Chris Sanchirico has recently identified a commonly overlooked downside of penalizing certain forms of obstruction: the higher the penalty for attempts to avoid detection of a wrong, the greater the wrongdoer's incentive to take additional steps to hide those attempts to avoid detection. (136) There are two negative efficiency implications. First, the resources expended on the added detection avoidance are "[f]rom a societal perspective ... deadweight loss." (137) Second, increasing the penalty for obstructive breach will be ineffective if that penalty does not deter the bad behavior, but instead results in successful efforts to obstruct recovery for the obstructive breach. That is, if Sanchirico's analysis is correct, extracompensatory remedies for obstructive breach have an inefficiency all their own. Penalizing attempts to obstruct recovery might not deter obstruction, but result instead in expenditures to obstruct recovery for the obstruction of recovery, and then attempts to obstruct any recovery for that obstruction, and so on. (138)

Sanchirico's argument, however, does not take full account of the fact that it becomes more and more difficult to avoid detection as one moves up the ladder to higher orders of obstruction. Fabrications become more complex and less plausible; the points of possible detection increase; the wrongdoer has fewer resources and options moving forward. A web of lies is less stable than a single lie, and the more extended and extensive the cover-up, the more costly and precarious it becomes. While in some cases penalties might make it worth investing to hide one's detection avoidance, rarely will it be worth investing significant additional resources to hide those investments. In many circumstances, even the first step will not be worth the effort. There is no a priori reason to expect that the parties will not be able to locate specific duties to cooperate in recovery that do not entail the special costs Sanchirico identifies.

I conclude that attaching penalties and punitive damages does not cost as much when attached to obstructive breach as when applied to first-order breach. When those remedies are necessary to secure effective duties not to obstruct, the added benefit of making the deal happen may be well worth the cost of the extracompensatory remedy.

B. Multipliers

That penalties and punitive damages for obstructive breach can be worth the price does not entail that they are the most cost-effective response to the problem of underenforcement. There is another, more familiar solution: damage multipliers. Multiplying verified first-order losses by the reciprocal of the probability of enforcement ratchets awards up to where the risk-neutral or risk-preferring promisee is fully insured against breach, and the promisor has the right incentives to perform. (139) Even if the benefits of an effective no-obstruct term outweigh the costs of its extracompensatory remedy, a first-order damage multiplier might bring the same or greater benefit at a lower price.

The relative costs of different contract terms are highly fact dependent and vary from transaction to transaction. There might well be situations where a multiplier is the cheaper alternative, and perhaps the law should allow parties to opt for multiplied damages. But multipliers are not free. While a complete analysis of the relative costs and benefits, across different transaction types, of damage multipliers versus extracompensatory remedies for obstructive breach would be interesting, here I will simply note some of the most significant and sometimes neglected costs of multipliers in the contract setting. I hope this will be sufficient to show that there are circumstances in which a duty to cooperate backed by penalties or punitive damages is the preferred alternative.

Multipliers come in two basic varieties. Static multipliers are fixed by law or by contract, while dynamic multipliers are determined after finding a first-order wrong on the basis of an additional finding as to the probability of enforcement. While a static multiplier corrects for the average or expected chance of nonenforcement, a dynamic multiplier corrects for the chance of nonenforcement in this case, taking into account the unique history of the transaction, including the defendant's behavior in it. (140)

Static multipliers are easy to administer, but have obvious drawbacks. Unlike penalties and punitive damages, which are meant to ensure performance, multipliers are supposed to ensure that the promisor fully internalizes the costs of his behavior (harm to the promisee) and therefore has the right incentives to perform or breach. If the multiplier is set by contract, the parties must predict the probability of enforcement in advance. When that prediction turns out to be wrong, the multiplier will result in too few or too many breaches. This is not a problem when it comes to penalties and punitive damages. By erring on the side of a greater sanction, the parties or the court can ensure that bad behavior is deterred for a range of enforcement probabilities. (141)

But the problem with static multipliers is not only epistemic. A static multiplier does not change depending on the parties' behavior. But the multiplier affects that behavior, and that changed behavior can in turn affect the actual probability of enforcement. Craswell describes potential benefits of this interplay. (142) When good behavior leads to a lower probability of enforcement (e.g., in the regulatory context), a static multiplier gives potential defendants more reason to be good: doing so helps avoid even greater liability. But there is also the effect Sanchirico emphasizes: the more costly the remedy, the greater the incentive to engage in bad behavior to avoid it. Where the promisee is particularly worried about obstruction, she will not want a static multiplier, which would give the breaching promisor all the more reason to try to escape liability--all the more reason to obstruct recovery. (143) It is true that the parties could choose an even higher multiplier to account for the new incentives--one that assumes obstruction. But now we are in the realm of the absurd: to secure efficient incentives for first-order breach, the parties contract for inefficient obstruction of first-order recovery. Rather than increasing the incentive to engage in such bad-faith behavior, it is better simply to prohibit it, which means a no-obstruct term backed by an extracompensatory remedy.

Dynamic multipliers do not create new incentives to obstruct. Quite the opposite: a perfectly dynamic multiplier renders obstruction otiose. The more the promisor obstructs recovery, the lower the probability of enforcement, the greater the multiplier, and the greater the potential damage award. Whether the breaching promisor helps or hinders recovery, the anticipated, risk-adjusted damage amount remains the same. A perfectly dynamic multiplier, therefore, both robs obstructive behavior of any benefit and perfectly insures the risk-neutral promisee against obstruction. Dynamic multipliers also share an advantage of generic fee-shifting and prejudgment-interest clauses: if it works, a dynamic multiplier applies to all postbreach obstructive behavior, whether anticipated in the contract or not. (144)

There is, however, that crucial caveat: if it works. Dynamic multipliers face new epistemic hurdles, for they work only if the promisor's obstructive behavior's effect on the probability of enforcement is both verifiable and quantifiable. There are reasons to expect that it will be neither.

First, verifiability. If it is possible that the promisor will be held liable for the underlying breach though his obstructive behavior escapes detection or proof, the dynamic multiplier will not work. The promisor gets the benefits of obstructing recovery without paying all the costs, for in some possible futures, his obstructive behavior will not be factored into the multiplier. The same applies if there is asymmetric information about the effectiveness of an obstructive tactic--if the promisor knows that it is more effective than can be verified in court. (145)

The second difficulty is quantifying the probability of enforcement. Even if the factfinder has perfect access to the totality of the facts surrounding the breach, including the defendant's obstructive behavior, she still must be able to extrapolate from those circumstances the numerical probability that, at the time of nonperformance, the breach would be discovered and verified. There is no algorithm for arriving at that number. And while empirical studies might help in arriving at static multipliers, (146) the facts and circumstances of each case are so different that they provide little assistance in calculating dynamic ones. Quantification requires a judgment call, based on the factfinder's background knowledge and situation sense. As such, it is not only inexact, but prey to familiar cognitive limitations and biases. (147)

It would be wrong to draw too strong a conclusion. The point of a damage multiplier is to give the promisor the right incentives. A multiplier works, therefore, not when it accurately measures the objective probability of enforcement, but when it reflects the promisor's perception of that probability. While cognitive biases might cause factfinders' judgments to err systematically, some of those errors could track promisor judgments. Moreover, while verification difficulties systematically push damages below the ideal, quantification errors might be equally distributed above and below it. The breaching promisor, not knowing where in that range assessed damages will fall, will assume the average or ideal measurement (as he perceives it). (148) Nonetheless, verification and quantification are significant costs of dynamic multipliers. It is simply expensive to judge after a finding of breach what the probability of that finding was at an earlier time- expensive both in terms of the resources required to arrive at a figure and in terms of the risk of error. Again, penalties and punitive damages for obstructive breach can avoid both the verification and the quantification problems by erring on the side of a higher penalty. It is easier to deter than to internalize.

Nor should we forget the other costs of supercompensatory damages for first-order breach identified in the previous section. Multipliers assume risk-neutral or risk-preferring parties, while there are good reasons to think that most contracting parties are risk averse. (149) Where the multiplier is dynamic, the vagaries of verification and quantification mean even more risk. And multiplied damages are likely to be especially salient or vivid. While perfectly rational and risk-neutral parties would not fear multiplied awards, in the real world they might well interact with the biases of risk-averse parties to deter efficient first-order breach. More targeted no-obstruct duties backed by penalties or punitive damages provide an attractive alternative. There may be contexts where first-order multipliers effectively address concerns about underenforcement, but there are almost certainly transactions where the parties would reasonably prefer to contract for effective duties not to obstruct.

CONCLUSION

The discussion in this Article has been fairly abstract, but its conclusions are of more than academic interest. Consider the Department of Defense's two contracts with Halliburton subsidiary KBR to restore Iraqi oil infrastructure, known as "RIO" and "RIO2." (150) A 2006 House Minority report concluded that KBR had significantly underperformed on both contracts. More importantly, it found that KBR "leadership demonstrated minimal cooperative attitude resolving problems," refusing to allow the government access to information that would verify its level of performance. (151) For instance, it appears that early in its attempt to drill a pipeline under the Tigris River, KBR learned that geological conditions would likely prevent its successful performance. KBR, which was charging the government on a cost-plus basis, did not inform the Department of Defense of its probable nonperformance, but continued work and even took steps to hide the difficulties it was encountering. (152)

Whether or not these reports are accurate (they have not been litigated), they illustrate why parties might want to write into their contract duties to cooperate in recovery for first-order breach. The RIO contracts probably include statutory terms requiring KBR to permit audits and other examination of records. (153) But ensuring performance and recovery for breach might well require more: contractual duties to keep accurate records of work done, to cooperate in investigations, and, in the case of the Tigris River project, immediately to report obstacles to performance. (154) This Article has shown that such contractual duties must be backed by extracompensatory remedies if they are to be effective.

