The article benefitted greatly from comments by David Enoch, Ehud Guttel, Alon Harel, Assaf Jacob, Christine Jolls, Alon Klement, Barak Medina, Liav Orgad, Ivan Reidel, Uzi Segal, Steven Shavell and Eyal Zamir. The article was presented in seminars at the Harvard Law School, the Hebrew University and the Interdisciplinary Center (IDC), Herzliya; I am grateful for the many valuable comments and observations made in those events. I am especially indebted to Ariel Porat for his insights and critical comments and to Richard Bronaugh for his perceptive editing and his important and helpful suggestions. For excellent research assistance I thank Stav Cohen. I also gratefully acknowledge the research support provided by the John M. Olin Center for Law, Economics and Business at Harvard Law School.

1. For a review of the Canadian law of such mistakes see, e.g., Waddams, SM, The Law of Contracts, 5th ed (Toronto: Canada Law Book, 2005) at 272-84Google Scholar [Waddams]; Fridman, GHL, The Law of Contract in Canada, 6th ed (Toronto: Carswell, 2011) at 239-82Google Scholar [Fridman]. US law refers to it as “mutual mistake.” See Restatement (Second) of Contracts, §152 (1981) [Restatement].

2. See Waddams, supra note 1 at 254-71. US law distinguishes between “frustration of purpose” on the one hand and “commercial impracticability” and “impossibility” on the other. Whereas the former refers to a contingency preventing the fulfllment of the contract’s purpose, the latter refers to a contingency rendering performance impossible or prohibitively costly. See Restatement §§261-72; Uniform Commercial Code §2-615 [UCC]; Farnsworth, E Allan, Farnsworth on Contracts, 3rd ed, Vol II (Aspen, 2004) at 624-81Google Scholar[Farnsworth]. In some cases, these doctrines have also been applied to misconceptions of facts that were pre-existing at the time of contracting. See Restatement §266.

3. In English common law, “common mistake” results in the contract being void, although in the past it could also render the contract voidable in equity. Frustration renders the contract voidable. See Furmston, MP, Cheshire, Fifoot & Furmston’s Law of Contract, 15th ed (Oxford: Oxford University Press, 2007) at 286-87, 299-300, 738-41Google Scholar; Peel, E, Treitel on the Law of Contract, 12th ed (London: Sweet & Maxwell, 2007) at 328-30, 972-80Google Scholar. In Canadian law common mistake also renders the contract void, although it is suggested that the traditional rule of voidability under equity may still apply. As under English law, frustration renders the contract voidable. See McCamus, J, The Law of Contracts (Toronto: Irwin Law, 2005) at 526-27, 545-52, 566-67, 603Google Scholar; Waddams, supra note 1 at 278; Fridman, supra note 1 at 251-52, 256-58, 637-38. Under US law, both common mistake and frustration render the contract voidable. See Farnsworth, supra note 2 at 610-11, 664. I use the term “relief” to refer to either voidness or voidability.

4. For an elaborate discussion of these arguments see Section III.

5. See, e.g., Savage, Leonard J, The Foundation of Statistics (New York: John Wiley & Sons, 1954)Google Scholar; Luce, R Duncan & Raiffa, Howard, Games and Decisions (New York: John Wiley & Sons, 1957) at 259-326Google Scholar; Dekel, Eddie, Lipman, Barton L & Rustichini, Aldo, “Standard StateSpace Models Preclude Unawareness” (1998) 66 Econometrica 158 CrossRefGoogle Scholar (demonstrating that un-awareness to a possible event cannot be modeled within the confines of a state-space model).

6. There are also more subtle ways in which fault enters contract doctrine. See, e.g., Ben-Shahar, Omri & Porat, Ariel, eds, Fault in American Contract Law (New York: Cambridge University Press, 2010).CrossRefGoogle Scholar

7. Supra note 1.

8. See, e.g., Fridman, supra note 1 at 244-52.

9. See Waddams, supra note 1 at 280-81; Restatement §154.

10. Ibid.

11. 33 NW 919 (Mich 1887).

12. Restatement §§261-64, §§267-72; UCC §2-615; Farnsworth, supra note 2 at 624-50, 660-73.

13. Fridman, supra note 1 at 617-20. See also Restatement §261 cmt. d.: “Performance may be impracticable because extreme and unreasonable difficulty, expense, injury, or loss to one of the parties will be involved A severe shortage of raw materials or of supplies due to war, embargo, local crop failure, unforeseen shutdown of major sources of supply, or the like, which either causes a marked increase in cost or prevents performance altogether may bring the case within the rule stated in this Section”.