The above analysis recommends several possible reforms. The most general and effective would be to create an exception to the rules against penalties and punitive damages allowing parties to attach those remedies to the breach of duties to cooperate in recovery. Less generally effective, but still a step in the right direction, would be to clarify the availability of contractually specified adverse inferences conditioned on obstructive breach. Finally, within the contract-rule framework, courts should at least permit, absent contractual language to the contrary, compensatory damages for the additional litigation costs and prejudgment interest for the delays that obstructive breach causes.

Courts or legislatures could allow parties to adopt some no-obstructive measures even without changing the law of contract by clarifying that the economic loss rule and other obstacles to tort liability in the contract context do not apply to fraudulent obstruction. Minnesota's codification of the economic loss rule specifies that it "shall not be interpreted to bar tort causes of action based upon fraud or fraudulent or intentional misrepresentation or limit remedies for those actions." (155) This is the better rule (though possibly the wrong default), for it enables the parties to craft information-sharing contractual duties to increase the likelihood that, in the case of breach, recovery will be had.

While I believe the analysis above fully supports these conclusions, I do not think it is the last word on duties to cooperate in recovery. At least three issues deserve further research. The first is an empirical hypothesis. This Article has considered only legal remedies for obstructive breach. In many contexts, such behavior will be subject to extralegal sanctions that are just as, or more, important to the parties. Most obviously, in long-term contractual relationships, obstructive breach may cause one side to end the relationship to the other's significant detriment. Alternatively, in some settings obstruction might be subject to reputational and other extralegal sanctions. The central conclusion of this paper--that in many cases the parties want no-obstruct duties, but that to be effective, such duties must be backed by extracompensatory remedies--suggests that we should expect to find such obligations where nonlegal sanctions can step in to fill the gap.

Second, having argued that the law should permit parties to choose extracompensatory remedies for obstructive breach, I have not yet fully analyzed whether or when extracompensatory remedies should be the default. The most generally effective default remedy for obstructive breach would be punitive damages. But one can imagine less radical measures, such as the wider use of adverse inferences. Determining the right default involves a variety of considerations, none of which are specific to the question of duties to compensate, but are worth exploring. (156)

Lastly, my analysis of why the law might want to permit parties to specify penalties or punitive damages for obstructive breach has been largely instrumentalist, focusing on how the parties can structure their contracts to get the best incentives for the least cost. I have not considered how those remedies fare under any of the nonconsequentialist theories of contract law, many of which have their own arguments for or against the traditional prohibitions on extracompensatory contract remedies. (157) And contractual duties not to obstruct raise nonconsequentialist considerations all their own. Such terms not only protect the promisee's entitlement to first-order damages, but also modify the way legal entitlements and sanctions are enforced. Whether to permit the parties to contract for such duties, therefore, depends in part on whether or to what extent private persons should be ceded the authority to determine the enforcement mechanisms that support legally determined transaction structures. (158) The answers to such questions turn on considerations of political authority and legitimacy that are specific to this area of contract law.

(1.) 507 S.E.2d 344, 347 (Va. 1998). In addition to the false certificates of compliance, Richmond claimed constructive fraud based on McDevitt's decision to seal the ends of the tubes with grout, giving the false impression that they had been filled. The Virginia Supreme Court held that these allegations were "nothing more than allegations of negligent performance of contractual duties," and therefore nonactionable as constructive fraud. Id.

(2.) There may be other harms as well, such as making recovery more expensive. I consider the effectiveness of compensatory damages for these harms in infra Subsection I.B.2.

(3.) This is a very thin concept of cooperation. My technical use of the term is obviously different from more demanding or normatively laden concepts of cooperation. See, e.g., Michael E. Bratman, Shared Cooperative Activity, in FACES OF INTENTION: SELECTED ESSAYS ON INTENTION AND AGENCY 93, 103-05 (1999).

(4.) The received wisdom should no longer be that compensatory damage measures are the only efficient remedies. See Richard R.W. Brooks, The Efficient Performance Hypothesis, 116 YALE L.J. 568 (2006); Alan Schwartz, The Case for Specific Performance, 89 YALE L.J. 271 (1979). Special transactional situations where the received wisdom does not apply are collected in Aaron S. Edlin & Alan Schwartz, Optimal Penalties in Contracts, 78 CHI.-KENT L. REV. 33 (2003).

(5.) The most significant recent decision is the Delaware Chancery Court's holding in ABRY Partners V, L.P. v. F&W Acquisition L.L.C. that "when a seller intentionally misrepresents a fact embodied in a contract--that is, when a seller lies--public policy will not permit a contractual provision to limit the remedy of the buyer to a capped damage claim." 891 A.2d 1032, 1036 (Del. Ch. 2006); see also id. at 1055-62 (discussing other holdings and authorities). For recent scholarly discussions, see Kevin Davis, Licensing Lies: Merger Clauses, the Parol Evidence Rule and Pre-Contractual Misrepresentations, 33 VAL. U. L. REV. 485 (1999); and Jeffrey M. Lipshaw, Of Fine Lines, Blunt Instruments, and Half-Truths: Business Acquisition Agreements and the Right To Lie, 32 DEL. J. CORP. L. (forthcoming 2007).

(6.) IAN AYRES & GREGORY KLASS, INSINCERE PROMISES : THE LAW OF MISREPRESENTED INTENT 8 (2005).

(7.) Robert E. Scott & George G. Triantis, Anticipating Litigation in Contract Design, 115 YALE L.J. 814, 857 (2006). For an example of the neglect in a famous source, see L.L. Fuller & William R. Perdue, Jr., The Reliance Interest in Contract Damages (pt. 1), 46 YALE L.J. 52, 58 (1936) ("If a contract represents a kind of private law, it is a law which usually says nothing at all about what shall be done when it is violated. A contract is in this respect like an imperfect statute which provides no penalties, and which leaves it to the courts to find a way to effectuate its purposes.").

(8.) The conclusions of this Article are consistent with results in other areas of the law. See, e.g., Alex Raskolnikov, Crime and Punishment in Taxation: Deceit, Deterrence, and the Self-Adjusting Penalty, 106 COLUM. L. REV. 569 (2006) (analyzing sanctions for tax evasion and avoidance) ; Ian C. Wiener, Note, Running Rampant: The Imposition of Sanctions and the Use of Force Against Fleeing Criminal Suspects, 80 GEO. L.J. 2175 (1992) (discussing optimal penalties for criminal flight).

(9.) Oliver Wendell Holmes, The Path of the Law, 10 HARV. L. REV. 457, 462 (1897), reprinted in 110 HARV. L. REV. 991, 995 (1997).

(10.) See, e.g., RICHARD POSNER, LAW, PRAGMATISM, AND DEMOCRACY 58 (2003) ("Holmes pointed out that in a regime in which the sanction for breach of contract is merely an award of compensatory damages to the victim, the entire practical effect of signing a contract is that by doing so one obtains an option to break it."); Avery Wiener Katz, The Option Element in Contracting, 90 VA. L. REV. 2187, 2202 (2004) ("In option terminology, we can restate Holmes's point by saying that the promisor holds a call option to buy her way out of the contract by paying a strike price equal to the value of court-awarded damages.") ; Robert E. Scott & George G. Triantis, Embedded Options and the Case Against Compensation in Contract Law, 104 COLUM. L. REV. 1428, 1429 n.1 (2004) ("It is well known that contract damages effectively give the promisor an option between performing the promise or breaching and paying damages. The classic statement is by Justice Holmes...."); Seanna Valentine Shiffrin, The Divergence of Contract and Promise, 120 HARV. L. REV. 708, 727 (2007) ("Justice Holmes famously declared that a contract to perform should be understood not as a promise to perform full stop, but as a promise either to perform or to pay damages.").

(11.) OLIVER W. HOLMES, THE COMMON LAW 302 (Boston, Little, Brown & Co. 1944) (1881); see also Joseph M. Perillo, Misreading Oliver Wendell Holmes on Efficient Breach and Tortious Interference, 68 FOKDHAM L. REV. 1085, 1085-93 (2000).

(12.) 686 P.2d 1158, 1166 (Cal. 1984), overruled by Freeman & Mills, Inc. v. Belcher Oil Co., 900 P.2d 669 (Cal. 1995).

(13.) See generally DORI KIMEL, FROM PROMISE TO CONTRACT: TOWARDS A LIBERAL THEORY OF CONTRACT (2003); Jules L. Coleman, Douglas D. Heckathorn & Steven M. Maser, A Bargaining Theory Approach to Default Provisions and Disclosure Rules in Contract Law, 12 HARV. J.L. & PUB. POL'Y 639 (1989). For a description of the way that trust can play a role in assuring cooperation in recovery, see ANNETTE C. BAIER, Trust and Its Vulnerabilities, in MORAL PREJUDICES: ESSAYS ON ETHICS 130, 138-39 (1994).