14. Ibid.

15. See Fridman, supra note 1 at 623-25.

16. 3 B & S 826, 122 Eng Rep 309 (1863).

17. 363 F 2d 312 (DC Cir 1966).

18. Ibid. (“While it may be an overstatement to say that increased cost and difficulty of performance never constitute impracticability, to justify relief there must be more of a variation between expected cost and the cost of performing by an available alternative than is present in this case.”)

19. Fridman, supra note 1 at 631-34; Restatement §265, cmt. a.

20. Ibid.

21. [1903] 2 KB 740.

22. See Posner, Richard A & Rosenfeld, Andrew M, “Impossibility and Related Doctrines in Contract Law: An Economic Analysis” (1977) 6 J Legal Stud 83 CrossRefGoogle Scholar.

23. See Restatement §261, illustration 6.

24. In a similar vein, the granting of relief may induce the promisee to act cooperatively with the promisor to prevent harm. See Porat, Ariel, “A Comparative Fault Defense in Contract Law” (2009) 107 Mich L Rev 1397 Google Scholar. Note, however, that since the promisee will still not be liable for the promisor’s lost benefit of the bargain, she will not internalize the full cost of her negligent behavior.

25. See Posner & Rosenfeld, supra note 22. Note that, here too, the promisee does not internalize the lost benefit of the promisor, and therefore her incentive falls short from the social optimum.

26. See Bruce, Christopher J, “An Economic Analysis of the Impossibility Doctrine” (1982) 11 J Legal Stud 311 CrossRefGoogle Scholar.

27. See Cooter, Robert, “Unity in Tort, Contract, and Property: The Model of Precaution” (1985) 73 Cal L Rev 1 at 12-13CrossRefGoogle Scholar; Wagner, Gerhard, “In Defense of the Impossibility Defense” (1995) 27 Loy U Chi LJ 55 Google Scholar.

28. See Shavell, Steven, “Damage Measures for Breach of Contract” (1980) 11 Bell J Econ 466 CrossRefGoogle Scholar.

29. See Goldberg, Victor P, “Impossibility and Related Excuses” (1988) 144 J Inst Theor Econ 100 Google Scholar.

30. Relief for common mistake may also raise concerns for inadequate disclosure of information, as well as inefficient investment in the acquisition of information. As demonstrated by Rasmusen and Ayres, from these perspectives it is preferable to either grant relief more generously (for unilateral mistakes as well as common mistakes) or to deny relief altogether. See Rasmusen, Eric & Ayres, Ian, “Mutual and Unilateral Mistake in Contract Law” (1993) 22 J Leg Stud 309 CrossRefGoogle Scholar.

31. When parties’ actions cannot be perfectly verified in court, no legal rule guarantees efficient behavior by both parties. To achieve efficiency, both parties would have to bear the full cost of the adverse contingency at the margin, as in the case in which the promisor bears full liability but pays damages to a third party. For a discussion of such a mechanism, see Cooter, Robert & Porat, Ariel, “Anti Insurance” (2002) 31 J Legal Stud 203 CrossRefGoogle Scholar.

32. Apparently, for that reason Michelle White has suggested abolishing the very distinction between standard breach and failure to perform due to an unforeseen contingency. See White, Michelle, “Contract Breach and Contract Discharge due to Impossibility: A Unified Theory” (1988) 17 J Legal Stud 353. CrossRefGoogle Scholar A similar policy prescription has been suggested by George Triantis, although for a different set of reasons. See Triantis, George, “Contractual Allocations of Unknown Risks: A Critique of the Doctrine of Commercial Impracticability” (1992) 42 UTLJ 450 CrossRefGoogle Scholar. Triantis argues that even when certain contingencies are not specifically contemplated, they may nevertheless be contractually allocated. Thus, for instance, even if a shipper and a carrier fail to consider the possibility that the delivery would become impossible due to the closure of the Suez Canal, they may still contemplate some broader category of risks that includes the realized contingency, e.g., that some misfortune would prevent the carrier from reaching its destination as planned. The contract, Triantis argues, allocates this broader category of risk even if the specific possibility of the Canal’s closure is not specifically contemplated. As all risks are thereby taken into account and contracted for, he concludes, contractual obligations should always be enforced

Yet this argument raises a difficulty. Although one might accept the proposition that all risks could be allocated, it does not follow that it must be optimal for the parties to do so if they are non-contemplated. In fact, the argument advanced here notes that, under the prescribed conditions, the welfare of contracting parties will be enhanced if the governing rule allows them to allocate only contemplated risks.