(14.) For a helpful overview of such mechanisms, albeit with an emphasis more on securing ex ante reliability rather than ex post compensation, see Ronald J. Mann, Verification Institutions in Financial Transactions, 87 GEO. L.J. 2225 (1999).

(15.) For the parties' ability to modify discovery procedures, see FED. R. CIV. P. 29. For the possible use of alternative burdens of proof, see Scott & Triantis, supra note 7, at 856-78.

(16.) AYRES & KLASS, supra note 6, at 165-68. This statement must be qualified by the already-noted reluctance of some courts to permit actions for fraud in the contractual setting. See infra Section III.B.

(17.) FED. R. CIV. P. 11(b)(1); see, e.g., CAL. CIV. PROC. CODE [section] 128.7 (West 2006); ILL. SUP. CT. R. 137; N.J. CT. R. 1:4-8 (West 2007) ; TEX. R. CIV. PROC. ANN. 13 (Vernon 2003). But see Chris William Sanchirico, Detection Avoidance, 81 N.Y.U. L. REV. 1331, 1384-86 (2006) (observing the limited scope of criminal and procedural sanctions for detection avoidance).

(18.) A rule can be said to be "quasi-mandatory" when the parties have a limited ability to modify it, though not to contract out of it entirely. Thus Rule 29 of the Federal Rules of Civil Procedure and many state analogs allow the parties to stipulate to discovery procedures, though such stipulations are subject to court review to ensure no undue frustration of the administration of justice. See, e.g., Franklin v. White, 493 N.E.2d 161 (Ind. 1986); Garden State Plaza Corp. v. S.S. Kresge Co., 189 A.2d 448 (N.J. Super. Ct. App. Div. 1963).

(19.) There is an analog to promisor duties to cooperate in recovery: promisee duties not to falsely allege breach. Many of the arguments below would apply equally to such duties, though I do not consider them as such.

(20.) Examples of noninformational duties to cooperate include requirements that the promisor keep funds available to pay any first-order damage award, that the promisor remain within the reach of a jurisdiction, or that the promisor voluntarily pay damages for clear breach. Many of my conclusions apply to these duties as well, though the details of the argument and the remedies required are different. For example, as Richard Craswell pointed out to me, a contractual liquidity requirement might be most effective if its breach triggered accelerated first-order remedies, such as the right to demand adequate assurance.

(21.) I follow here the convention of using "observability" to refer to the parties' ability to detect some fact, and "verifiability" to refer to the ease with which that fact can be demonstrated in court.

(22.) See Exxon Corp. v. Dep't of Conservation & Natural Res., 859 So. 2d 1096 (Ala. 2002); Post-Judgment Order, Alabama v. Exxon Corp., No. 99-2368 (Ala. Cir. Ct. Mar. 29, 2004).

(23.) 859 So. 2d at 1100; Post-Judgment Order, supra note 22, at 9.

(24.) Post-Judgment Order, supra note 22, at 21, 52-54.

(25.) Id. at 19-22.

(26.) Id. at 21.

(27.) Id. at 21-22.

(28.) Id. at 20.

(29.) Id. at 21 (citations omitted). A similar description of the same dynamic can be found in the legislative history of a California bill that would have penalized recording companies for underpaying and misreporting royalties. The bill's sponsor, Senator Murray, explained: "[T]his bill addresses the core issue that has allowed record companies to under-report royalty earnings without any penalty.... Under the current structure, there is no disincentive or penalty if record companies do not properly account royalties to artists; therefore, bad behavior is rewarded." Cal. Assem. Comm. on Arts, Entertainment, Sports, Tourism, and Internet Media, Bill Analysis, S. 1034, 2002-2003 Leg., Reg. Sess., at 1-2 (Cal. 2003) (as amended May 12, 2003) (statement of Sen. Murray), available at http://info.sen.ca.gov/pub/03-04/bill/sen/sb_1001-1050 /sb_1034_cfa_20030630_102556_asm_comm.html.

(30.) Post-judgment Order, supra note 22, at 32; see also Reis v. Peabody Coal Co., 997 S.W.2d 49, 68, 72 (Mo. Ct. App. 1999) (rejecting the fraud defendant's argument that the plaintiff incurred no actual damages because plaintiff had already recovered for breach of the royalty contract in arbitration).

(31.) Post-Judgment Order, supra note 22, at 19-37, 62.

(32.) Exxon Corp. v. Dep't of Conservation & Natural Res., No. 1031167 (Ala. filed Apr. 30, 2004).

(33.) RESTATEMENT (SECOND) OF CONTRACTS [section] 350 cmt. b (1981).

(34.) Discouraging frivolous lawsuits is but one of the claimed advantages of fee shifting. See, e.g., W. Kent Davis, The International View of Attorney Fees in Civil Suits: Why Is the United States the "Odd Man Out" in How It Pays Its Lawyers?, 16 ARIZ. J. INT'L & COMP. L. 361, 399-430 (1999) (arguing that fee shifting can reduce the number of lawsuits and ensure access to litigation); Amy Farmer & Paul Pecorino, A Reputation for Being a Nuisance: Frivolous Lawsuits and Fee Shifting in a Repeated Play Game, 18 INT'L REV. L. & ECON. 147 (1998) (finding that fee shifting can reduce frivolous litigation costs, where lawyer reputation is relevant); James W. Hughes & Edward A. Snyder, Litigation and Settlement Under the English and American Rules: Theory and Evidence, 38 J.L. & ECON. 225 (1995) (finding that efficient settlement, costs, and successful plaintiff suits are more common under the English rule); Walter Olson & David Bernstein, Loser-Pays: Where Next?, 55 MD. L. REV. 1161, 1161-63 (1996) (arguing that fee shifting can diminish speculative lawsuits, can decrease litigation costs, and is more fair); John F. Vargo, The American Rule on Attorney Fee Allocation: The Injured Person's Access to Justice, 42 AM. U. L. REV. 1567, 1590-93 (1993) (describing common arguments against the American rule).

(35.) The plausibility of recovery of attorney fees for obstructive breach is demonstrated by some contemporary reactions to California's experiment with punitive damages for obstruction of recovery in Seaman's Direct Buying Service, Inc. v. Standard Oil Co., 686 P.2d 1158 (Cal. 1984). See C. Delos Putz, Jr., & Nona Klippen, Commercial Bad Faith: Attorney Fees--Not Tom Liability--Is the Remedy for "Stonewalling," 21 U.S.F. L. REV. 419, 499 (1987); Dana Rae Landsdorf, Note, California's Detortification of Contract Law: Is the Seaman's Tort Dead?, 26 LOY. L.A. L. REV. 213, 238 (1992) ("Contracting parties who deny--in bad faith--the existence of a contract should be liable for all damages proximately caused and resulting from such conduct. For example, the breaching party should not only be liable for compensatory damages but also for the injured party's attorney's fees because the non-breaching party does not expect to incur attorney's fees from the transaction.").

(36.) It might be thought that no court would award litigation costs without proof of first-order breach, the idea being that such recovery would give promisees a perverse incentive to sue. That thought overlooks the fact that compensatory litigation-cost damages for breach of an anti-obstruction clause should be limited to costs attributable to the defendant's obstructive behavior. Litigation-cost recovery for breach of the recordkeeping clause would not give Developer a free suit on the first-order breach. It would, however, support her expenditure of extra resources to determine whether there was a first-order breach-expenses that would be unnecessary but for Contractor's failure to keep the required records. Developer should be permitted to recover these costs whether or not she sues for the underlying breach, and if she does sue on the underlying breach, whether or not she prevails in that suit.

(37.) See supra text accompanying note 29.

(38.) RESTATEMENT (SECOND) OF CONTRACTS [section] 356(1) (2006).

(39.) $1,000 = (.1)($10,000).

(40.) $5,000 = (.5)($10,000).

(41.) RESTATEMENT (SECOND) OF CONTRACTS [section] 356 cmt. b (2006); see E. ALLAN FARNSWORTH, CONTRACTS [section] 12.18, at 814-25 (4th ed. 2004).

(42.) See RESTATEMENT (SECOND) OF CONTRACTS [section] 356 illus. 4 (2006); see also id. [section] 356(1) ("Damages for breach by either party may be liquidated in the agreement but only at an amount that is reasonable in the light of the anticipated or actual loss caused by the breach and the difficulties of proof of loss." (emphasis added)).

(43.) 24 WILLISTON ON CONTRACTS [section] 65:33 (4th ed. 2006).

(44.) Specific performance also qualifies as extracompensatory under my definition, but timing issues prevent it from being effective against most sorts of obstruction. An injunction ordering subsequent performance cannot undo the damage caused by destroyed records, hidden or unreported first-order nonperformance, delay tactics, and most other forms of obstruction.

One exception is contractual duties to permit audits. Thus 10 U.S.C. [section] 2313 (2000), which gives the Department of Defense the right to audit records of certain contractors, also gives it the power, enforceable by the district court, to subpoena those records. This is equivalent to a mandatory audit term supported by specific performance. California has recently legislated mandatory auditing rights in recording contracts, though without specifying the remedy for their breach. CAL. CIV. CODE [section] 2501 (West Supp. 2007). At some point the California courts will have to decide what the appropriate remedy is. My analysis suggests that it should be specific performance, an adverse inference, or some other extracompensatory measure.

Yet other remedial options are available where the promisor is a corporate entity and liability for first-order and obstructive breaches (or rewards for cooperation) can be assigned separately to the corporation and its agents. See Jennifer Arlen & Reinier Kraakman, Controlling Corporate Misconduct: An Analysis of Corporate Liability Regimes, 72 N.Y.U.L. REV. 687 (2997). Such a solution is comparable to putting enforcement in the hands of a third party. See supra text accompanying note 14.

(45.) I will follow Farnsworth and use "terminate" (where one could also say "rescind" or "cancel") to refer to the nonbreaching party's right to end the contract. See FARNSWORTH, supra note 41, [section] 8.15 n.2.

(46.) According to California Senator Murray, a former entertainment lawyer,
   even though the obligation to accurately account for royalty
   earnings is a material part of the contract, most if not all
   recording artist contracts provide that a breach of the obligation
   to account for or pay royalties is not a material breach, leaving
   the artist with no real recourse except to settle the claim on
   perhaps a percentage of what is owed or else conduct protracted and
   expensive litigation at the possible expense of the artist's
   career.


Cal. Assem. Comm. on Arts, Entertainment, Sports, Tourism, and Internet Media, Bill Analysis, S. 1034, 2002-2003 Leg., Reg. Sess., at 1-2 (Cal. 2003) (third reading) (paraphrasing Sen. Murray), available at http://info.sen.ca.gov/pub/03-04/bill/sen/sb_1001-1050 /sb_1034_cfa_20030514_100846_sen_floor.html.

(47.) For the power of labels to dictate terms, and how this plays out in terms of artists' ability to get accurate accountings or otherwise monitor label performance, see Wendy V. Bartholomew, Fiduciary Duty: Can It Help Calm the Fears of Underpaid Artists?, 6 VAND. J. ENT. L. & PRAC. 246 (2004); and Corrina Cree Clover, Note, Accounting Accountability: Should Record Labels Have a Fiduciary Duty To Report Accurate Royalties to Recording Artists?, 23 LOY. L.A. ENT. L. REV. 395 (2003).

(48.) FARNSWORTH, supra note 41, [section] 8.16.

(49.) Richard Craswell, Property Rules and Liability Rules in Unconscionability and Related Doctrines, 60 U. CHI. L. REV. 1, 7 (1993).

(50.) Id. at 7-8 (emphasis omitted).

(51.) For a description of how courts and prosecutors use adverse inferences to discourage obstruction in the criminal context, see Sanchirico, supra note 17, at 1378-82.

(52.) Scott & Triantis, supra note 7, at 857-58 (footnote omitted). See generally id. at 856-78 (describing decisions).

(53.) Thus Scott and Triantis identify only three ways by which the parties might
   clarify, reverse, or fine-tune the default allocation in their
   contract.... The first approach is by direct allocation of burden;
   the second is by predesignating whom the plaintiff will be in the
   event of a dispute; and the third is by framing the substantive
   provisions governing, for example, the right to assign or terminate
   a contract.


Id. at 866. I am suggesting a fourth: rather than direct allocation of the burden, conditional allocation. Scott and Triantis do describe an instance in which the court determined that a prima facie showing of bad faith shifted the burden of proof--though not at the behest of the contract. Id. at 876-77 (discussing Int'l Harvester Co. v. Calvin, 353 So. 2d 144 (Fla. Dist. Ct. App. 1977)).

(54.) See generally 29 AM. JUR. 2D Evidence [section] 244 (1944 & Supp. 2006). The evidentiary rule requires proof of scienter: "a conscious awareness of the existence of the dispute and that the act done will destroy evidence or access to evidence." Id. This is interesting in light of my argument in Sections III.B and IV.A that scienter requirements add value to no-obstruct duties.

(55.) See, e.g., Keith N. Hylton, Punitive Damages and the Economic Theory of Penalties, 87 GEO. L.J. 421, 425-39 (1998).

(56.) ($10,000)(.5) = $5,000.

(57.) ($5,000)/(.2) = $25,000.

(58.) In many jurisdictions, punitive damages may only be awarded where there is proof of actual loss. See 25 C.J.S. Damages [section] 197 (2002) ("As a general rule, in order to recover exemplary or punitive damages, actual damages must be shown, or there must be a basis for the recovery of compensatory damages." (footnote omitted)); 1 JOHN J. KIRCHER & CHRISTINE M. WISEMAN, PUNITIVE DAMAGES: LAW AND PRACTICE [section] 5:21 (2d ed. Supp. 2003) ("Abundant authority exists to support the proposition that a finding must be entered entitling the plaintiff to actual damages before that plaintiff will be allowed to recover punitive damages."). This rule is open to criticism. See RESTATEMENT (SECOND) OF TORTS [section] 908 cmt. c (1979) ("Although ... the extent of the harm may be considered in determining their amount, it is not essential to the recovery of punitive damages that the plaintiff should have suffered any harm, either pecuniary or physical."); Jane Mallor & Barry Roberts, Punitive Damages: Toward a Principled Approach, 31 HASTINGS L.J. 639, 666-67 (1980) (arguing against the rule that punitive damages bear a reasonable relationship to actual harm). In jurisdictions that follow the first rule, the secondary harms of obstructive breach might satisfy the actual harm requirement. And the punitive damages in cases where there is proof of harm can be further increased to take account of the ex ante possibility that such proof was not going to be available.

(59.) RESTATEMENT (SECOND) OF CONTRACTS [section] 355 (1981). For developments in the law of punitive damages for breach of contract, see William S. Dodge, The Case for Punitive Damages in Contracts, 48 DUKE L.J. 629, 636-51 (1999).

(60.) RESTATEMENT (SECOND) OF CONTRACTS [section] 356(1) (1981). For descriptions of practical complexities and differences between jurisdictions, see Kenneth W. Clarkson, Roger LeRoy Miller & Timothy J. Muris, Liquidated Damages v. Penalties: Sense or Nonsense?, 1978 WIS. L. REV. 351, 352-57; Ian R. Macneil, Power of Contract and Agreed Remedies, 47 CORNELL L.Q. 495, 501-13 (1962); and Eric L. Talley, Note, Contract Renegotiation, Mechanism Design, and the Liquidated Damages Rule, 46 STAN. L. REV. 1195, 1200-05 (1994).

(61.) Not all states follow these rules. South Carolina, for example, permits punitive damages for breach of contract accompanied by "any act characterized by dishonesty in fact, unfair dealing, or the unlawful appropriation of another's property by design," a rule that might well cover many forms of obstructive behavior. Perry v. Green, 437 S.E.2d 150, 152 (S.C. Ct. App. 1993); see also Dodge, supra note 59, at 649-50 (describing similar rules in Idaho, Mississippi, and New Mexico). It is unclear whether the courts in these states treat such punitive damages as mandatory remedies or as defaults the parties could opt out of, much less whether courts would permit the parties to specify the bad acts that should trigger them.

(62.) See infra text accompanying notes 123-128.

(63.) The argument from freedom of contract is suggested by Daniel Friedmann in The Efficient Breach Fallacy, 18 J. LEGAL STUD. 1, 23 (1989). Samuel Rea describes the difficulty the choice metric poses for economic explanations of the rule against penalties:
   The perplexing aspect of the penalties-liquidated damages
   distinction is that there is a conflict between the apparent, but
   express, intention of the parties and the subsequent action of the
   courts.

   Economists are inclined to believe that courts are contributing
   to inefficiency when they upset the agreed upon terms of a contract.


Samuel A. Rea, Jr., Efficiency Implications of Penalties and Liquidated Damages, 13 J. LEGAL STUD. 147, 148 (1984); see also Lake River Corp. v. Carborundum Co., 769 F.2d 1284, 1289 (7th Cir. 1985) (Posner, J.) ("[T]he parties (always assuming they are fully competent) will, in deciding whether to include a penalty clause in their contract, weigh the gains against the costs ... and will include the clause only if the benefits exceed those costs .... "); Charles J. Goetz & Robert E. Scott, Liquidated Damages, Penalties and the Just Compensation Principle: Some Notes on an Enforcement Model and a Theory of Efficient Breach, 77 COLUM. L. REV. 554, 555 n.12 (1977) (describing how the penalty doctrine is "anomalous in terms of the theoretical underpinnings of modern contract law").

(64.) Thus Alan Schwartz argues:
   Courts do not have to prevent promisees from obtaining penalty
   clauses if promisees do not want penalty clauses. The ex ante rule
   is not merely unnecessary: judicial review produces mischief.
   Courts sometimes mistake compensatory damage measures for
   penalties, and so have found that particular liquidated damage
   clauses would inevitably overcompensate promisees when those
   clauses only protected the expectation. Thus, the ex ante branch of
   the liquidated damage rule should be abandoned.


Alan Schwartz, The Myth That Promisees Prefer Supercompensatory Remedies: An Analysis of Contracting for Damage Measures, 100 YALE L.J. 369, 370 (1990) ; see also Goetz & Scott, supra note 63, at 578-93 (arguing that many penalty clauses are designed to compensate for damages or create efficient incentives that remain inscrutable to courts); Schwartz, supra, at 383-87.

(65.) These developments are described in Edlin & Schwartz, supra note 4, at 43-52.

(66.) See Goetz & Scott, supra note 63, at 554.