33. While the new model uses a binary language of “contemplated” or “non-contemplated” contingencies, it is possible, of course, to also conceive of intermediate states of cognition. Awareness of possible risks is perhaps most accurately described as a continuum: There are various degrees of awareness to what one does not know, and as to where one should look in order to discover new, relevant information. The model’s language should therefore not be taken to suggest a dichotomy. Rather, the term “non-contemplation” should be taken to mean that the parties are sufficiently unaware of what they do not know, such that they cannot incorporate the risk—or the actions to be taken in view of the risk—into their agreement. Since the point of the analysis is to identify the desirable legal consequences of such risks, the various cognitive states that produce this result are lumped together in the interest of convenience.

Of course, determination of whether a particular contingency has been contemplated may produce problems of verification, which may in turn affect the properties of the desirable legal rule. This difficulty and its ramifications are discussed in Section III.B.2, infra.

34. See, e.g., Perillo, Joseph M, Calamari and Perillo on Contracts, 5th ed (St. Paul: West Group, 2003) at § 9-30Google Scholar; Gergen, Mark P, “A Defense of Judicial Reconstruction of Contracts” (1995) 71 Ind LJ 45 at 51-59Google Scholar.

35. In the presence of reliance investments, that statement must be qualifed. See infra Section V. 1 for a discussion.

36. See, e.g., Berliner, Baruch, Limits of Insurability of Risks, (Englewood Cliffs: Prentice-Hall, 1982)Google Scholar (analyzing the various reasons for which insurance might not be voluntarily provided, among them the unpredictability of risk); Skogh, Goran, “Development Risks, Strict Liability and the Insurability of Industrial Hazards” (1998) 87 Geneva Pap. Risk Ins. 237 Google Scholar (explaining that the risk associated with certain industrial hazards is unquantifiable and thus uninsurable.); Kunreuther, Howard & Hogarth, Robin M, “How Does Ambiguity Affect Insurance Decisions?” in Dionne, Georges, ed, Contributions to Insurance Economics, (Norwell, MA: Kluwer Academic, 1992)Google Scholar (providing experimental support to the claim that ambiguity in the measure of risk explains insurance firms’ reluctance to offer insurance); Kunreuther, Howard, Hogarth, Robin & Meszaros, Jacqueline, “Insurer Ambiguity and Market Failure” (1993) 7 J Risk & Uncertainty 71 CrossRefGoogle Scholar (reporting experimental results indicating that demanded insurance premiums rise sharply when probability is ambiguous; Boardman, Michelle E, “Known Unknowns: The Illusion of Terrorism Insurance” (2005) 93 Geo LJ 783 Google Scholar (describing the risk of terrorism as a “known unknown” rendering it impossible to insure in a non-regulated market); Skogh, Goran, “Risk-Sharing Institutions for Unpredictable Losses” (1999) 155 J Inst Theoretical Econ 505 Google Scholar (identifying alternative methods for coping with unpredictable risks when insurance is unavailable).

37. Pub L 107-297, 116 Stat 2322 (2002).

38. For further discussion, see Boardman, supra note 36, and Kendall, Jane, “The Incalculable Risk: How the World Trade Center Disaster Accelerated the Evolution of Insurance Terrorism Exclusions” (2002) 36 U Rich L Rev 569 Google Scholar.

39. See Kunreuther, Howard, “The Role of Actuaries and Underwriters in Insuring Ambiguous Risks” (1989) 9 Risk Analysis 319 CrossRefGoogle Scholar (explaining the unavailability of insurance against environmental pollution by the ambiguity associated with that risk.) See also Abraham, Kenneth S, “Environmental Liability and the Limits of Insurance” (1988) 88 Colum L Rev 942 CrossRefGoogle Scholar (discussing the unpredictability of liability imposed for harm associated with environmental pollution).

40. Tol, Richard SJ, “Climate Change and Insurance: A Critical Appraisal” (1988) 26 Energy Policy 257 CrossRefGoogle Scholar (arguing that risks associated with climate change are unquantifiable and, therefore, uninsurable).