(67.) See, e.g., Klinger v. Adams County Sch. Dist. No. 50, 130 P.3d 1027, 1034 (Colo. 2006) ("[A] contract provision for liquidated damages is invalid as a penalty if it is unreasonably large for the expected loss from a breach of contract."); Bragdon v. Twenty-Five Twelve Assocs. P'ship, 856 A.2d 1165, 1173 (D.C. 2004) ("Punitive damages will not lie for breach of contract, even if it is proven that the breach was willful, wanton, or malicious."); Dist. Cablevision P'ship v. Bassin, 828 A.2d 714, 724 (D.C. 2003) ("Agreements to pay fixed sums plainly without reasonable relation to any probable damage which may follow a breach will not be enforced."); TAL Fin. Corp. v. CSC Consulting, Inc., 844 N.E.2d 1085, 1093 (Mass. 2006) ("[L]iquidated damages will not be enforced if the sum is 'grossly disproportionate to a reasonable estimate of actual damages' made at the time of contract formation." (quoting Kelley v. Marx, 705 N.E. 2d 1114, 1116 (Mass. 1999)) ; Ins. Co. of the W. v. Gibson Tile Co., 234 P.3d 698, 703 (Nev. 2006) ("[T]he award of punitive damages cannot be based upon a cause of action sounding solely in contract.") ; Smith v. Grand Canyon Expeditions Co., 84 P.3d 1154, 1161 (Utah 2003) ("[P]unitive damages are recoverable only for torts, not for breach of contract.").

(68.) AYRES & KLASS, supra note 6.

(69.) See supra note 5.

(70.) The three techniques I discuss are not the only ones available. Recording artists have tried and generally failed to impose on recording companies a fiduciary duty to account for royalty payments. See S. 1034, 2002-2003 Leg., Reg. Sess. (Cal. 2003). Like fraud, breach of fiduciary duty can permit the recovery of punitive damages, effectively deterring obstructive behavior. A downside of fiduciary duties as compared to fraud liability is their lack of specificity. Rather than the relatively clear duty to be honest, there is a relatively amorphous duty of loyalty. See Deborah A. DeMott, Breach of Fiduciary Duty: On Justifiable Expectations of Loyalty and Their Consequences, 48 ARIZ. L. REV. 925, 934-40 (2006); D. Gordon Smith, The Critical Resource Theory of Fiduciary Duty, 55 VAND. L. REV. 1399, 1406-11 (2002).

(71.) 686 P.2d 1158, 1167 (Cal. 1984).

(72.) Id.

(73.) For surveys of the variety of lower-court interpretations of Seaman's, see, for example, Freeman & Mills, Inc. v. Belcher Oil Co., 900 P.2d 669 (Cal. 1995); Landsdorf, supra note 35, at 222-35. The furthest extension of Seaman's by the lower courts appears to have been in Koehrer v. Superior Court, 226 Cal. Rptr. 820 (Ct. App. 1986), which "extend[ed] Seaman's to bad faith attempt to deprive employee of contractual benefits." Freeman & Mills, 900 P.2d at 675.

(74.) Freeman & Mills, 900 P.2d at 676-77. Dodge argues that the real turning point was as early as 1988, when the California Supreme Court declined to extend Seaman's to employment contracts. Dodge, supra note 59, at 642 (discussing Foley v. Interactive Data Corp., 765 P.2d 373 (Cal. z988)).

(75.) 686 P.2d at 1167; id. at 1174 (Bird, C.J., concurring in part and dissenting in part).

(76.) See, e.g., Dodge, supra note 59, at 638 n.33; E. Allan Farnsworth, Developments in Contract Law During the 1980's: The Top Ten, 41 CASE W. RES. L. REV. 203, 205 (1990); Friedmann, supra note 63, at 19 n.62; Curtis Bridgeman, Note, Corrective Justice in Contract Law: Is There a Case for Punitive Damages?, 56 VAND. L. REV. 237, 270-71 (2003). But see Shiffrin, supra note 10, at 723 n.27 (recognizing the limited scope of Seaman's).

(77.) See, e.g., John A. Sebert, Jr., Punitive and Nonpecuniary Damages in Actions Based upon Contract: Toward Achieving the Objective of Full Compensation, 33 UCLA L. REV. 1565, 2600-54 (1986).

(78.) Old Am., Inc. v. Microtech Int'l, Inc., 872 F.2d 312, 324-15 (9th Cir. 2989) (Kozinski, J., concurring).

(79.) 686 P.2d at 1167.

(80.) Michael Dorff is therefore wrong when, criticizing the application of punitive damages in Seaman's, he states that "the consequences of denying the contract's existence would be virtually identical to the consequences of any other complete contract breach, and the parties would be expected to negotiate about the risks of a complete breach." Michael Dorff, Attaching Tort Claims to Contract Actions: An Economic Analysis of Contort, 28 SETON HALL L. REV. 390, 423 (1997).

(81.) Oki Am., 872 F.2d at 315 (Kozinski, J., concurring).

(82.) See supra note 73. Kozinski correctly diagnosed this problem:
   [Seaman's] created a cause of action so nebulous in outline and so
   unpredictable in application that it more resembles a brick thrown
   from a third story window than a rule of law. Seaman's gives nary a
   hint as to how to distinguish a bad faith denial that a contract
   exists, from a dispute over contract terms, from a permissible
   attempt to rescind a contract, or from a loosely worded disclaimer
   of continued contractual responsibility.


Oki Am., 872 F.2d at 315 (Kozinski, J., concurring) (internal quotation marks omitted).

(83.) See Oki Am., 872 F.2d at 315 (Kozinski, J., concurring) ("Seaman's throws kerosene on the litigation bonfire by holding out the allure of punitive damages, a golden carrot that entices into court parties who might otherwise be inclined to resolve their differences.") ; Lynch & Freytag v. Cooper, 267 Cal. Rptr. 189, 195-96 (Ct. App. 1990) (suggesting that minor contract action was litigated overzealously due to the "allure of punitive damages").

(84.) Cf. Samuel W. Buell, Novel Criminal Fraud, 81 N.Y.U.L. REV. 1971, 1988-92 (2006) (collecting examples of similar observations about the law of fraud).

(85.) 686 P.2d at 1167.

(86.) In fact, Seaman's recognized the parties' ability "to shape the contours of their agreement and to include provisions for attorney fees and liquidated damages in the event of breach.... [T]hey are free, within reasonable limits at least, to agree upon the standards by which application of the covenant [of good faith] is to be measured." 686 P.2d at 1167. While allowing that the availability of party choice counseled caution in applying tort remedies in the contract setting, the court did not discuss whether parties might opt-out of or contractually modify the Seaman's tort.

(87.) W. PAGE KEETON ET AL., PROSSER AND KEETON ON TORTS [section] 105, at 728 (5th ed. 1984).

(88.) See, e.g., Morrill v. Becton, Dickinson & Co., 747 F.2d 1217, 1222 (8th Cir. 1984) (holding that evidence was insufficient to establish requisite scienter for some claimed misrepresentations as to sales figures connected to royalty payments).

(89.) Post-Judgment Order, supra note 22, at 1, 20-25, 62.

(90.) See, e.g., Morrill, 747 F.2d at 1220 (false royalty statements to inventor); Gregory v. Chem. Waste Mgmt., Inc., 38 F. Supp. 2d 598 (W.D. Tenn. 1996) (false waste-disposal facility royalty statements); Reis v. Peabody Coal Co., 997 S.W.2d 49, 68, 72 (Mo. Ct. App. 1999) (false royalty statements and misrepresentations as to compliance with most-favored-nation clause); Apple Records, Inc. v. Capitol Records, Inc., 137 A.D.2d 50, 56 (N.Y. App. Div. 1988) (false royalty statements in recording contract); Oestreicher v. Am. Nat'l Stores, Inc., 225 S.E.2d 797 (N.C. 1976) (false royalty sales statements); Grynberg v. Citation Oil & Gas Corp., 573 N.W.2d 493 (S.D. 1997) (false oil royalty statements). For another nonroyalty case, see Krantz v. Chateau Stores of Canada Ltd., 683 N.Y.S.2d 24 (App. Div. 1998) (false statement of annual profits used as basis for bonus).

(91.) Richmond Metro. Auth. v. McDevitt St. Bovis, Inc., 507 S.E.2d 344 (Va. 1998).

(92.) 646 F.2d 224, 225-27 (5th Cir. Unit A May 1981).

(93.) 102 P.3d 268 (Cal. 2004).

(94.) Id. at 271.

(95.) See Robinson Helicopter Co. v. Dana Corp., 129 Cal. Rptr. 2d 682, 686 n.6 (Ct. App. 2003).

(96.) Robinson Helicopter, 102 P.3d at 274-75.

(97.) For an even better example, see the discussion of the RIO and RIO2 contracts in the Conclusion.

(98.) See RESTATEMENT (SECOND) OF TORTS [section] 550 (1997).

(99.) I have been able to locate only one case applying fraudulent suppression based on a contractual duty to disclose. Gregory v. Chemical Waste Management, Inc., 38 F. Supp. 2d 598 (W.D. Tenn. 1996), held that there was a duty to disclose where the contract warranted that the defendant's representations "do not and will not include any untrue statement of material fact or fail to include any material fact, all to the end that such statements are not misleading." Id. at 612 (emphasis omitted). The court concluded that the defendant breached the resultant duty to disclose when it did not reveal a change in its manner of calculating royalty payments. Id. The fraud claim in that case, however, was also supported by affirmative misrepresentations in the royalty statements. Id. at 611-12. The reason there are not more examples of fraudulent concealment based on contractual duties to disclose is not difficult to discern: the vast majority of cases and commentary on fraud by silence concern fraud in the inducement, before a contract exists. See, e.g., 37 AM. JUR. 2D Fraud and Deceit [section] 204 (2006); Christopher T. Wonnell, The Structure of a General Theory of Nondisclosure, 41 CASE W. RES. L. REV. 329 (1991).