41. See Skogh, supra note 36.

42. See, e.g., Katzman, Martin T, “Pollution Liability Insurance and Catastrophic Environmental Risk” (1988) 55 J Risk & Insurance 75 CrossRefGoogle Scholar.

43. Moizer, Peter & Smith, Lisa H, “UK Auditor Liability: An Uninsurable Risk?” (1998) 2 Int’l J Audit 197 Google Scholar (explaining that difficulty to insure arises from lack of evidence regarding the magnitude of risk).

44. Inability on the part of private insurers to estimate the risk of foods has triggered the enactment of the National Flood Insurance Act of 1968, Pub L No 90-448, 82 Stat 572, codified at 42 U.S.C. 4001 et seq., under which the government undertakes the role of an insurer.

45. See Kunreuther, Howard C & Kleindorfer, Paul R, “Insuring Against Country Risks: Descriptive and Prescriptive Aspects” in Herring, R ed, Managing International Risks (Cambridge: Cambridge University Press, 1983)Google Scholar. The problem of uninsurability has been mitigated by the provision of government insurance through the Overseas Private Investment Corporation. See 22 USC. §§ 2191-2200a.

46. This is not to preclude the possibility of other reasons that may prevent insurance coverage for some risks. One such reason, for instance, might be that some risks are not independent, and are therefore difficult to diversify even by an insurer. The point is that the difficulty to quantify risk is accepted as a reason for which an insurance market may not emerge. See supra note 36.

47. Hogarth, Robin & Kunreuther, Howard, “Ambiguity and Insurance Decisions” (1985) 75 Am Econ Rev 386 Google Scholar.

48. In a similar vein, Laure Cabantous asked 78 professional actuaries to price an insurance premium for two types of risk, each with equal likelihood: In the first, a loss would occur with a probability of 0.2%, whereas in the second a loss of the same magnitude would occur with a probability of either 0.1% or 0.3%. As the two risks are actuarially the same, one would expect the premium to be set identically. However, the underwriters priced the second risk at a much higher rate: Whereas the first risk was priced with a loading factor of 35% of the actuarial value, the second was priced at more than twice that amount, at 78%. See Cabantous, Laure, “Ambiguity Aversion in the Field of Insurance: Insurers’ Attitudes to Imprecise and Conflicting Probability Estimates” (2007) 62 Theory & Decision 219 CrossRefGoogle Scholar. See also Kunreuther et al, supra note 36, finding that actuaries, underwriters, and reinsurers recommend considerably higher premiums to be charged for ambiguous risks, and some express outright reluctance to offer any insurance at all. For similar results, see Hogarth, Robin M & Kunreuther, Howard, “Risk, Ambiguity and Insurance” (1989) 2 J Risk & Uncertainty 5 CrossRefGoogle Scholar; Hogarth, Robin M & Kunreuther, Howard, “Pricing Insurance and Warranties, Ambiguity and Correlated Risks” (1992) 17 Geneva Pap Risk Ins 35 CrossRefGoogle Scholar. See also Hershey, John C, Kunreuther, Howard C & Schoemaker, Paul JH, “Sources of Bias in Assessment Procedures for Utility Functions” (1982) 28 Manage Sci 936 CrossRefGoogle Scholar and Thaler, Richard, “Towards a Positive Theory of Consumer Choice” (1980) 1 J Econ Behavior & Org 39 CrossRefGoogle Scholar.

49. Interestingly, this remains so under the assumptions of the conventional model, even if the contract is incomplete as a result of positive drafting costs. Even an incomplete contract distributes surplus in accordance with the parties’ wishes, for if it were otherwise the parties could simply adjust the price. Thus, under the assumptions of the conventional model, even an incomplete contract generates an accurate representation of the parties’ wishes.

50. 63 AD 3d 1145 [2d Dept. 2009].

51. In this particular case, the husband sought reformation rather than release. The alternative of reformation is attractive when the desired distribution of gains is verifiable, as the court saves the parties the trouble of renegotiations and the pitfalls associated with a possible change in relative bargaining power. However, often the desired distribution of gains is not clearly verifiable to the court, in which case the potential cost of judicial error may tip the scales against it.

52. If attitudes are optimistic, then an opposite type of inefficiency may generally occur, namely that parties will contract even when contracting is inefficient. As in the case of pessimistic attitudes, relief will prevent the inefficiency by excluding the risk of non-contemplated events from the domain of contractual obligation. The possibility of inefficient contracting, however, exists only if the parties rely, contrary to the working assumption underlying this section.