(100.) See RESTATEMENT (THIRD) OF ECONOMIC TORTS AND RELATED WRONGS [section] 8 (Preliminary Draft No. 2, 2006).

(101.) Robinson Helicopter, 129 Cal. Rptr. 2d at 684.

(102.) As Justice Traynor explained in the seminal case, "[t]he history of the doctrine of strict liability in tort indicates that it was designed, not to undermine the warranty provisions of the sales act or of the

Uniform Commercial Code but, rather, to govern the distinct problem of physical injuries." Seely v. White Motor Co., 403 P.2d 145, 149 (Cal. 1965) (Traynor, C.J.).

(103.) HTP, Ltd. v. Lineas Aereas Costarricenses, S.A., 685 So. 2d 1238, 1239 (Fla. 1996). The leading case adopting this approach is Huron Tool & Engineering Co. v. Precision Consulting Services, Inc., 532 N.W.2d 541, 544-45 (Mich. Ct. App. 1995). See also Dinsmore Instrument Co. v. Bombardier, Inc., 999 F. Supp. 968, 971-72 (E.D. Mich. 1998); Valleyside Dairy Farms, Inc. v. A.O. Smith Corp., 944 F. Supp. 612, 616-17 (W.D. Mich. 1995); Kaloti Enters., Inc. v. Kellogg Sales Co., 699 N.W.2d 205, 219-21 (Wis. 2005). For academic discussions, see Jean Braucher, Deception, Economic Loss and Mass-Market Consumers: Consumer Protection Statutes as Persuasive Authority in the Common Law of Fraud, 48 ARIZ. L. REV. 829 (2006); Steven C. Tourek, Thomas H. Byrd & Charles J. Schoenwetter, Bucking the "Trend": The Uniform Commercial Code, the Economic Loss Doctrine, and Common Law Causes of Action for Fraud and Misrepresentation, 84 IOWA L. REV. 875 (1999); R. Joseph Barton, Note, Drowning in a Sea of Contract: Application of the Economic Loss Rule to Fraud and Negligent Misrepresentation Claims, 41 WM. & MARY L. REV. 1789 (2000).

(104.) 129 Cal. Rptr. 2d at 695; see Tourek et al., supra note 103, at 895-912; Barton, supra note 103, at 1802-12. Some courts have taken the further step of applying the economic loss rule to "foreclose tort claims against the defendant who fraudulently induces the contract by representations about the character and quality of the goods or services sold." Dan B. Dobbs, An Introduction to Non-Statutory Economic Loss Claims, 48 ARIZ. L. REV. 713, 729 (2006).

(105.) See, e.g., Robinson Helicopter, 129 Cal. Rptr. 2d at 697.

(106.) Huron Tool, 532 N.W.2d at 545.

(107.) 129 Cal. Rptr. 2d at 697; see also Grynberg v. Citation Oil & Gas Corp., 573 N.W.2d 493, 510 (S.D. 1997) (Amundson, J., dissenting in part) (arguing that there is no fraud liability where the contract "impliedly create[d] the obligation to make true factual representations"). But see Grynberg, 573 N.W.2d at 501 (majority opinion) (holding that false royalty statements, though required by contract, violated a separate obligation to refrain from invading the property of others by fraud).

(108.) Robinson Helicopter, 102 P.3d at 274-75.

(109.) Id. at 276 (emphasis added). Other factors emphasized by the California Supreme Court likely to distinguish Robinson Helicopter from future cases include the impact of Dana's breach on public safety and the fact that the improper clutches put Robinson out of compliance with FAA regulations. See id. at 274 n.7.

(110.) Saratoga Fishing Co. v. J.M. Martinac & Co., 520 U.S. 875, 880 (1997) ; see also, e.g., Clark v. Rowe, 701 N.E.2d 624, 626 (Mass. 1998) ("When the economic loss rule has been applied, the parties usually were in a position to bargain freely concerning the allocation of risk ...."); Spring Motors Distribs., Inc. v. Ford Motor Co., 489 A.2d 660, 672 (N.J. 1985) ("Contract principles ... are generally more appropriate for determining claims for consequential damage that the parties have, or could have, addressed in their agreement.") ; Grynberg v. Questar Pipeline Co., 70 P.3d 1, 11 (Utah 2003) ("Whether the doctrine is asserted in terms of economic loss or independent duty, the underlying reasoning remains the same: tort law should govern the duties and liabilities imposed by legislatures and courts upon non-consenting members of society, and contract law should govern the bargained-for duties and liabilities of persons who exercise freedom of contract."); RESTATEMENT (THIRD) OF ECONOMIC TORTS AND RELATED WRONGS [section] 8 cmt. b (Preliminary Draft No. 2, 2006) (noting that rationales for the economic loss rule include "[e]ncourag[ing] private ordering" and "preserving the priority of contract law").

(111.) The above analysis is significantly different from other arguments for excluding fraud from the economic loss rule. For instance, Tourek, Boyd, and Schoenwetter maintain that the rule should not apply to fraud because, among other things, "[a] s a practical matter, it is difficult to see how a party can effectively anticipate and therefore negotiate the allocation of risks for every type of misrepresentation and every form of deceit that may be perpetrated by the other party to an agreement." Tourek et al., supra note 103, at 916; see also Stoughton Trailers, Inc. v. Henkel Corp., 965 F. Supp. 1227, 1236 (W.D. Wis. 1997) ("A party to a contract cannot rationally calculate the possibility that the other party will deliberately misrepresent terms critical to that contract."); Dobbs, supra note 104, at 731 ("Where the parties have not agreed to subject a given risk or dispute to the contract's terms, to apply the economic loss rule is not to honor the contract but rather to impose a contract limitation where none was intended...."). My argument is exactly the opposite: when it comes to contractually required representations, the parties might anticipate and prefer allocating the risk of misrepresentation in accordance with the law of fraud, but the expansive reading of the economic loss rule prevents them from doing so.

(112.) See HTP, Ltd. v. Lineas Aereas Costarricenses S.A., 685 So. 2d 1238 (Fla. 1996); Huron Tool & Eng'g Co. v. Precision Consulting Servs., Inc., 532 N.W.2d 541 (Mich. Ct. App. 1995); Kaloti Enters., Inc. v. Kellogg Sales Co., 699 N.W.2d 205 (Wis. 2005) ; see also AKA Distrib. Co. v. Whirlpool Corp., 137 F.3d 1083 (8th Cir. 1998) (applying Minnesota law); Cooper Power Sys., Inc. v. Union Carbide Chems. & Plastics Co., 123 F.3d 675 (7th Cir. 1997) (applying Wisconsin law).

(113.) For a summary overview, see Dorff, supra note 80, at 407-11. For other examples of how these doctrines have been applied to concealment of breach cases, see Catherine Paskoff Chang, Note, Two Wrongs Can Make Two Rights: Why Courts Should Allow Tortious Recovery for Intentional Concealment of Contract Breach, 39 COLUM. J.L. & Soc. PROBS. 47, 66-72 (2005).

Not all courts exclude liability from the contract setting. See, e.g., Morrill v. Becton, Dickinson & Co., 747 F.2d 1217, 1222 (8th Cir. 1984) ("[U]nder Missouri law, liability in tort may co-exist with liability in contract arising out of the same events."); Life Ins. Co. v. Murray Inv. Co., 646 F.2d 224, 228 (5th Cir. Unit A May 1981) ("[U]nder Texas law, if the act complained of constitutes both a breach of contract and a willful tort, exemplary damages are recoverable."). And the tendency to exclude tort is of relatively recent origin. See WILLIAM LLOYD PROSSER, The Borderland of Tort and Contract, in SELECTED TOPICS ON THE LAW OF TORTS 380 (1954). Prosser himself was in favor of a broad overlap between tort and contract: "When the ghosts of case and assumpsit walk hand in hand at midnight, it is sometimes a convenient and comforting thing to have a borderland in which they may lose themselves." Id. at 452.

(114.) 507 S.E.2d at 347; see also Americana Petroleum Corp. v. Northville Indus. Corp., 606 N.Y.S.2d 906, 908 (App. Div. 1994) ("[I]n order to state a cause of action sounding in fraud, the plaintiff must allege a breach of duty which is collateral or extraneous to the contract between the parties.").

(115.) Cimino v. FirsTier Bank, 530 N.W.2d 606, 613 (Neb. 1995).

(116.) Textron Fin. Corp. v. Nationwide Mut. Ins. Co., 684 N.E.2d 1261, 1271 (Ohio Ct. App. 1996) (emphasis omitted); see also Chachere v. Drake, 941 S.W.2d 193, 197 (Tex. App. 1996) ("The mere availability of a tort-based theory of recovery is not sufficient; actual damages sustained from an independent tort must be proven before punitive damages are available.").

(117.) AYRES & KLASS, supra note 6, at 59-82.

(118.) Sony Music Entm't Inc. v. Robison, No. 01 Civ. 6415, 2002 WL 272406, at * 2 (S.D.N.Y. Feb. 26, 2002). Representatives of the Beatles have recently made similar allegations against Capitol Records. See Apple Corps Ltd. v. Capitol Records, Inc., No. 05-604385, slip op. at 45 (N.Y. Sup. Ct., Aug. 23, 2006).

(119.) Sony Music, 2002 WL 272406, at *2. The district court dismissed the claim based on a line of cases of dubious authority that suggest promissory fraud is not actionable in New York. See Ian Ayres & Gregory Klass, Promissory Fraud, N.Y. ST. B. ASS'N J., May 2006, at 26, 27.

(120.) 901 So. 2d 1, 6-9 (Ala. 2004).

(121.) 891 A.2d 1032, 1036 (Del. Ch. 2006); see also id. at 1056-62 (discussing authorities); RESTATEMENT (SECOND) OF CONTRACTS [section] 195(1) (2006) ("A term exempting a party from tort liability for harm caused intentionally or recklessly is unenforceable on grounds of public policy."); Davis, supra note 5, at 488-92, 513-20 (summarizing relevant holdings).

(122.) "With each royalty payment it sent to the State, Exxon effectively represented that (i) it was paying royalty from gross when it in fact was not, (ii) its gross was less than it in fact was, (iii) it was paying the State for all gas produced when it in fact was not and/or (iv) it owed the State less than it in fact did." Post-Judgment Order, supra note 22, at 26. The trial court's theory of fraud was unfortunately expansive in several other directions as well. For one thing, the court permitted a finding of fraudulent suppression based on the fact that Exxon withheld or provided only partial royalty reports. Id. at 27; see also Grynberg v. Citation Oil & Gas Corp., 573 N.W.2d 493, 498-99 (S.D. 1997) (royalty reports not itemized and withheld were contrary to contract). Unlike the case where a breaching promisor is contractually required to report breach and remains silent, here the absence of required reports or information should have put Alabama on notice that there might be a first-order breach. Elsewhere the trial court appears to endorse a broader action for bad faith breach: "Alabama law surely does not permit a business party such as Exxon to breach a contract as it sees fit without fear of punitive damages." Post-Judgment Order, supra note 22, at 23.

(123.) See, e.g., Rea, supra note 63, at 159 ("[T]he parties to a contract are unlikely to agree ex ante to damages that exceed the expected loss."); Schwartz, supra note 64, at 370 ("[P]romisees do not want contractual damage measures that would grant more than their lost expectation.").

(124.) For more detailed versions, see Richard Craswell, Contract Remedies, Renegotiation, and the Theory of Efficient Breach, 61 S. CAL. L. REV. 629 (1988); Rea, supra note 63, at 151-63; Schwartz, supra note 64, at 372-83; and Talley, supra note 60, at 1212-18. The core argument presented here omits some considerations, such as the use of penalties as signals, their employment as strategic barriers to entry, and their effect on price and selection, as well as economic arguments that, in special circumstances, extracompensatory remedies can promote efficient investment. For a detailed overview of the literature, see Edlin & Schwartz, supra note 4

(125.) See RICHARD A. POSNER, ECONOMIC ANALYSIS OF LAW 117-18 (6th ed. 2003); Talley, supra note 60, at 1218-41. But see Clarkson et al., supra note 60, at 360-62; Dodge, supra note 59, at 632-33, 666-76 (arguing that renegotiation is possible) ; Goetz & Scott, supra note 63, at 587.

(126.) This moral hazard problem is emphasized by Clarkson et al., supra note 60, at 368-72. See also Craswell, supra note 124, at 646-53, 656-61; Rea, supra note 63, at 155.

(127.) See Craswell, supra note 124, at 642-45; Richard Craswell, Deterrence and Damages: The Multiplier Principle and Its Alternatives, 97 MICH. L. REV. 2185, 2230 (1999); Rea, supra note 63, at 152-54; Schwartz, supra note 64, at 392.

(128.) Samuel Rea summarizes the arguments as follows: "Penalty clauses are costly because they (1) induce excessive precautions by the promisor, (2) induce deficient precautions by the promisee, (3) overinsure the promisee, and (4) expose the promisor to additional risk." Rea, supra note 63, at 156.

(129.) See Sanchirico, supra note 17, at 1352-60.

(130.) For a detailed analysis of the concept of opportunistic breach, see Timothy J. Muris, Opportunistic Behavior and the Law of Contracts, 65 MINN. L. REV. 521, 521-26 (1981). See also Charles J. Goetz & Robert E. Scott, Principles of Relational Contracts, 67 VA. L. REV. 1089, 1139 n.118 (1981) ("[O]pportunistic behavior only redistributes portions of an already allocated contractual pie."); Barry Perlstein, Crossing the Contract-Tort Boundary: An Economic Argument for the Imposition of Extracompensatory Damages for Opportunistic Breach of Contract, 58 BROOK. L. REV. 877, 880 (1992) ("The opportunistic actor creates more value for himself, but only by taking an equivalent amount or more from others.").

(131.) Muris, supra note 130, at 524.

(132.) POSNER, supra note 125, at 118; see also Dodge, supra note 59, at 654-62 (arguing that punitive damages for opportunistic breach result in greater efficiency); Richard A. Posner, Common-Law Economic Torts: An Economic and Legal Analysis, 48 ARIZ. L. REV. 735, 745-47 (2006) (discussing when punitive damages for breach are appropriate).

Several academics have suggested permitting penalty clauses for opportunistic breach only, though they have proposed different tests for distinguishing opportunistic from efficient breach. See Dodge, supra note 59, at 633 n.14 (listing authors who support punitive damages for opportunistic breach). To the extent that obstructive breaches are very likely to be opportunistic, my suggestion is consonant with these recommendations.

(133.) Efforts to avoid obstruction are wasted when they cost the promisor more than they benefit the promisee (by providing her needed assurance that she will recover for a first-order breach). At the time of formation, neither party wants a term that will result in such inefficient behavior, which decreases the total value of the transaction that the parties have to divide between themselves. In other words, the promisor is likely to pass some or all of the costs of wasted effort on to the promisee in the form of a higher price.

(134.) See Dodge, supra note 59, at 651-98 (recommending punitive damages for willful, as opposed to involuntary, breach).

(135.) This is another way of putting the familiar observation that extracompensatory remedies can be a cost-effective means of communicating a high probability of performance. See POSNER, supra note 125, at 128; Rea, supra note 63, at 156-57.

(136.) Sanchirico, supra note 17, at 1352-82. One difference between my category of obstruction of recovery and Sanchirico's detection avoidance is that Sanchirico is mainly interested in criminal and regulatory sanctions. As a result, he focuses on how to prevent attempts to avoid punitive first-order sanctions, a problem with a slightly different structure than trying to prevent attempts to avoid compensatory damages.

(137.) Id. at 1337; see also id. at 1352-60 (describing costs of detection avoidance).

(138.) The "and so on" leads Sanchirico to make another, more ambitious claim:
   Sanctioning activity X encourages another activity X+1 in the form
   of effort exerted to avoid detection of X by those who still choose
   to engage in X.

   Stating the principle in these general terms makes clear that it
   is recursive. Because the formula applies to any activity X, we are
   free to substitute "detection avoidance" itself for X, whereby it
   begets an X+1 equal to effort exerted to avoid detection of
   detection avoidance. Indeed, nothing stops us from returning to the
   formula with "detection avoidance of detection avoidance,"
   substituting this for X, and generating, as X+1, effort exerted to
   avoid detection of detection avoidance of detection avoidance. And
   we may continue like this ad infinitum, repeatedly inputting the
   last application's output.


Id. at 1368. I do not think the recursivity claim is essential to Sanchirico's argument as a whole. In any case, it appears overstated. One typically says that a series is "recursive" only if its definition applies the same function to each member of the series to arrive at the next. But as Sanchirico recognizes, the incentives to engage in detection avoidance vary from one level to the next. Id. at 1369. A wrongdoer's reasons to engage in X+1-level detection avoidance depend not only on the penalty for X-level detection avoidance, but also on the out-of-pocket cost of X+1-level detection avoidance, its chances of success, and its probability of detection. The latter factors vary stochastically from one level to the next, depending on the context. There is no a priori reason to think that there will not be some level of detection avoidance where the next higher level will, as a factual matter, be so difficult that it will not be worth the cost (perhaps even without a higher-order penalty), putting an end to the regress.

(139.) On the potential economic utility of damage multipliers, see Robert D. Cooter, Punitive Damages for Deterrence: When and How Much?, 40 ALA. L. REV. 1143 (1989) ; and A. Mitchell Polinsky & Steven Shavell, Punitive Damages: An Economic Analysis, 111 HARV. L. REV. 869 (1998). For some of the complications in establishing the proper multiplier and alternative approaches, see Craswell, supra note 127; and Keith N. Hylton & Thomas J. Miceli, Should Tort Damages Be Multiplied?, 21 J.L. ECON. & OP, G. 388 (2005).

The statement that multiplied compensatory damages give the promisor the right incentive to perform glosses over a number of complexities in the attempt to use a single remedy to induce both sides to behave efficiently throughout the entire transaction. See generally Craswell, supra note 124; Richard Craswell, In That Case, What Is the Question? Economics and the Demands of Contract Theory, 112 YALE L.J. 903, 907-10 (2003).