53. The phenomenon, widely known as the “Ellsberg Paradox,” has first been demonstrated in Ellsberg, Daniel, “Risk, ambiguity and the Savage axioms” (1961) 75 Q J Econ 643 CrossRefGoogle Scholar. For surveys of the literature, see Camerer, Colin & Weber, Martin, “Recent developments in modeling preferences: Uncertainty and ambiguity” (1992) 5 J Risk & Uncertainty 325 CrossRefGoogle Scholar; Keren, Gideon & Gerritsen, Léonie EM, “On the Robustness and Possible Accounts of Ambiguity Aversion” (1999) 103 Acta Psychol 149 CrossRefGoogle Scholar; Mukerji, Sujoy, “A Survey of Some Applications of the Idea of Ambiguity Aversion in Economics” (2000) 24 Int’l J Approximate Reasoning 221.CrossRefGoogle Scholar

Within the decision-theory literature, the notion of ambiguity has been formalized as a setting in which the likelihood of a future occurrence is consistent with several probability distributions. Accordingly, aversion to ambiguity has been interpreted as the tendency to base one’s estimate on the most pessimistic of those distributions. See Gilboa, Itzhak & Schmeidler, David, “Maxmin expected utility with a non-unique prior” (1989) 18 J Math Econ 141.CrossRefGoogle Scholar For a nontechnical review see Mukerji, ibid at 223-24.

54. This assumption captures the notion that in real-world contracting situations, the possibility of a non-contemplated contingency is always present, and cannot be eliminated simply by way of postponing the moment of contracting to a future time.

55. The threshold value for which this will occur is when p=Ga/(Ga+Gb) for Aaron and q=Gb/ (Ga+Gb) for Betsy.

56. See, e.g., Waddams, supra note 1 at 256-69, 272-81; Fridman, supra note 1 at 239-44, 645-48; White, J & Summers, R, Uniform Commercial Code, 2d ed (1980) at 132-33Google Scholar; Hillman, Robert A, “An Analysis of the Cessation of Contractual Relations” (1983) 68 Cornell L Rev 617 at 650-55Google Scholar.

57. That is the classic “holdup problem.” For the seminal work on this issue, see Williamson, Oliver E, Markets and Hierarchies: Analysis and Anti-Trust Implications (New York: Free Press, 1975)Google Scholar; Williamson, Oliver E, “Transaction Cost Economics: The Governance of Contractual Relations” (1979) 22 J L & Econ 233 CrossRefGoogle Scholar; Klein, Benjamin, Crawford, Robert & Alchian, Armen, “Vertical Integration, Appropriable Rents, and the Competitive Contracting Process” (1978) 21 J L Econ 297 CrossRefGoogle Scholar.

58. Supra note 11.

59. Supra note 21.

60. The fact that Krell could still earn the rental fee was viewed by several commentators as providing the primary justification for the outcome in this case. See, e.g., Wladis, John D, “Common Law and Uncommon Events: The Development of the Doctrine of Impossibility of Performance in English Contract Law” (1987) 75 Geo LJ 1575 at 1619Google Scholar; Birmingham, Robert L, “Why is There Taylor v Caldwell? Three Propositions about Impracticability” (1989) 23 USF L Rev 379 at 393Google Scholar.

61. Supra note 17.

62. See Posner & Rosenfeld, supra note 22 at 103-05.

63. See Restatement § 261, illustrations 9 and 10; in the following cases relief was not awarded, but it is explicitly reasoned that it would have been awarded if the error were more severe: American Trading & Prod Corp v Shell Intl Marine, 453 F 2d 939 (2d Cir. 1972) (“increase of less than one third” over agreed price “is not sufficient to constitute commercial impracticability, under either American or English authority”); Transatlantic Fin Corp v United States, supra note 17 (“While it may be an overstatement to say that increased cost and difficulty of performance never constitute impracticability, to justify relief there must be more of a variation between expected cost and the cost of performing by an available alternative than is present in this case.”)

64. For instance, in the case of Glidden Co v Hellenic Lines, 275 F2d 253 (2d Cir NY 1960), it was determined affirmatively that the parties had in fact contemplated the risk of closure of the Canal, and relief was denied on that basis.

65. Friedman, Milton, “The Methodology of Positive Economics” in Friedman, Milton, Essays In Positive Economics (Chicago: University of Chicago Press, 1953) 3 at 15.Google Scholar

66. Ibid.