(140.) In practice, most multipliers are not simply dynamic or static, but somewhere in between, depending on how fine-grained or fact-dependent they are. The idea of a "perfectly" dynamic multiplier is a theoretical fiction, useful in my analysis, but not to be mistaken as a description of how any multiplier actually works.

Richard Craswell reports: "[M]ost legal systems do not use multipliers that are calculated case by case. Instead, they use multipliers set at the same level for all defendants, or fines set at the same level for all defendants, or compensatory damages with no multipliers at all." Craswell, supra note 127, at 2198. Craswell's article provides the most comprehensive examination of the relative merits of static and dynamic multipliers. Other proponents of multipliers often ignore the issue. Polinsky and Shavell, who are relatively optimistic about the ability of factfinders to calculate the probability of enforcement, still recognize the danger of error and recommend that "the legislature ... set damages multipliers for separate categories of wrongful conduct, based on rough assessments of the different chances of escaping liability in the various settings." Polinsky & Shavell, supra note 139, at 893.

(141.) This point is a variation on Robert Cooter's observation that when "lawmakers can identify socially desirable behavior, but are prone to error in assessing the cost of deviations from it, then sanctions are preferable to prices." Robert Cooter, Prices and Sanctions, 84 COLUM. L. REV. 1523, 1524 (1984). The multiplier option shows that the relevant difficulty can include not only "assessing the costs of deviation," but also assessing the probability of enforcement.

(142.) Craswell, supra note 127, at 2193-94.

(143.) Alan Schwartz has identified a related problem with static multipliers: because the probability that the promisee will sue is endogenous, a function of the damage measure itself, the parties will find it difficult to agree on the right multiplier. Schwartz, supra note 64, at 399-40l. And Ian Weiner notices a similar effect in criminal law: "Raising the sanction for the original crime may be counterproductive because it will increase the incentives for criminals to avoid arrest." Ian C. Wiener, Running Rampant: The Imposition of Sanctions and the Use of Force Against Fleeing Criminal Suspects, 80 GEO. L.J. 2175, 2182 (1992).

(144.) Jennifer Aden has observed that variable fines for corporate wrongdoing could, in theory, achieve similar benefits in the criminal context. Jennifer Aden, The Potentially Perverse Effects of Corporate Criminal Liability, 23 J. LEGAL STUD. 833, 849 (1994). Like me, Aden thinks the costs of case-by-case calculations of the probability of enforcement "place this solution beyond the realm of realistic alternatives." Id.

(145.) A society of econometric superheroes could address imperfect verification of obstruction by adding to the generic first-order multiplier a generic obstruction multiplier. Damages for first-order breach would then be multiplied twice. The first multiplier would take account of the probability of enforcement of the first-order breach, given all that is known about the situation, while proof of obstruction would trigger a second multiplier, designed to account for the possibility that those obstructive tactics might have gone undetected or unproven.

Mere mortals will find this a difficult solution. First, calculating the appropriate multiplier would be more complex than simply multiplying by the reciprocal of the probability that the obstructive behavior in question would go unverified. Such simple multipliers work only where nonverification results in no damage award. In a dynamic multiplier regime, no proof of obstructive behavior does not result in no award, but in an award that is too low. Second, the problem of quantification is even more intractable in the case of obstruction multipliers, which require the factfinder to evaluate the likelihood that bad behavior that is designed to be hidden but has been discovered and proven might not have been.

(146.) See Keith N. Hylton & Thomas J. Miceli, Should Tort Damages Be Multiplied?, 21 J.L. ECON. & ORG. 388, 403-11 (2005) (deriving generic multipliers from empirical studies of different categories of tort litigation).

(147.) See generally Daniel Kahneman et al., Shared Outrage and Erratic Awards: The Psychology of Punitive Damages, 16 J. RISK & UNCERTAINTY 49 (1998); Cass R. Sunstein et al., Assessing Punitive Damages (with Notes on Cognition and Valuation in Law), 107 YALE L.J. 2071, 2111-12 (1998); W. Kip Viscusi, The Social Costs of Punitive Damages Against Corporations in Environmental and Safety Torts, 87 GEO. L.J. 285, 327-32 (1998). But see David Luban, A Flawed Case Against Punitive Damages, 87 GEO. L.J. 359, 369-70 (1998) (questioning Viscusi's analysis of juror biases).

(148.) See Polinsky & Shavell, supra note 139, at 892.

(149.) See Craswell, supra note 124, 664-65 (discussing risk aversion and the effect of multipliers on precaution decisions); Polinsky & Shavell, supra note 139, at 886-87 (noting effect of risk-aversion on optimal multipliers).

(150.) See generally MINORITY STAFF OF THE SPEC. INVESTIGATIONS DIV. OF THE H. COMM. ON GOV'T REFORM, HALLIBURTON'S PERFORMANCE UNDER THE RESTORE IRAQ OIL 2 CONTRACT (2006).

(151.) Id. at 13.

(152.) James Glanz, Rebuilding of Iraqi Oil Pipeline as Disaster Waiting To Happen, N.Y. TIMES, Apr. 25, 2006, at A1. The Times quotes a senior Oil Ministry official, who "said he began hearing rumors from Iraqis in the ministry in Baghdad that something had gone terribly wrong, but the company itself seemed determined not to clarify what had happened. 'We couldn't get a good status report,' Mr. Vogler said. 'We kept asking for it.... We couldn't get one.'" Id. The Army Corps of Engineers colonel in charge of the project told the Times that "KBR provided him with optimistic assessments nearly to the end of the line ... and he was convinced that the project would be a success." Id.

(153.) 10 U.S.C. [section] 2313(a)-(c) (2007).

(154.) Section 2313 stipulates that it is not to be read to require government contractors to keep any records other than those they would otherwise maintain in the normal course of business. Id. [section] 2313(c)(3).

(155.) MINN. STAT. [section] 604.10(e) (2000); see also Tourek et al., supra note 103, at 927-38 (describing history of Minnesota statute).

(156.) See generally Ian Ayres & Robert Gertner, Majoritarian vs. Minoritarian Defaults, 51 STAN. L. REV. 1591 (1999).

(157.) See, e.g., KIMEL, supra note 13, at l00-l09 (arguing against specific performance remedy based on Mill's harm principle); Ian R. Macneil, supra note 60 (applying reliance theory to justify rule against penalty clauses); Ernest J. Weinrib, Punishment and Disgorgement as Contract Remedies, 78 CHI.-KENT L. REV. 55, 84-102 (2003) (maintaining that principles of corrective justice argue against punitive damages for breach of contract). But see Curtis Bridgeman, Note, Corrective Justice in Contract Law: Is There a Case for Punitive Damages?, 56 VAND. L. REV. 237 (2003) (applying corrective justice considerations to argue for punitive damages in contract).

Special mention should be made of Daniel Markovits's suggestion that punitive damages might be appropriate whenever a promisor "refuses to pay the compensatory expectation damages that vindicating contractual collaboration requires." Daniel Markovits, Contract and Collaboration, 113 YALE L.J. 1417, 1510 (2004). What Markovits is describing here seems to be, in my terminology, a duty not to obstruct, perhaps akin to that created by Seaman's. But I believe that Markovits's argument entails that both the obligation and the remedy are mandatory, since both derive from the promisor's moral duty to treat the promisee as an end in herself. Cf. id. at 1505-08 (arguing that the parties should not be allowed to contract for reliance or other sub-expectation damage measures).

(158.) I borrow this way of posing the question from Alan Klevorick:
   [T]he critical observation is that the explication of why some acts
   are crimes while others are not requires an inquiry into the
   legitimation of the transaction structure. It forces one to
   confront questions like: Why does the collectivity have the right
   to decide the terms on which particular transactions will take
   place under different circumstances? Why do some rights reside in
   the individual while others rest with the state?


Alan Klevorick, The Economics of Crime, in 27 CRIMINAL JUSTICE: NOMOS 288, 303 (J. Roland Pennock & John W. Chapman eds., 1985); see also Jules Coleman, Crimes and Transactions, 88 CAL. L. REV. 921, 924 (2000) ("[S]omeone who violates the transaction structure has asserted an authority that he does not possess. The power to set the terms of legitimate transfer resides in the political sovereignty and not in ordinary folk.").

Associate Professor, Georgetown University Law Center. I want to thank Jennifer Arlen, Guido Calabresi, Richard Craswell, Kevin Davis, Chris Elmendorf, John Mikhail, Jed Purdy, Chris Sanchirico, Christian Turner, Bill Vukowich, Ethan Yale, and Kathy Zeiler, each of whom added to the ideas in this Article through discussion or comment. I owe special thanks to Ian Ayres, who suggested that I write down these thoughts during a walk in the Arizona desert. The Article also benefited from faculty workshops at the University of Virginia Law School and Georgetown University Law Center, and from discussion at the Stanford/Yale Junior Faculty Forum. Jason Daniels provided excellent research assistance. Finally, I thank Sanjukta Misra, who not only provided input along the way, but is the work's sine qua non.
COPYRIGHT 2007 Yale University, School of Law
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2007 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Klass, Gregory
Publication:Yale Law Journal
Date:Oct 1, 2007
Words:30348
Previous Article:Sheltering deprivations: FEMA, Section 408 housing, and procedural redesign.
Next Article:"I did not come here to defend myself": responding to war on terror detainees' attempts to dismiss counsel and boycott the trial.
Topics:

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters