Indiana Law Review


Expectations, Loss Distribution and Commercial

Impracticability

Steven Walt*

I. Introduction

The doctrine of commercial impracticability is easy to state and

difficult to apply. Simply stated, the doctrine of commercial impracti-

cability **excuses a seller from timely delivery of goods contracted for,

where his performance has become commercially impracticable because

of unforeseen supervening circumstances not within the contemplation

of the parties at the time of contracting. "• The doctrine is difficult to

apply because of the vague statutory language of section 2-615 of the

Uniform Commercial Code.^ Two of the crucial terms, **basic assump-
tions" and * impracticable," are undefined by the statute and provide

no standards. Nor do the accompanying comments. In fact, the official
comment states that section 2-615 **deliberately refrains from any effort

* Assistant Professor of Law, University of San Diego School of Law. B.A.,

Kalamazoo College, 1976; M.A., University of Chicago, 1978; Ph.D., University of Chicago,

1984; J.D., Yale Law School, 1988.
I thank Larry Alexander, Jules Coleman, and Scott Widen for helpful discussions. I am
particularly grateful to Alan Schwartz for discussion and comments on a previous draft.

Brian Paton provided valuable research assistance. The usual disclaimer applies.

1. U.C.C. §2-615 comment 1 (1987).
2. U.C.C. § 2-615. Excuse by Failure of Presupposed Conditions.

Except so far as a seller may have assumed a greater obligation and subject
to the preceding section on substituted performance:

(a) Delay in delivery or non-delivery in whole or in part by a seller who
complies with paragraphs (b) and (c) is not a breach of his duty under

a contract for sale if performance as agreed has been made impracticable

by the occurrence of a contingency the non-occurrence of which was

a basic assumption on which the contract was made or by compliance

in good faith with any applicable foreign or domestic governmental

regulation or order whether or not it later proves to be invalid.

(b) Where the causes mentioned in paragraph (a) affect only a part of

the seller's capacity to perform, he must allocate production and

deliveries among his customers but may at his option include regular
customers not then under contract as well as his own requirements
for further manufacture. He may so allocate in any manner which is
fair and reasonable.

(c) The seller must notify the buyer seasonably that there will be delay

or non-delivery and, when allocation is required under paragraph (b),

of the estimated quota thus made available for the buyer.

U.C.C. § 2-615.



66 INDIANA LAW REVIEW [Vol. 24:65

at an exhaustive expression of contingencies and is to be interpreted in

all cases sought to be brought within its scope in terms of its underlying

reason and purpose.*'^

Section 2-615 requires four conditions to be satisfied before a party's

performance is excused:"^ (1) the occurrence of a supervening contingency;^

(2) the nonoccurrence of the resulting contingency was a basic assumption

upon formation of the contract;^ (3) the occurrence of the contingency
rendered the agreed performance impracticable;^ and (4) the occurrence

of the contingency was not a risk assumed by the performing party.*

Excuse is also granted when contractual performance is rendered im-
practicable **by compliance in good faith with any applicable foreign

or domestic governmental regulation or order whether or not it later

proves to be invalid.*''

By treating conditions (2) and (4) as equivalent, •^ judicial gloss on
section 2-615 excuses performance when: (1) an unforeseeable contingency

has occurred; (2) the risk of the contingency was not allocated by the

parties; and (3) the occurrence of the contingency rendered performance

commercially impracticable.^' Because some imprecise terms remain in

the judicial reformulation, difficulties in application also remain. Pre-

3. U.C.C. §2-615 comment 1.

4. See generally J. White & R. Summers, Uniform Commercial Code § 3-9, at
155 (3d ed. 1988).

5. U.C.C. § 2-615(a); cf. Restatement (Second) of Contracts § 261 (1979)

(containing similar element).

6. U.C.C. § 2-615(a); cf. Iowa Elec. Light & Power Co. v. Atlas Corp., 467 F.
Supp. 129, 134 (N.D. Iowa 1978); Restatement (Second) of Contracts § 261 (1979)

(containing similar element).

7. Iowa Elec. Light & Power Co., 467 F. Supp. at 134.
8. Id.

9. U.C.C. § 2-615(a); see Eastern Air Lines, Inc. v. McDonnell Douglas Corp.,

532 F.2d 957 (5th Cir. 1976) (government request to give specific military projects priority

over civilian production constituted a "regulation or order" under § 2-615).

The statute and most case law applies only to sellers. See Northern Ind. Pub. Serv.

Co. V. Carbon County Coal Co., 799 F.2d 265 (7th Cir. 1986). For an application to

buyers, see Nora Springs Coop Co. v. Brandau, 247 N.W.2d 744 (Iowa 1974) and U.C.C.

§ 2-615 comment 9 (1987). Cf. Lawrance v. Elmore Bean Warehouse, Inc., 108 Idaho

892, 894, 702 P.2d 930, 932 (Idaho Ct. App. 1985) (dicta stating that § 2-615 of the

Idaho Uniform Commercial Code is available to buyers); Northern 111. Gas Co. v. Energy

Coop., Inc., 122 111. App. 3d 940, 954, 461 N.E.2d 1049, 1060 (1984) (trial court's

allowance of buyer's invocation of § 2-615 of the 111. Commercial Code affirmed).
10. United States v. Wegematic Corp., 360 F.2d 674, 676 (2d Cir. 1966).

11. Transatlantic Fin. Corp. v. United States, 363 F.2d 312, 315 (D.C. Cir. 1966);

cf Neal-Cooper Grain Co. v. Texas Gulf Sulphur Co., 508 F.2d 283, 298 (7th Cir. 1974);

Iowa Elec. Light & Power Co., 467 F. Supp. 129. See also R. Hillman, J. McDonnell
& S. NiCKLES, Common Law And Equity Under The Uniform Commercial Code 7.03,
at 7-15 to 7-17 (1985) (characterizing the conditions for the excuse of impracticability).



1990] COMMERCIAL IMPRACTICABILITY 67

sumably, these difficulties have induced the courts to construe section

2-615 narrowly.*2 As a general rule, courts never excuse sellers' per-
formance when the market-contract differential is positive. A seller's
performance is rarely excused in any case, and then only when per-
formance would result in a large out-of-pocket loss.'' The determination
of the extent of such a loss has not been exact: input cost increases of

1000 percent have been held to excuse performance while increases of

less than 100 percent have not excused performance. '"* Vagueness in the

statutory term "impracticability" provides no principled basis for these

decisions.

Only part of the difficulty in applying section 2-615 is terminological

vagueness. Part is also the absence of a good understanding of contractual

behavior concerning price adjustment.'' A fixed-price contract allocates
the gains from performance of the agreement between the parties. Changes

in the cost of performance over time potentially alter the initial distri-

bution of contractual gains. The parties can take account of the changes

ex ante by adopting a mechanism for altering the contract price. Al-

ternatively, the decision can be postponed or assigned to a third party.

Or the parties can subsequently adjust the contract price without resorting
to the contract at all.'^ Selection of an adjustment mechanism or its

12. Iowa Elec. Light & Power Co., 467 F. Supp. at 134 n.7; see also Jennie-O
Foods, Inc. V. United States, 580 F.2d 440, 409 (Ct. CI. 1978) ("[Tlhis court has not

applied [the doctrine of commercial impracticability] with frequency or enthusiasm.'*);

McGinnis v. Cayton, 312 S.E.2d 765, 775 (W. Va. 1984) (•*[T]he commercial impracticability

doctrine is recognized, but rarely allowed as an excuse for nonperformance.").

13. See, e.g., Jennie-O Foods, 580 F.2d at 409; McGinnis, 312 S.E.2d at 775;

Schwartz, Sales Law and Inflations, 50 S. Cal. L. Rev. 1, 4-5 (1976). See generally
Prance, Commercial Impracticability: A Textual and Economic Analysis of Section 2-615
of the Uniform Commercial Code, 19 Ind. L. Rev. 457 (1986); Stroh, The Failure of

the Doctrine of Impracticability, 5 Corp. L. Rev. 195 (1982).

14. Cf. Transatlantic Fin. Corp., 363 F.2d 312 (12.27<7o); American Trading &
Production Corp. v. Shell Int'l Marine, Ltd., 343 F. Supp. 91 (S.D.N.Y. 1971) (increase

of less than 33<7o); Mineral Park Land Co. v. Howard, 172 Cal. 289, 156 P. 458 (1916)
(l,000<7o); Maple Farms, Inc. v. City School District, 76 Misc. 2d 1080, 352 N.Y.S.2d

784 (N.Y. Sup. Ct. 1974) (23*7o increase not basis for excuse from performance). See also

Wladis, Impracticability as Risk Allocation: The Effect of Changed Circumstances upon

Contract Obligations for the Sale of Goods, 22 Ga. L. Rev. 503, 610-12 (1988); Note.

U.C.C. § 2-615: Defining Impracticability Due to Increased Expense, 32 U. Fla. L. Rev.
516, 535-36 n.l62 (1980).

15. See, e.g., Goldberg, Price Adjustment in Long-Term Contracts, 1985 Wis. L.

Rev. 527, 542; Scott, Conflict and Cooperation in Long-Term Contracts, 75 Calbf. L.

Rev. 2005, 2007 (1987); cf. Narasimhan, Of Expectations, Incomplete Contracting, and
the Bargain Principle, 74 Calif. L. Rev. 1123 (1986) (current doctrines of excuse and

modiHcation dependent on a model of contracting in which all contingencies are allocated

by contractual provision).

16. See Macaulay, Non-Contractual Relations in Business: A Preliminary Study,



68 INDIANA LAW REVIEW [Vol. 24:65

alternatives depends, among other things, on the parties* ability to
forecast the cost of performance over time, the characteristics of the

subject matter of the contract, the opportunities for strategic behavior,

and the reliability of enforcement.'^ Identification of these factors does

not constitute to a theory of the form price adjustment takes. Without

such a theory, courts lack an explanatory ground for intervention by a

finding of commercial impracticability.

Another part of the difficulty is normative. Application of section

2-615 requires a principle of justice for distributing losses due to input

cost increases between the contracting parties. Assume that a seller
contracts with a buyer to provide a good for $1.00. The seller's production
costs at the time are $.80. Subsequently the seller's costs rise to $1.20

and the market price of the good rises to $1.50. Should the seller bear

the entire cost of the price increase of $.50 ($.50 = $1.50 - $1.00)?
Alternatively, should the seller only bear the input price increase in

excess of the contract price with the buyer bearing the rest of the price

increase of the good?

Under the second alternative, the seller would incur a loss of $.20

(-$.20 = $1.00 - $1.20) and the buyer would incur a loss of $.30 ($.30
= $1.50 - $1.20). The buyer's loss of $.30 is the difference between
the market-contract differential ($.50) and that diffential adjusted for

the input cost increase ($.20). Another alternative would be for the

buyer to share in part or all of the price increase. Perhaps the seller

and the buyer should equally share the price increase?

Assume, to simphfy matters, that the seller agreed to bear the risk

of a price variation of +-$.10 (i.e., $.90 < $1.00 < $1.10). If the
seller's performance is excused, the buyer would lose $.10 (-$.10 =

$1.00 - $1.10), the amount of the agreed-to price increase. If the seller's

performance is enforced, the seller would lose $.40 (-$.40 = $1.10 -

$1.50), the amount of the unagreed-to price increase. And if the seller's
performance is enforced and the unagreed-to increment of the price

increase is split equally, the seller would lose $.20 (-$.20 = ($1.10 -

$1.50)/2) and the buyer would gain $.20 ($.20 = ($.50 - $.10)/2). The
selection of one of the above three alternatives is dependent upon the

adoption of a normative principle for distributing loss. That normative

28 Am. Soc. Rev. 55 (1963); cf. White, Contract Law in Modern Commercial Transactions,
An Artifact of Twentieth Century Business Life?, 22 Washburn L.J. 1 (1982) (survey
evidence suggests that chemical companies ignore both contractual and statutory provisions

in allocating products when capacity to fully perform is affected).

17. See Goldberg, supra note 15; Goldberg, Relational Exchange, 23 Am. Behav.

Sci. 337 (1980); Goldberg & Erickson, Quantity and Price Adjustment in Long-Term
Contracts: A Case Study of Petroleum Coke, 30 J.L. & Econ. 369 (1987); Joskow, Price
Adjustment in Long-Term Contracts: The Case of Coal, 31 J.L. & Econ. 47 (1988).



1990] COMMERCIAL IMPRACTICABILITY 69

principle can be considered a principle of justice between the contracting

parties. Because neither section 2-615 nor case law contains an explicit

principle of justice, justifications of case outcomes are defective. Case

outcomes themselves are difficult to predict.'*

In what follows, I offer a principle of justice for applying section

2-615. The principle is * internal" to the Uniform Commercial Code in
two respects: It is consistent with the statutory provisions, and it is

suggested by both the comments to section 2-615 and the applicable

case law. Section II briefly states a criterion of **internalness.*' It applies

the criterion by considering and rejecting a Rawlsian principle of loss

distribution. Section III outlines the standard doctrine of commercial

impractibility as applied. It shows what has already been noted: that

foreseeability is taken to be a normatively significant factor in excusing

sellers* performance. Section IV offers a principle of loss distribution

under which foreseeability is treated as normatively justified. Section V
presents a serious problem for analyses that make use of notions of
foreseeability, denominated as the description problem. A solution to
the problem in commercial settings is then presented. Section VI addresses

some difficulties in applying the proposed principle of loss distribution.
A conclusion and brief comparison with accepted judicial practice are
provided in Section VII.

II. Internalness

There is one constraint on admissible normative principles of justice
under section 2-615: at a minimum, they must be compatible with
plausible elements of institutional (judicial) competence. There is nothing

unusual here. Any adjudicatory principle is under the same constraint.
Institutional competence has two components. One is the informational
prerequisites to be satisfied by a decision maker — a court in this
instance. I delay discussion of this component until Section VI. The
other component is one of institutional authority. Here the component
concerns both the statutory and judicial basis for adopting a normative

principle to apply section 2-615.

Normative principles are "internal" if they are consistent with and

supported by both bases. **Internal" normative principles are required

18. Mueller, Contract Remedies: Business Fact and Legal Fantasy, 1967 Wis. L.
Rev. 833, 837 ("[T]he courts have managed to reach the best solutions on an individual

case basis by the seat of their pants."). This is ironic, considering that Llewellyn, the

drafter of § 8.7 of the Revised Sales Act which was the forerunner to § 2-615 and its

accompanying comments, intended to eliminate uncertainty in case outcomes. See Hawkland,

The Energy Crisis and Section 2-615 of the Uniform Commercial Code, 79 Com. L. J.
77 (1974) (citing the private notes of Karl Llewellyn). Uncertainty in case outcomes has

been reduced, but in an unintended way: a seller's performance is seldom excused.



70 INDIANA LAW REVIEW [Vol. 24:65

to satisfy conditions of institutional authority. The notion of support
can be characterized, but not precisely defined. It is more than logical
consistency and less than a member of a set of deductive consequences.
Because **internalness*' depends on an authoritative basis, institutional

competence is not just a matter of having a comparative advantage in

access to information. Institutional competence is not just a matter of

being less subject to informational constraints.'^ It also requires support

by an * internal" normative adjudicatory principle.

Rawlsian principles of loss distribution illustrate ** external'* prin-

ciples. Consider first a version of such principles.^° The version depends
on applying a Rawlsian choice situation and decisional machinery to
commercial contexts. On that application, seller and buyer are in a state
of partial ignorance. Both are ignorant of existing legal rules for assigning

losses due to unforeseeable supervening events. But both parties know
that they have entered into a contract and that there is a positive

probability of such events occurring. Seller and buyer are to select a
principle of loss distribution.

The Rawlsian analogue should be familiar. Seller and buyer are in
an original position, subject to a "thin'* veil of ignorance. Given the

Rawlsian analogue, both are to adopt the maximin rule.^' The maximin
rule is equivalent to the minimax rule because losses are being dis-

tributed.2^ Under the minimax rule, every act aj. . .sl^ performed in states
of the world Sj. . .,Sn, is assigned a negative utility payoff Ujj, i = 1,. . .,
m, j = 1,..., n. An index is assigned to each act a^ such that it
minimizes the numbers Uil,Ui2. . .jUjn. The minimax rule instructs decision

19. See G. Calabresi, A Common Law for the Age of Statutes 51 (1982) for
a limitation of institutional competence in discerning majoritarian preference. The limitation

in effect confines competence to informational access. The operative notion of competence

above is broader, because it includes institutional authority. For a use of the notion of

institutional competence which is indifferent to the distinction between informational and

normative competence, see Summers, General Equitable Principles Under Section 1-103

Of The Uniform Commercial Code, 72 Nw. U.L. Rev. 906, 912-13 (1978) (**Indeed, with
respect to most commercial issues, legislators have no institutional competence superior

to that of courts.").

20. By calling the principle "Rawlsian," I do not mean to suggest that Rawls
would endorse it. For an endorsement of this principle of loss distribution, as well as

the model of which it is a part, see Harrison, A Case for Loss Sharing, 56 S. Cal. L.
Rev. 573, 595-99 (1983). See also Gillette, Commercial Rationality and the Duty to Adjust

Long-Term Contracts, 69 Minn. L. Rev. 521, 569-70 (1985).
21. Harrison, supra note 20, at 597. Harrison advocates the maximin rule as an

appropriate decision rule for individual choice under uncertainty. See infra note 22 and

accompanying text. For an application of the maximin rule to two-person games, with a

concomitant caution about its appropriateness, see M. Shubik, Game Theory in the
Social Sciences 218-221, 368-69 (1982).

22. R. Luce & H. Raiffa, Games and Decisions 279 (1957).



1990] COMMERCIAL IMPRACTICABILITY 71

makers to select the act that minimizes the maximum loss from selecting
an act. That is, the decision maker is instructed to select that aj with

the lowest associated index. Applied to the example at the beginning,

in which each payoff equals a(Uij), the buyer and the sellers' decision

matrix looks like this:

State

Strategy

equal share (aj)

no share (a2)

partial share (aj)

loss assigned to other loss assigned to self

(s.) (S2)

25 25

50

10 40

Figure 1

The minimax rule would select aj, the equal share strategy,^^ because
only ai minimizes buyer and sellers' maximum loss. Strategy a.^ would
not be selected because, by assumption, the seller and the buyer are

ignorant of their respective identities in the analogue original position.

Thus, the maximum loss under aj is 40, not 10. The equal share strategy,
represented by ai, is the Rawlsian principle of loss distribution.

The Rawlsian principle is an **external" one on the criterion presented
three paragraphs back. It is consistent with the provisions of the Uniform

Commercial Code. However, it is supported by neither case law nor the

Code's provisions or comments. There is no mention of equal share

provisions or allocative principles between merchants.^"* Likewise, of

23. A rule prescribing the maximization of expected utility (minimizing expected
disutility) would be indifferent between a.-a, if pCs/a,) = pCs/aj) = pCs/a,) = p(Si) and

P(Si) = p(s2). For endorsement of the equal share distribution as being ''eminently sensible,"
see Mueller, supra note 18, at 837.

24. Most of the provisions of Article Two pertaining to merchants concern the
formation of contracts. See, e.g., U.C.C. §§2-201(2) (between merchants, a written

confirmation signed by sender satisfies the Statute of Frauds writing requirement), 2-205

(offer by a merchant in writing giving assurance that offer will not be revoked is irrevocable

for the stated time or if no time is stated for a reasonable time), 2-207(2) (between

merchants, terms in an offer-varying acceptance become part of the contract unless the

offer precludes the variance of terms, the terms materially alter contract, or the offeror

seasonably notifies, the offeree of an objection to their inclusion). The few allocative

provisions pertaining to merchants have nothing to do with the merchant's state of

ignorance. See, e.g., U.C.C. §§ 2-509(3) (in instances not covered by subsections (1) or

(2) of § 2-509, the risk of loss shifts to the buyer upon his receipt of the goods provided

the seller is a merchant), 2-5 10(2),(3) (the seller or the buyer may treat the risk of loss
as resting with the breaching party to the extent of any deficiency in the nonbreaching

party's insurance coverage).



72 INDIANA LAW REVIEW [Vol. 24:65

course, the assumption of partial ignorance is not supported." In fact,

given that many provisions and comments make reference to prior case
law, 2^ the assumption is unsupported. Ignorance of an existing rule for

allocating loss is a necessary condition for the applicability of the analogue

original position. By referring to prior case law, the comments to section
2-615 make knowledge of prior legal rules pertinent. So, despite the
imprecision of the operative notion of support discussed above, a Rawl-

sian principle has no basis in the relevant statutory or case law.

It is also unjustified. To begin with, the application of Rawlsian
principles to commercial settings is inappropriate. Rawls's choice situation

and decisional machinery are applicable only to the selection of principles

governing the basic structure of society. Rawls takes the basic structure

to include the major political, social, and economic institutions. Com-
mercial contracts are frequently local transactions. Principles of loss

distribution for such transactions, therefore, do not involve the basic

structure of society. Rawls himself has emphasized the distinction between

local transactions and the basic structure. ^^ It is a nonsequitur to transfer

Rawls's choice situation and decisional machinery to the selection of

principles governing local transactions. Rawls states that *'the maximin

criterion is not meant to apply to small-scale situations, say, to how a
doctor should treat his patients or a university its students. For these

situations different principles will presumably be necessary. Maximin is

a macro not a micro principle. **^^ He further asserts that "what are

25. See, e.g., U.C.C. §2-509(3) (the risk of loss shifts to the buyer upon the

buyer's receipt of goods provided that the seller is a merchant; otherwise, the risk of loss

passes to the buyer upon seller's tender of delivery). Provisions of Article Nine concerning

priorities of security interests in the same collateral are also indifferent to a party's

ignorance. See, e.g., U.C.C. §§ 9-312(5) (1987) (first party to file or perfect a security

interest has priority even if party having priority knew of existence of competing security

interest), 9-312(4) comment 5, example 1 (purchase money security interest in noninventory

has priority over competing nonpurchase money security interest if properly perfected;

knowledge of competing security interest is irrelevant), 9-307(1) (buyer in the ordinary

course of business from a nonmerchant farmer takes free of a security interest created

by his seller even if buyer knows of its existence).

26. E.g., U.C.C. § 1-103 (indirectly via reference to "the principles of law and

equity"); see also U.C.C. § 2-615, comment 4 and cited cases.
27. See Rawls's writings following A Theory of Justice (1971): Rawls, Some

Reasons for the Maximin Criterion, 64 Am. Econ. Rev. 141, 141-42 (1974); Rawls, A
Weil-Ordered Society, in Politics, Philosophy and Society, Fefth Series 1, 16 (P. Laslett

& J. Fishkin eds. 1979); Rawls, The Basic Structure as Subject, in Values and Morals
47, 59-60 (A. Goldman & J. Kim eds. 1978); Rawls, The Basic Liberties and Their Priority,
in Liberty, Equality and Law: Selected Tanner Lectures on Moral Philosophy 1,
15 (S. McMurrin ed. 1987) [hereinafter Basic Liberties]', Rawls, The Idea of an Overlapping

Consensus, 1 Oxford J. Legal Stud. 1, 3, 3 n.3 (1987).
28. Rawls, Some Reasons for the Maximin Criterion, supra note 27, at 142.



1990] COMMERCIAL IMPRACTICABILITY 73

fair terms for joint-partnerships and for associations, or for small groups

and teams, are not suitable for social cooperation/ '^9

Suppose, however, that Rawls is wrong here. That is, suppose that

Rawls's choice situation applies to local transactions, including com-

mercial contracts. The seller and the buyer are in a state of partial

ignorance, as discussed above. The minimax rule still would not be

adopted as a decision rule, however, because not all the features for

adopting the rule are satisfied. Adoption of the minimax rule requires

that: (1) the choice situation is such that probability assignments are

absent or not well defined; (2) the decision maker is not concerned with

avoiding losses above those assigned by the minimax payoff index; and

(3) other strategy choices have payoffs that the decision maker finds

unacceptable.^^ The problem that arises is that in commercial settings

features (2) and (3) are not present. The absence of each feature is

discussed below.

Feature (2): Assume that seller and buyer are firms whose capital

structure includes publicly traded common stock and debt. Now the
common stock of leveraged firms can be treated as a call option on
the firm's assets. ^^ The value of an option is the difference between the

stock price and its exercise price. An option will not be exercised if its
exercise price exceeds the stock price because its value will never be

below zero. This means that the value of an option increases as the

stock becomes more volatile. Given the same mean return, an option
on stock with a greater variance is more valuable than a stock with less

variance.^^ Thus, the option with the greater variance will be preferred.

29. Basic Liberties, supra note 27, at 15 (emphasis added).

30. Cf. J. Rawls, A Theory of Justice 154-56 (1971). The statement of the
features in the text is cast in terms of the minimax rule because losses are involved. Rawls

is unclear whether features (l)-(3) are jointly sufficient or only individually necessary. The

argument below does not depend on either treatment because it rejects features (2) and

(3).

Note that feature (1) also is not satisfied because the utihty payoffs in Figure 1

already reflect attitudes toward risk. In particular, the payoffs are constructed by reference

to probability assignments to outcomes: objective probabilities if Von Neuman-Morgenstern
functions are used, and subjective probabihties if Ramsey functions are used. This is one

reason why Rawls makes use of a notion of primary goods, not utilities, in evaluating

social states. J. Rawls, supra, at 155. I do not discuss the point because it is well known

in the literature. See, e.g.. Arrow, Some Ordinalist Utilitarian Notes on Rawls' Theory

of Justice, 70 J. Phil. 251, 256-57 (1973).

31. See R. Brealey & S. Myers, Principles of Corporate Finance 430-31, 442-
43 (2d ed. 1984); V. Brudney & M. Chirelstein, Corporate Finance: Cases and
Materials 290-92 (3d ed. 1987) for a summary of the points made in this and the next

two sentences. For experimental evidence on risk preferences in choice situations approx-

imating a Rawlsian original position, see infra note 35.

32. This is true even if the probability density functions do not describe normal

distributions.



74 INDIANA LAW REVIEW [Vol. 24:65

Because common stock in leveraged firms can be treated as options,
common stockholders will prefer mean-equivalent, more volatile projects.
They will be concerned with gains above those guaranteed by the maximin
payoff index. Equivalently, they will not be concerned with avoiding

losses above those that the minimax payoff index guarantees. Therefore,

feature (2) is not present in leveraged firms.

To be sure, leveraged firms themselves will adopt restrictive covenants
to avoid common stockholders from exploiting debt holders by acting
on their preference. This is irrelevant. For the point is not that common
stockholders will not be prevented from accepting mean-equivalent, more
volatile projects. It is simply that they have a preference for doing so.

And this is sufficient to show that feature (2) is not present. Put simply,
if the decision maker is a firm and its managers* choices reflect stock-

holder preferences, the firm will be risk-neutral or risk-preferring.

Feature (3): Assume, as was assumed above, that Rawls*s choice

situation applies both to the basic structure and local transactions.

Application of the minimax rule in the original position is taken to yield

a set of payoffs acceptable to the parties. By assumption, those are the
payoffs guaranteed by adopting the minimax rule, as appHed to the

basic structure. The parties in a state of partial ignorance know this
much. After all, they adopted the minimax rule. Because the parties

know that they are guaranteed acceptable payoffs, they have no reason
to find that other strategy choices in local transactions have unacceptable

payoffs.

In fact, just the opposite occurs. The parties know that other strategy
choices cannot have unacceptable payoffs in local transactions because

the minimax rule has already been adopted and appHed to the basic

structure. This guarantees acceptable payoffs. Thus, the parties have no

reason to adopt the rule for local transactions. To apply it to local
transactions is to apply the minimax rule twice. Double appHcation is

not, by itself, implausible. However, it is implausible when one of the

features necessary for adopting the minimax rule for local transactions

in the first place is removed by its application to the basic structure.

Application of the minimax rule to the basic structure removes feature

(3) as a reason for adopting the rule for local transactions. Hence, the

minimax rule will not be adopted in commercial contexts. Nor, therefore,

will a Rawlsian principle of loss distribution that requires equal sharing

be adopted.

Assume, finally, that the minimax rule is not applied to the basic

structure. The seller and the buyer in commercial contexts still would

not adopt the minimax rule. This is because features (2) and (3) are

not present in commercial contexts, such contexts being local transactions.

Feature (2)*s absence in commercial contexts is independent of the min-

imax rule being applied to the basic structure. The argument concerning



1990] COMMERCIAL IMPRACTICABILITY 75

feature (2) above establishes this independence. That argument only

assumes that the class of sellers and buyers includes leveraged firms in

which not all common stockholders are also debt holders in the same
enterprise. The assumption is a weak and reasonable one and entails
nothing about the minimax rule's appHcation to the basic structure.

Hence, feature (2) is not present in local transactions even if the minimax

rule is not in force at the level of the basic structure.

Neither is feature (3) present in local transactions. Recall that feature

(3) involves strategy choices other than the minimax rule which have

payoffs that the decision maker finds unacceptable. The decision makers

are seller and buyer in the analogue of the original position. The seller

and the buyer include firms consisting of common stockholders. Note
that strategy choices other than the minimax rule have payoffs which

these stockholders find acceptable. In particular, common stockholders
would find a rule assigning the full loss resulting from unforeseeable

supervening events to one party (e.g., seller) acceptable. Acceptability,

at least where common stockholders are concerned, is not a direct function
of the losses consequent of a single strategy choice. Presumably it depends

on two factors: the proportion of those losses to the stockholder's net

wealth and the stockholder's other investments. Alter the net wealth of

a seller or buyer, and the loss distribution rules in Figure 1 may be
ranked differently. Familiar data about choice behavior in decisions

involving **smaH" or **large" losses support this observation.^^ Alter the

portfolio of investments of a seller or buyer, and the ranking of rules

in Figure 1 again may change.
The application of both factors to the common stock shareholder

should be apparent. If the stockholder's investment in the firm of the

seller or buyer is **smaU" relative to her net wealth, losses in excess of

those prescribed by the minimax rule may be acceptable. Whether they
are in fact acceptable, of course, depends on the investment and net

wealth of the particular stockholder, and whether the common stock-
holder diversifies to reduce systematic and unsystematic risk.^^ In the

limiting case in which one investment is a perfect hedge for another

investment, a stockholder will be indifferent between all loss distribution

rules. For instance, in Figure 1, if the stockholder has equally valued

shares in seller and buyer, then her net loss will be the same in a,-a3:

50. Thus, loss distribution rules ai-aj will have equally acceptable payoffs.

Obviously, whether either factor identified obtains equally acceptable

33. See K. Arrow, Essays in the Theory of Risk Bearing 93 (1971).
34. See R. Brealey 8l S. Myers, supra note 31, at 123-26; H. Raiffa, Decision

Analysis: Introductory Lectures on Choices Under Uncertainty 97-100 (1968); W.
Sharpe, Portfolio Theory and Capital Markets 146-50 (1970).



76 INDIANA LAW REVIEW [Vol. 24:65

payoffs is an empirical matter. But the point here is modest: feature

(3) may not be present at the level of local transactions even if the
minimax rule is inapplicable at the level of the basic structure. ^^ Hence,
again, assessments of the respective scope of the minimax rule are
independent of each other.

III. Standard Doctrine of Commercial Impracticability

It is worth emphasizing that a specific normative principle of loss

distribution is required. The principle must be specific in two ways.
First, a well-defined prescriptive basis for assigning losses as between

seller and buyer is needed. Second, the principle must be asymmetric

in its prescriptive force. If a loss is justifiably assigned to the seller

(buyer), then it is not also justifiably assigned to the buyer (seller) on

the same basis. Candidate principles that are cast only in broad terms

are unhelpful. They exhibit one of two defects. The defects correspond

to the ways in which a principle of loss distribution must be specific.

A candidate principle itself can be unjustified. This occurs when a
principle contains no normatively compelling basis for assigning loss.

Alternatively, the candidate principle can justify an assignment of loss

to one party when it equally justifies the loss assignment to the other
party.

*

'Assign loss to the wealthier party as between seller and buyer,'*

without more, is an unjustified principle of loss distribution. Wealth

alone is not a compelling basis for the assignment. **Assign loss to the

party capable of avoiding or reducing the risk of non-performance,"

without more, also justifies assigning the loss to seller and buyer.

Commentators' candidate principles of loss distribution often are

cast in unhelpfully broad terms. One suggestion is that sophisticated
contractors undertaking an unqualified duty of performance should bear

the entire loss because it is '*most fitting. "^^ Another suggestion is that.

35. Some experimental evidence suggests that subjects selecting distributive principles
in choice situations approximating a Rawlsian original position reject the minimax rule.

Frohlich, Oppenheimer & Eavey, Laboratory Results on Rawls's Distributive Justice, 17
Brit. J. Pol. Sci, 1, 11 (1987) [hereinafter Laboratory Results]; Frolich, Oppenheimer &
Eavey, Choices of Principles of Distributive Justice in Experimental Groups, 31 Am. J.

Pol. Sci. 606, 617 (1987) [hereinafter Distributive Justice]. In most cases, subjects instead

select a distributive principle which requires maximizing average income subject to a

constraint on minimum income. Laboratory Results, supra, at 11; Distributive Justice,
supra, at 620. A distributive principle requiring maximization of average income frequently
is selected when higher variance payoffs are present and sometimes selected when losses

are to be distributed. Distributive Justice, supra, at 620-21.

36. Trakman, Frustrated Contracts and Legal Fictions, 46 Mod. L. Rev. 39, 51

(1983); cf Halpern, Application of the Doctrine of Commercial Impracticability: Searching

for "The Wisdom of Solomon," 135 U. Pa. L. Rev. 1123, 1130 (1987) (urging the use
of "flexible and workable tools" in applying the impracticability doctrine). Contra Trakman,

supra, at 53.



1990] COMMERCIAL IMPRACTICABILITY 77

absent the contractors' expectations, **basic principles of fairness and

justice should be applied* *^^ to assign loss. Or it is urged that, given
accurate estimates of probability, negotiation, and litigation costs, the

performing ** party on whom the loss initially falls cannot complain. "^^

The suggestion relies on the fact that the parties will bear the cost of

identifying and shifting the risk of loss up to the point where the

marginal cost of doing so equals the marginal benefit risk reduction

yields. Each prescription appeals just to general normative notions. And
each is thereby defective.

**Most fitting'' may not be a normative basis at all. Even if it is
one, the basis is not well-defined. The seller's cost of performance

increases as a result of increases in input costs. No provision for increases
in the contract price may have existed in the initial agreement. What
is normatively **most fitting" about assigning the consequent loss to

seller? Basic principles of justice and fairness provide no reason for any

particular allocation of losses at all. No prescriptively significant feature
of the transaction or the parties is identified. The notions appealed to,
standing alone, simply are vague.

Further, the requirement of asymmetrical justificatory force is vi-

olated. It may be equally **fair" or **just" to assign consequent losses
to either seller or buyer. The appeal to a notion of not being able to

complain, by itself, is unhelpful. Supplementation with an assumption

is needed. The assumption here^^ is that a loss remains with the performing

37. Farnsworth, Disputes Over Omission in Contracts, 68 Colum. L. Rev. 860,

881 (1968).

38. Gillette, supra note 20, at 537.

39. See id. at 538 (the failure to allocate a risk "may constitute a decision by the
party who will suffer from the risk's materialization that the expected loss from the risk
is not worth the resources that would have to be invested to identify it and allocate it

expressly"); cf. Priest, Breach and Remedy for the Tender of Nonconforming Goods
Under the Uniform Commercial Code: An Economic Approach, 91 Harv. L. Rev. 960,
963 (1978) ("To the extent that the parties neglect specification in the contract, they leave

to courts the task of deducing the attributes upon which they 'agreed.'").

The suggestion is that all risks can be allocated by contractual provision. This

supposition may not always be true. Some low probabihty events may be unforeseeable,
in which case the risk of their occurrence will not be allocated. See Bishop, The Contract-

Tort Boundary and the Economics of Insurance, 12 J. Legal Stud. 241, 249 (1983). The

risk of other events, although foreseeable, cannot be allocated due to the conceptual

difficulty of constructing an appropriate contractual provision. See, e.g., Harris Corp. v.

National Radio and Television, 691 F.2d 1344, 1358 (11th Cir. 1982) (argument that

customer under a standby letter of credit could have protected itself against risk of a

fraudulent demand by the beneficiary "is to ignore the realities of the drafting of commercial

documents"). Market price formulas for adjusting contract price when the good is a

nonhomogeneous commodity are instances of inadequate provisions. See, e.g., Joskow,

supra note 17, at 54-55 (market price formulas in long-term coal contracts fail to define



78 INDIANA LAW REVIEW [Vol. 24:65

party if the risk of that loss has not otherwise been allocated. However,

the assumption is unsound.

Absent a supplementary principle, there is no reason to require the

performing party to bear the risk of loss in the first place. There is no

reason why an otherwise unallocated risk of loss should not be allocated
by the parties* ex post negotiations.'*^ To impose the loss on the per-
forming party unless its risk is shifted is to presuppose that that party

is to bear the risk initially. The same risk, after all, could be assigned
to the nonperforming party initially. In the case of the nonperforming

party, given the above assumption, the nonperforming party also would

have no complaint. Hence, unsupplemented, this candidate principle

violates the requirement of asymmetrical justificatory force. Given dif-

ferent initial assignments with regard to the bearer of the risk of loss,

both the seller and the buyer could not (could) complain.

General Uniform Commercial Code provisions and comments, the
comments to section 2-615, and case law support the adoption of a
normative principle based on foreseeability. The first two sources support
its adoption indirectly. Section 1-103 provides that, "unless displaced

by the particular provisions of*' the Uniform Commercial Code, principles

of equity ** supplement its provisions.*''** Because section 2-615 does not

displace them,'*^ equitable principles apply. Section 1-102(3) provides that

the provisions of the Uniform Commercial Code may be varied by
agreement.'*^ Therefore, absent agreement to the contrary, equitable prin-

ciples apply under section 1-103.

Comment 6 to section 2-615 states that **[i]n situations in which
neither sense nor justice is served . . . when the issue is posed in flat
terms of ^excuse* or *no excuse', adjustment under the" Act's provisions

is necessary, including recourse to equitable principles.'*^ This appears

an appropriate market price norm due to geographic variations, quality, and transportation

costs; a completely contingent contract will be difficult or impossible to draft in such

cases). The inability to allocate the risk of some events by contract has been taken to

define relational contracts. See Goetz & Scott, Principles of Relational Contracts, 67 Va.
L. Rev. 1089, 1091 (1981). The criticism in the paragraph below does not depend on

denying Gillette's deniable supposition.

40. C/. Hillman, Court Adjustment of Long-Term Contracts: An Analysis Under
Modern Contract Law, 1987 Duke L.J. 1, 12-13 ("In a fixed-price contract, . . . there
is no persuasive reason why even foreseeable risks must fall automatically on a party
rather than trigger an adjustment duty.").

41. See U.C.C. § 1-103 (1987).

42. See U.C.C. § 2-615 comment 6 (recourse to general equitable principles ad-
vancing commercial standards and good faith to be used when issue cannot be posed as

one of "excuse" or "no excuse").

43. See U.C.C. § 1-102(3); cf U.C.C. §§ l-102(l)-(2); 1-102 comment 3.
44. U.C.C. § 2-615 comment 6.



1990] COMMERCIAL IMPRACTICABILITY 79

to allow appeal to equitable principles that have just outcomes in cases

involving section 2-615/^ Direct support comes from other comments to

section 2-615. Comment 1 specifies that section 2-615 applies when
performance becomes commercially impracticable through unforeseeable

supervening events.'** Cgmment 4 specifies that performance is excused
when the increased cost is due to some unforeseeable contingency that
alters the "essential nature" of performance."*^ Putting the indirect and

direct support together: in applying section 2-615, courts can appeal to

principles, including equitable principles, to achieve just outcomes on

the basis of foreseeability.

Case law further supports this unexciting conclusion. Courts, in

applying section 2-615, typically focus on the foreseeability of a super-

vening event. "** As the court in In re Westinghouse Electric Corp. Uranium
Contracts Litigation succinctly put it: **Where the promisor had no
reason to anticipate a supervening event which radically increases the

difficulty of performance, or which renders performance impossible, it

is manifestly unfair to hold him to the agreement. '*"*' Foreseeability of

an event is taken to be logically distinct from risk allocation with respect

to that event.^° Hence, the distinctness of the statutorily specified con-

ditions in section 2-615 is recognized. In practice, however, courts take

the foreseeability to be contingently connected to risk allocations.^' A

45. C/., e.g.. Opera Co. of Boston, Inc. v. Wolf Trap Found, for the Performing

Arts, 817 F.2d 1094, 1100 (4th Cir. 1987) ("the modern and prevailing doctrine of

impossibility of performance as a defense to a breach of contract" is essentially equitable

in character).

46. U.C.C. § 2-615 comment 1.

47. U.C.C. § 2-615 comment 4.

48. See, e.g., Alimenta (U.S.A.), Inc. v. Gibbs Nathaniel (Canada) Ltd., 802 F.2d

1362 (11th Cir. 1986); Cook v. Deltona Corp., 753 F.2d 1552, 1558 (11th Cir. 1985);
Waldinger Corp. v. CRS Group Engineers, Inc., 775 F.2d 781, 786 (7th Cir. 1985); Roy
V. Stephen Pontiac-Cadillac, Inc., 15 Conn. App. 101, 105, 543 A.2d 775, 778 (1988);

Nora Springs Coop Co. v. Brandau, 247 N.W.2d 744, 747-48 (Iowa 1976); Mishara Constr.
Co., Inc. V. Transit-Mixed Concrete Corp., 365 Mass. 122, 128, 310 N.E.2d 363, 367

(1974).

Courts also focus on foreseeability in determining contractual duties under agreements

not governed by Article 2 of the Uniform Commercial Code. Contractual modification

provides a well known instance. See Restatement (Second) of Contracts § 89 (1981)
("A promise modifying a duty under a contract not fully performed on either side is
binding. . .if the modification is fair and equitable in view of circumstances not anticipated

by the parties when the contract was made. . . ."). See also Recker v. Gustafson, 279

N.W.2d 744 (Iowa 1979); Angel v. Murray, 113 R.I. 482, 322 A.2d 630 (1974).
49. 517 F. Supp. 440, 454 (E.D. Va. 1981), partially rev'd, 826 F.2d 239 (4th Cir.

1987).

50. Transatlantic Fin. Corp. v. United States, 363 F.2d 312, 319 (D.C. Cir. 1966).

51. See, e.g.. Northern Ind. Pub. Serv. Co. v. Carbon County Coal Co., 799 F.2d



80 INDIANA LAW REVIEW [Vol. 24:65

party who could foresee the occurrence of a supervening event is either
found to have assumed the risk of its occurrence or held to have assumed
that risk." That is, foreseeability is taken either to be evidence of risk

allocation or to be a reason for imposing the pertinent risk on a party.
In either case, foreseeability is being used to allocate the cost of price

increases between the contracting parties. Hence, standard applications

take foreseeability to be normatively primary.

As a matter of applying section 2-615, this treatment is of course
unobjectionable. Because unforeseeability is a necessary condition for

excusing a seller's performance, a court need only determine foreseeability

to order enforcement of the contract. Determining the parties' risk

allocation is superfluous. The omission presumably is justified by a
sometimes-stated premiss: A party who reasonably anticipates a contin-
gency bears the responsibility of using its knowledge to allocate risk by

contractual provision." Foreseeability, therefore, is a normatively sig-

nificant feature of the contracting parties. However, by itself foresee-

ability is not normatively significant and is an insufficient reason to

impose liability for a price increase. Contracting parties are not liable,

contractually or otherwise, for all the foreseeable risks of their conduct.

Rather, performance is required only with respect to those foreseeable

risks that were allocated by the parties* bargain. The following Section
offers a normative principle connecting the parties* bargain and fore-

seeability.

IV. Loss Distribution: Foreseeability Normatively Justified

Under which condition is it permissible to impose liability on a party

for the occurrence of a harm (a cost or a loss)? This is a general
normative question, not peculiar to issues of commercial impracticability.

Applied to commercial impracticability, the question is equivalent to

another: Under what conditions should a contracting party bear the loss

265, 278 (7th Cir. 1986); Asphalt Int'l, Inc. v. Enterprise Shipping Corp., 667 F.2d 261,

265 (2d Cir. 1981); In re Westinghouse, 517 F. Supp. at 454; Iowa Elec. Light & Power
Co. V Atlas Corp., 467 F. Supp. 129, 133 (N.D. Iowa 1978); Nora Springs Coop Co.,
247 N.W.2d at 747.

52. See, e.g.. Cook v. Deltona Corp., 753 F.2d 1552 (11th Cir. 1985); United

States V. Wegematic Corp., 360 F.2d 674 (2d Cir. 1966); Berg v. Erickson, 234 F. 817

(8th Cir. 1916).

53. See Nora Springs Coop Co., lAl N.W.2d at 747; cf. Lloyd v. Murphy, 25

Cal. 2d 48, 54, 153 P.2d 47, 50 (1944) ("If it was foreseeable there should have been

provision for it in the contract, and the absence of such a provision gives rise to the

inference that the risk was assumed."); Farnsworth, supra note 37, at 876 ("courts properly

try to realize the actual expectations of the parties even when they have not reduced them

to contract language.").



1990] COMMERCIAL IMPRACTICABILITY 81

for the increased cost of an input or the diminution in the value of

performance? More simply: Under what conditions should a party bear
the loss for an event rendering one party's performance uneconomical?

To say that contractual liability is permissibly imposed when deserved
is unhelpful. It does not specify the conditions under which a party

deserves to bear a loss (a cost or a harm). It does not follow that a

deserved loss, however specified, ought to be imposed. Supplementary

premisses are required.

One suggestion appeals to the notion of consent. A party deserves
to bear those losses to which she has consented. Absent express consent,

the notion of consent can be taken to be equivalent to ex ante com-

pensation. This is Posner*s suggestion. ^"^ Because consent clearly entails

foreseeability, the normative requirement of foreseeability in Section III

is satisfied. An argument for imposing loss goes as follows:
(1) Parties should only bear the losses they deserve.

(2) Deserved losses are losses to which a party consents.

(3) Parties consent to losses, the risk of which they are compensated

to bear ex ante.

(4) Therefore, parties should bear the loss, the risk of which they

are compensated to bear ex ante.

Premiss (3) contains Posner's crucial suggestion. ^^ The premiss also has

been subject to severe criticism. Commentators^^ have pointed out the

failure: It does not follow from the fact that a party is compensated,

ex ante or ex post, that that party consented to the resulting loss.

Thomson has noted the failure of truth-functionally valid inferences in
modal contexts involving consent, as in the foliowing:^^

If [C(if p then q) and p], then Cp.

54. Posner, The Ethical and Political Basis of the Efficiency Norm in Common
Law Adjudication, 8 Hofstra L. Rev. 487, 492 (1980); see also Schwartz, supra note
13, at 8.

55. Posner, supra note 54, at 492.

56. E.g., Coleman, Efficiency, Utility, and Wealth Maximization, 8 Hofstra L.

Rev. 509, 535-36, 536 n.45 (1980); Dworkin, Why Efficiency?, 8 Hofstra L. Rev. 563,
575-76 (1980).

57. Cf. J. Thomson, Rights, Restitution and Risk 189 (1986). The above is an
instance of the failure of modus ponens in modal contexts. For other well-known inferential
failures in (alethic) modal contexts, see L. Linsky, Oblique Contexts 98-99 (1983); W.
QuiNE, From a Logical Point of View 141-42 (2d ed. 1980); W. Quine, Word and
Object 151 (1960). For an implicit recognition of the above failure in tort law, when

consent is at issue, see Simons, Assumption of Risk and Consent in the Law of Torts:
A Theory of Full Preference, 67 B.U.L. Rev. 213, 217 (1987) ("Participants in a game
know that injuries often occur. And they choose to participate in the game with that
knowledge. Do they legally consent to the risks of all injuries that they can foresee? 1
seriously doubt that they do.").



82 INDIANA LAW REVIEW [Vol. 24:65

Let **C" be the one-place, non-alethic modal operator ** Jones consents

to . . . ." Suppose Jones consents to losing a dollar if a card marked
"7" is drawn from the deck. A card marked **7" is drawn. Jones may
or may not consent to losing a dollar in the circumstances. His consent
cannot be inferred just from the truth of the previous two sentences.

Hence, sentence variables within modal operators cannot be detached

from those operators, as the inference in (3) above requires. This con-

clusion has a normative corollary. The corollary is that the justificatory

basis of (4) is not consent. Contrary to premiss (3), parties need not

consent to a loss, the risk of which they are compensated to bear ex

ante. Modality aside, Posner's implicit inference is still invalid: Invalidity

obviously results even if **consent propositions" are interpreted truth-

functionally. For from the fact that you have consented to the risk that

events p or q or r will occur, it does not follow that if q occurs, you

have consented to its occurrence. Detachment of a disjunct is invalid

here as well.

Nevertheless, the unsoundness of premiss (3), Posner*s suggestion,

is not fatal to the above argument. Premiss (3) can be replaced with

the following:

(3)* Parties cannot justifiably object to losses, the risk of which

they are compensated to bear ex ante,

the remainder of the above argument being altered accordingly. The

justification of premiss (3)' is as follows. Using Posner's example,^*

compare the outcome of a lottery with being the victim of a crime.

Suppose one is compensated ex ante in both cases. Absent fraud or

duress, the lottery **victim" has no grounds for complaint. On what
grounds can the victim of a crime complain? The victim of the crime

may not complain on the basis of there being a person-assailant in the
crime case and not in the lottery case. A person could function as a
^'randomizer" in a lottery (e.g.y whichever number Jones selects wins).

Thus, the presence of a person inducing the loss or risk of loss is

irrelevant.

Further, the complaint cannot be based on the sort of loss a party

was compensated to bear. Suppose I am given $100 to bear a .10 risk
of the loss of a watch valued at $1000. Whether the loss is classified

as monetary or as a loss of a valuable and exchangeable good seems

irrelevant. Also irrelevant is the claim, again, that the party had an

unequal prospect of being a crime victim vis-a-vis other parties. The

victim's prospects in fact could be equal to those of other potential

victims. Finally, it is irrelevant that the party failed to bargain with the

58. Posner, supra note 54, at 492.



1990] COMMERCIAL IMPRACTICABILITY 83

crime victim. If transaction costs are high, neither party would prefer

to bargain. The objection therefore must be that the party did not
consent to the transfer. And that does not show that the transfer itself
was objectionable, as noted two paragraphs up.

Instead, the crime victim's basis of complaint concerns the acceptable

alternatives to bearing a risk, even if compensated. In particular, the

victim has no opportunity to refuse to bear the risk. Alternatively, if

refusal is possible, the consequence is deemed unacceptable. It may be
deemed unacceptable for a number of reasons. The outcome itself is
deemed unacceptable. (Contrast this with the alternative outcomes of

receiving $100 or $1000, no other consequences attached.) Or the outcome
would be unacceptable were collective action problems not present. Al-

ternatively, it may be acceptable to the choosing party, but weakens the
position of others selecting an option yielding the same outcome. The

victim also does not have the opportunity to bargain for a different,

and presumably higher, level of compensation. Therefore, a fuller but

still rough formulation of (3)' would be:

(3)'' If the party expects (foresees) an event to occur with prob-

ability p; if she is compensated for that event's occurrence (to

her); and if she has the opportunity to refuse the compensation

or bargain for a different level of compensation, either being

acceptable alternatives; then if the event occurs, she cannot

justifiably object.

Precisely enumerating the conditions in (3)" is difficult. Difficulties in

stating what constitutes an acceptable alternative, either in refusing to

bargain or in altering the compensation level, are apparent. Buyers may
in fact have no opportunity to bargain for a different compensation

level at all. Compensation levels can be set by standard form contracts.

And a buyer may have relatively unique demands for a particular level.''
If both possibilities occur, the marginal cost to a seller of offering a

nonstandard compensation level will exceed the marginal benefit to the

buyer. Hence, sellers will not offer such buyers the opportunity to bargain

for a nonstandard level of compensation. They will respond only to

aggregate demand, not to relatively unique buyer demand.^ This is not

a sufficient condition for the unacceptability of an alternative.

59. See Schwartz, A Reexamination of Nonsubstantive Unconscionabilityy 63 Va.
L. Rev. 1053, 1064-66 (1977).

60. Id. at 1064-66; Schwartz, Seller Unequal Bargaining Power and the Judicial

Process, 49 Ind. L.J. 367, 377-78, 386 (1974); see also Trebilcock, The Doctrine of

Inequality of Bargaining Power: Post-Benthamite Economics in the House of Lords, 26

U. Toronto L.J. 359 (1976).



84 INDIANA LAW REVIEW [Vol. 24:65

Specifying all the conditions in (3)" is unnecessary here. For in

commercial settings, the parties typically have the acceptable alternatives

of not entering into the contract or of bargaining for a different com-

pensation level. Indeed, cases concerning issues of commercial imprac-

ticability often involve nonstandard contracts that are the product of

elaborate negotiations.^' In commercial settings involving merchants, pro-

visions concerning the unconscionability of contract clauses also apply

via section 2-302.^^ Because a compensation level is specified by a contract

clause, section 2-302 is applicable independently of section 2-615. Thus,

the condition in (3)*' concerning acceptable alternatives is superfluous

in commercial settings. Only the above conditions concerning expectations

and compensation are left. Hence, in commercial settings only compen-

sated-for expectations (foresight) of risk are important.

A party cannot justifiably object to bearing the loss of foreseeable
events, the risk of which she was compensated ex ante. To require
performance given unforeseeable supervening events is objectionable be-

cause performance under the circumstances is an unbargained for ad-

vantage.^^ It is, therefore, undeserved according to the above argument.

Call that part of premiss (3)** that concerns compensation and foresight

the compensation-expectation principle.

A simple model^ of production under risk supports use of the
compensation-expectation principle here. Let C be the total cost to the
seller of providing a unit of a particular good to buyer. Let q be the

61. See ALCOA v. Essex Group, Inc., 499 F. Supp. 53 (W.D. Pa. 1980) and
Eastern Air Lines, Inc. v. Gulf Oil Corp., 415 F. Supp. 429 (S.D. Fla. 1975) for two

notable instances.

62. U.C.C. § 2-302(1) states: "If the court as a matter of law finds the contract

or any clause of the contract to have been unconscionable at the time it was made the
court may refuse to enforce the contract, or it may enforce the remainder of the contract
without the unconscionable clause, or it may so limit the application of any unconscionable
clause as to avoid any unconscionable result." U.C.C. § 2-302(1) (1987).

63. See, e.g., Waldinger Corp. v. CRS Group Engineers, Inc., 775 F.2d 781, 786
(7th Cir. 1985); Roy v. Stephen Pontiac-Cadillac, Inc., 15 Conn. App. 101, 105, 543

A.2d 775, 778 (1988); Mishara Constr. Co., Inc. v. Transit-Mixed Concrete Corp., 365

Mass. 122, 128, 310 N.E.2d 363, 367 (1974).

64. The model is a variant on the efficient insurer model offered by Goetz and

Scott in another context. See Goetz & 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, 579-83 (1977). It differs from Goetz and Scotts' model

principally in that the probability of input price increases is treated as being stochastically

independent of the seller and the buyers' actions. The treatment therefore assumes that

there is no question of moral hazard. For a treatment of warranties in which consumers

and manufacturers allocate investment in precautions and insurance according to com-

parative cost, see generally Priest, A Theory of the Consumer Product Warranty, 90 Yale
L.J. 1297 (1981).



1990] COMMERCIAL IMPRACTICABILITY 85

cost to the seller of purchasing input i at the time of production. A
unit of every required input 1, 2. . ., n is needed to produce a unit of

the produced good. Seller's total production cost, unadjusted for the

risk of input price changes, is summarized as C = c, . . . + Cj . . . +
Cn. There is a prospect that input prices will increase after the seller

and the buyer have contracted. The seller must be compensated to bear
the risk of that prospect. The risk is also a production cost to the seller.
Two elements enter into the seller's risk-bearing cost. One is the expected
value of the seller's loss on the contract should input prices increase.

Define seller's expected value of loss as the product of p(c + 1), c, 1
> 0, p being the probability that an input will cost a given amount at
the time of production, c a particular input price, and 1 a random
variable with a positive mean. A second element is the resources expended
in determining the values of p, c, and 1, respectively. Let this expenditure

be described by r. Then seller's expected risk-bearing cost^^ for an input
i is E(Ci) = (1 - p)Ci + [p(c + l)i 4- r]. Assuming that risk-bearing
for input prices is a competitively priced activity with a normal rate of

profit, TT, seller's risk-bearing compensation for input i is E(Ci)(l + it).
E(Ci)(l + tt) is the amount which the seller will demand given the
probability of an increase in the price of input i. And E(C) is the
amount which the seller will demand to produce the good given the
probability of price increases across all inputs.

Take now just input i. Buyer will pay seller E(Cii)(l + tt) if E(Ci)
for seller is less than E(Ci) for buyer (i.e., Egfe) < Ebfe)).^ Because
E(Ci) = (1 - p)Ci + [p(c + l)i + r)], buyer will do so if (1 - p)Ci +
[p(c + l)i + r)] for seller is less than that for buyer. That is, E^Cj) <
Eb(Ci) obtains just when the seller's expected risk-bearing cost is less
than the buyer's expected risk-bearing cost. The values of p, c and 1
are taken to be stochastically independent of the seller and the buyers'

behavior. ^^ As price-takers in a competitive input market, both cannot
affect the values of p, c, and 1. Thus, those values remain constant as

between the seller and the buyer. The value of r, the resources expended

to determine p, c, and 1, may vary between the two parties. Hence
Es(Ci)(l + tt) < Eb(Ci)(l + tt) when r takes values such that r^ < r^.
When E^CiXl + tt) < Eb(Ci)(l + tt), the buyer will pay the seller E(Ci)(l
+ tt) to bear the risk described by p. In such a case, the seller's
expectations concerning input i's price increases are already compensated.

65. Both the seller and the buyer are assumed to be risk-neutral, so U(c) = E(c)
= c.

66. Cf. Priest, supra note 64, at 1313 ("A manufacturer . . . offers market insurance
for those losses or items of service for which market insurance is less costly than insurance

or allocative investments by the consumer himself.").

67. See generally Priest, supra note 64.



86 INDIANA LAW REVIEW [Vol. 24:65

The same also is true for all other inputs 1, 2,. . ., n. Hence, in the
limiting case, the seller's total production cost, adjusted for risk across

all inputs, is summarized by E(C) = E(c,)(l + tt). . . + E(Ci)(l + tt)
. . . + E(Cn)(l + tt). As represented by the E(c)(l + tt) terms, the
seller is compensated for any expectations which concern input prices.

Given the normative principle contained in condition (3)'*, courts

are to allocate losses due to supervening events on the basis of desert.^®

Courts are not to allocate loss in a manner such that social cost or the

contracting parties* joint costs are minimized. Cost minimization is only

contingently related to the compensation-expectation principle and, there-

fore, contingently related to desert. Even when the social costs and the
parties' joint costs coincide, the compensation-expectation principle jus-

tifies the allocation of loss. Cost minimization, however, does not provide

the justification. It is, at best, evidence that a party has been compensated

ex ante to bear a risk of loss. Courts instead are to identify a feature

of the seller and the buyer by which loss justifiably can be imposed.

That feature is captured by the compensation-expectation principle :^^ A
party cannot justifiably object to an event's occurrence if that occurrence

is expected with a probability p, and the party is compensated for the

event's occurrence ex ante. Applied to the distribution of loss, the

principle reads: A party cannot justifiably object to the imposition of
a loss (cost) if that loss is expected with a probability p, and if the

party is compensated for the risk of the loss ex ante. When the com-
pensation-expectation principle is satisfied, the party deserves to bear

the compensated-for loss. Conversely, imposition of any greater or lesser

loss is undeserved. Courts using the compensation-expectation principle

are to allocate loss according to desert. ^° To do so requires a determination
of what was foreseeable to the parties at the time the contract was

68. Schwartz calls this directive, based on the consent principle, the "desert case."

Schwartz, supra note 13, at 11. The directive in the text, based on the compensation-

expectation principle, appeals to a different justification.

69. As construed, the compensation-expectation principle is the basis for claiming

that a party does or does not deserve to bear a particular loss. C/., e.g.y J. Feinberg,

Doing and Deserving 58-61 (1970); G. Sher, Desert 7-9 (1987) (all stating the condition
that desert claims be justifled by reference to some characteristic or action of the party

who is the subject of the claims); Kleinig, The Concept of Desert, 8 Am. Phil. Q. 71
(1971). Cost minimization is an irrelevant basis according to the compensation-expectation

principle. The argument in the text therefore avoids the criticism that calling a proposal

incorporating cost minimization a "desert case" is both "unnecessary and misleading."

Gillette, supra note 20, at 578.

70. Strictly, condition (3)", of which the compensation-expectation principle is a

part, is to be used. Because consideration of the other conditions of condition (3)" is

unnecessary in preponderate commercial contexts, the use of the compensation-principle

alone is warranted. See supra notes 61-62 and accompanying text.



1990] COMMERCIAL IMPRACTICABILITY 87

entered into. This is required even if it is assumed, as courts do,^' that

what was foreseeable was impounded in the contract price via a prob-

ability function.

To illustrate: Suppose seller's input cost is $.80 at the time of
contracting. Suppose also that seller and buyer expect post-contract input

prices to vary +-$.10 and to be fully reflected in the market price of

the output. Seller will charge a premium to bear the risk of the price
increase. Assume that the contract price is $1.00, which includes seller's
normal profit and the risk premium. Seller has been compensated ex

ante to bear a risk of a ten percent price increase or decrease. Finally,

suppose that if seller has a foreseeable input price increase of +-$.50,

she would have agreed to a contract price of $1.50. The post-contract

price in fact rises to $1.50. The seller's deserved loss is $.10 (-$.10 =

$1.00 - $1.10), her compensated-for expectation. Ten cents ($.10) is the

amount of risk which she was paid to bear. The seller's undeserved loss

is $.40 (-$.40 = $1.10 - $1.50), the uncompensated-for, unforeseeable
risk borne. Under the compensation-expectation principle above, the

assignment of any part of the $.40 loss to seller is undeserved. It should

not be imposed on seller.

The argument here should not be misunderstood. It is not that buyer

deserves to bear only the $.40 loss and no more. After all, the buyer

was not compensated ex ante to bear any price increase over +-$.10

either. Based upon the compensation-expectation principle, the buyer

could justifiably complain about having the $.40 loss imposed on her.

The asserted argument is that the compensation-expectation principle, in

conjunction with section 2-615, justifies the imposition of the loss because

the compensation-expectation only applies to sellers. This result occurs

because section 2-615 expressly excuses the seller's performance. It does

so, given its judicial gloss,^^ by appeal to unforeseeable supervening

events. Both the statute and the case law treat the seller's foresight as

an important factor. ^^ Further, comment 6 to section 2-615 allows ad-
justments to be made when the issue cannot satisfactorily be framed in
terms of '^excuse" or *'no excuse."^'* It allows distribution of loss between

the seller and the buyer. (Comment 6 applies to the above illustration.)

71. See supra notes 39-43 and accompanying text for the standard analysis. See

also Florida Power & Light Co. v. Westinghouse Elec. Corp., 826 F.2d 239, 270-71 (4th
Cir. 1987) (the strongest evidence that the parties expected spent nuclear fuel to be

reprocessed was the contract price); In re Westinghouse Elec. Corp. Uranium Contracts

Litig., 517 F. Supp. 440, 444 (E.D. Va. 1981) (contract price for per kilowatt hour

reflected the parties' anticipation of the reprocessing of spent nuclear fuel).

72. See supra text accompanying notes 10-11.

73. See, e.g.. Cook v. Deltona Corp., 753 F.2d 1552 (11th Cir. 1985).
74. U.C.C. § 2-615 comment 6 (1987).



88 INDIANA LAW REVIEW [Vol. 24:65

Section 2-615 and its case law provide a basis for limiting the com-

pensation-expectation principle to sellers. ^^ Comment 6 provides a basis
for distributing loss between the buyer and the seller. The former basis

justifies imposing part of the loss on seller. The latter basis, in conjunction
with the compensation-expectation principle, justifies imposition of part

of the loss on buyer.

The operative notion of desert also should be noted. Desert sometimes
is distinguished from entitlement.^^ Entitlements are claims said to require

the existence of a set of legal or institutional rules. ^^ Desert is taken to

bear no necessary connection to such rules. The distinction is not adopted

here. Desert claims may or may not be distinct from claims to entitle-
ments.^^ Because only the notion of desert is doing any justificatory

work above, no commitment to a thesis of conceptual difference need

be made. Further, the operative notion of desert is non-polar^^ with

respect to loss. That is, it does not follow from the fact that the seller

does not deserve to bear a given loss that the buyer deserves to bear

that loss. Perhaps no party deserves to bear it. Desert, in conjunction

with a statutorily and judicially applied directive, justifies the imposition

75. The compensation-expectation principle by itself is neutral as between the seller

and the buyer. Hence, U.C.C. § 2-615 and case law apart, the buyer's performance also

could be excused by its application. Cf. Northern Ind. Pub. Serv. Co. v. Carbon County

Coal Co., 799 F.2d 265, 274-75 (7th Cir. 1986) (buyer-public utility sought excuse from

performance of a fixed-price contract under impracticability doctrine when electricity costs

declined). Suppose, again, that the contract price is $1.00 and that the seller bears a risk

of a +-$.10 variance in price. The market price declines to $.50. A loss of $.50, represented
by the market-contract differential, is to be allocated. Unsupplemented, the compensation-

expectation principle justifies imposing a deserved loss of $.40 on the buyer (-$.40 =

$.50 - $.90) and $.10 on seller (-$.10 = $.90 - $1.00). The buyer's risk-adjusted deserved

loss is $.40, the size of the risk-adjusted market-contract differential. Section 2-615 and

case law preclude the buyer from relying on the compensation-expectation principle. Cf.

id. at 276-77 (section 2-615 and Indiana's official comments to § 2-615 appear to make

§ 2-615 applicable only to sellers); Swift Textiles Inc. v. Lawson, 135 Ga. App. 799, 804-

06, 219 S.E.2d 167, 170-71 (1975) (citing commentary limiting the availability of § 2-615

to sellers). Hence, the buyer would have to bear the entire $.50 loss represented by the

unadjusted market-contract differential. Section 2-615 and case law, not the compensation-

expectation principle itself, concern only the seller's loss. For an analogous limitation at

the constitutional level, see Kaplow, An Economic Analysis of Legal Transitions, 99 Harv.
L. Rev. 509, 554-55 (1986) (the takings clause of the Constitution is limited to a requirement

that losers be compensated, there is no requirement that the gainers compensate losers

or incur a pro rata tax on those gains).

76. See J. Feinberg, supra note 69, at 57; W. Sadurski, Giving Desert Its Due
118-20 (1985); J. Kleinig, supra note 69.

77. See supra note 76.

78. Arguments for the distinction rely on appeal to the ordinary use of words.

See supra note 76. Even advocates of the distinction sometimes elide desert and entitlement

claims. See, e.g., J. Feinberg, supra note 69, at 59, 62.

79. I borrow the term from J. Feinberg, supra note 69, at 62.



1990] COMMERCIAL IMPRACTICABILITY 89

of loss on a party. Seller does not deserve to bear a particular loss.

Buyer may not deserve to bear the loss either. Nonetheless, the loss has
to be distributed as between seller and buyer.*^ Given section 2-615, its

comments and pertinent case law, and the fact that the seller does not
deserve to bear the loss, the loss justifiably is imposed on buyer.

V. Foreseeability: The Description Problem

There is a problem here. It concerns the use of the notion of

foreseeability. The problem is that the foreseeabihty of a risk depends
on how that risk is described. **Foreseeability contexts,'' to use the
jargon, are intensional. Roughly, a sentential context is intensional if

there is some singular term, coextensive predicate, or truth-preserving
sentence such that its replacement in that context alters the truth value

of the sentence.

For example, let C(x) be a sentential context introduced by "C'
Let s be a particular sentence with a corresponding truth-value. Then
C(x) is an intensional context if C(s)'s truth-value varies with truth-

preserving replacements of s or constituents of s. Foreseeability concerns

that which an information set impUes or makes probable about future

events or contingencies. An informational set consists of a set of prop-
ositions or proposition-like items. Whether an event is foreseeable relative

to that set depends on the propositions available given the available

descriptions. It is only relative to those descriptions that an event is

implied or made probable.
Suppose X is the event of being exposed to uranium. Consider the

following sentences:

1. x is harmful to persons.

2. Exposure to uranium is harmful to persons.

3. Unprotected contact with this ore is harmful to persons.

All three sentences are true by virtue of the same event (exposure to

uranium). Each remains true when "exposure to uranium," **x,'' and
**contact with this ore" are substituted for one another in the respective

sentences. However, the truth values of the following sentences may
vary:

It is foreseeable that x is harmful to persons.

It is foreseeable that exposure to uranium is harmful to persons.

It is foreseeable that unprotected contact with this ore is harmful

to persons.

80. For discussion of the justification of the requirement that compensation for

loss occur between parties to a transaction, see Walt, Causation and Victim-Particularized

Compensation (unpublished manuscript on file with author).



90 INDIANA LAW REVIEW [Vol. 24:65

The problem is to select among available descriptions of the same event
in a nonarbitrary way such that the event is foreseeable under a de-
scription. Call this the description problem.

The description problem has an empirical instantiation. It is present
in Tversky and Kahneman's experimental results on the framing effects
of decision problems.*' A decision frame is the conception under which
a decision maker represents acts, outcomes, and conditional probabilities.

Tversky and Kahneman show that framing effects can induce different
decision behavior in otherwise identical choice problems.*^ With respect

to the framing of outcomes, a significant majority of subjects select

option A over B and option C over D in the following problem:
There are 600 deaths expected from a disease.

A: If program A is adopted, 200 will be saved.
B: If program B is adopted, there is a .33 probability that 600

will be saved, and a .66 probability that no one will be

saved.

C: If program C is adopted, 400 people will die.
D: If program D is adopted, there is a .33 probability that no

one will die, and a .66 probability that 600 people will die.

Options A and C and B and D, respectively, present different descriptions
(frames) of statistically equivalent alternatives. Tversky and Kahneman
note that subjects are risk-averse when outcomes are described as gains
and are risk-seekers when outcomes are described as losses.*^ Experi-
mentally induced demand for insurance is also subject to the same framing
effect.*'* The same phenomenon has been found among negotiators in
bargaining situations.*^

81. See, e.g., Tversky, A Critique of Expected Utility Theory: Descriptive and
Normative Considerations, 9 Erkenntnis 163, 166-68 (1975) [hereinafter Critique]; Tversky

& Kahneman, Can Normative and Descriptive Analysis be Reconciled?, Working Paper
RR-4, Center for Philosophy and Public Policy, March 1987 [hereinafter Normative and

Descriptive Analysis]; Tversky & Kahneman, Rational Choice and the Framing ofDecisions,
59 J. Bus. 251 (1986) [hereinafter Rational Choice]; Tversky & Kahneman, The Framing
of Decisions and the Psychology of Choice, 211 Sci. 453 (1981) [hereinafter Framing of
Decisions]. See also Bierman, The Allais Paradox: A Framing Perspective, 34 Beh. Sci.
46 (1989).

82. See Normative and Descriptive Analysis, supra note 81; Rational Choice, supra

note 81; Framing of Decisions, supra note 81.

83. See Critique, supra note 81; Framing of Decisions, supra note 81.

84. Hershey, Kunreuther & Schoemaker, Sources of Bias in Assessment Procedures
for Utility Functions, 28 Momt. Sci. 936 (1982); Hershey & Schoemaker, Risk Taking
and Problem Context in the Domain of Losses: An Expectd Utility Analysis, 47 J. Risk



1990] COMMERCIAL IMPRACTICABILITY 91

Framing affects probability assignments as well. It is present in what

Tversky and Kahneman term a pseudo-certainty effect: The measure of
certainty assigned to an outcome when it in fact is dependent on a
contingency.®^ Pseudo-certainty can be induced by a sequential formu-

lation of decision problems. Here is an example:

Problem I: The following is a two-stage game. In the first stage,
there is a .75 chance to end the game without winning anything,
and a .25 chance to move into the second stage. If you reach the
second stage you have a choice between:

A: A sure win of $30.
B: An .80 chance to win $45.
Which of the above do you prefer?
Problem 2: Which of the following do you prefer?
C: A .25 chance to win $30.
D: A .20 chance to win $45.

The majority of subjects select A over B in Problem 1 and D over C
in Problem 2, even though the probabilities in options A and C are
identical (A: .25 = 1.0 x .25), as are the probabilities in options B and
D (D: .20 = .25 x .80). Tversky and Kahneman note that these choices
violate a principle of invariance or extensionality: Preferences between

options should be independent of their description.®^ They also note that

the principle is a necessary condition for a normatively adequate theory

of choice.®®

The description problem is not peculiar to the compensation-expec-

tation principle outlined in Section IV. It applies to any normative

Ins. Ill (1980); Schoemaker & Kunreuther, An Experimental Study of Insurance Decisions,
46 J. Risk Ins. 603 (1979).

85. Bazerman, Negotiator Judgment, 27 Am. Behav. Sci. 211, 213-14 (1983);
Bazerman & Carroll, Negotiator Cognition, 7 Res. Organizational Behav. 247 (1986).
Other researchers have shown the presence of framing effects among physicians with respect
to choices among surgical procedures. McNeil, Pauker, Sox & Tversky, On the Elicitation
of Preferences for Alternative Therapies, 306 New Eng. J. Med. 1259 (1982). For a
review of some of the empirical literature on decision theory, which warns against uncritical

acceptance of the findings in designing legal rules, see Scott, Error and Rationality in

Individual Decision Making: An Essay on the Relationship Between Cognitive Illusions
and the Management of Choices, 59 S. Cal. L. Rev. 329 (1986).

Experimental evidence suggests that people's choice behavior is largely independent

of the order in which frames are presented and that frame presentation is not a good

predictor of choice behavior. Fischhoff, Predicting Frames, 11 J. Exp. Psy. 103 (1983).

86. Framing of Decisions, supra note 81, at 453.

87. Normative and Descriptive Analysis, supra note 81, at 2, 4; cf. Arrow, Risk

Perception in Psychology and Economics, 20 Econ. Inq. 9 (1982).

88. Normative and Descriptive Analysis, supra note 81.



92 INDIANA LAW REVIEW [Vol. 24:65

principle of loss allocation employing a notion of foreseeability, directly

or indirectly. In particular, it applies to Posner and Rosenfields' in-

junction to discharge performance only when the nonperforming party
is the superior risk bearer.*'* The injunction initially requires asking which
party assesses the probability that a risk will materialize. This is equivalent

to asking which party foresees the occurrence of an event.

In this context, a probability is best understood as a measure of a

party's uncertainty about the truth or falsity of a proposition about an

event. However, a proposition about an event is an expression of that

event under particular descriptions. Hence, an individual's uncertainty

can vary with the propositions under which the same event is described.

Therefore, if the foreseeability of an outcome depends on the available

descriptions, so too does the assessment of the magnitude of risk. Both

what is foreseeable and what is probable varies with those descriptions.

Notice that the description-dependence of foreseeability is not con-

fined to a class of decision makers. In particular, it is not confined to

unsophisticated individuals. Granted, Tversky's and Kahneman's exper-

imental results concern lay persons in artificial contexts. They neither

concern commercially sophisticated individuals or corporations,^ nor

involve long-term commercial contracts.^' However, some experimental

results indicate the presence of framing effects among commercially
sophisticated individuals. ^^

89. Northern Ind. Pub. Serv. Co. v. Carbon County Coal Co., 799 F.2d 265, 278

(7th Cir. 1986) (Posner, J.); Posner & Rosenfield, Impossibility and Related Doctrines in
Contract Law: An Economic Analysis^ 6 J. Legal Stud. 83, 90 (1977); cf. Speidel,
Excusable Nonperformance in Sales Contracts: Some Thoughts About Risk Management,
32 S.C.L. Rev. 241, 273 (1980).

90. See Bazerman & Carroll, supra note 85 (used negotiators); Hershey, Kunreuther
& Shoemaker, supra note 84 (undergraduates and M.B.A. students with below average
quantitative backgrounds). Cf. Hershey & Shoemaker, supra note 84 (subjects were M.B.A.
students with below average quantitative background); Kaplow, supra note 75, at 549,

561 n.l46 ("as a rough approximation" institutional investors and corporations can be

expected to make accurate probability assessments); McNeil, Pauker, Sox & Tversky, supra
note 85 (used radiologists); Schoemaker & Kunreuther, supra note 84 (experimental subjects
were undergraduates and individual clients of an insurance agency); Framing of Decisions,

supra note 81 (experimental subjects were undergraduates). Arguably, both professions

could be deemed sophisticated decision makers for purposes of the decision problems

presented to each. The argument in this paragraph of the text does not depend on making

the claim.

91. Cf. Gillette, supra note 20, at 549 ("There is reason to believe, however, that

cognitive dissonance does not pose a serious problem in the context of long-term commercial

contracts.").

92. See Schurr, Effects of Gain and Loss Decision Frames on Risky Purchase

Negotiations, 72 J. App. Psy. 351, 357 (1987) (risk-taking behavior among both M.B.A.
students and professional buyers affected by decision frames, where gain and loss frames



1990] COMMERCIAL IMPRACTICABILITY 93

Empirical evidence aside, description-dependence is present across

classes of decision makers. The reason has already been noted: fore-

seeability of an event is relative to a set of propositions, under the

descriptions available. Commercially sophisticated individuals or cor-

porations, like lay persons, are presented with a set of propositions. Of
course, different sets may be presented. The description problem is the
problem of selecting among elements of that set in a nonarbitrary way
such that an event is foreseeable. It is a problem that is present at the

negotiation stage of a contract, whether short- or long-term.

Commercially sophisticated individuals or corporations may render
the description problem unimportant. This occurs if both decision makers

are not subject to framing and pseudo-certainty effects of the type

identified by Tversky and Kahneman.^^ The decision maker's probability

and utility assignments would remain invariant across extensionally equiv-

alent descriptions. Application of the compensation-expectation principle

are net profit and expense, respectively); cf. Bazerman, The Relevance of Kahneman and
Tversky 's Concept of Framing to Organizational Behavior, 10 J. Mgmt. 333, 338 (1984)
("combination of information format presented and the individual's framing tendencies

will affect risk attitudes and subsequent escalation behavior of managers, politicians, and

academicians"); Slovic, Fischhoff & Lichtenstein, Regulation of Risk: A Psychological
Perspective, in Regulatory Policy and the Social Sciences 241 (R. Noll ed. 1985)
(experts not immune to cognitive error).

93. No empirical studies are available concerning the presence of framing effects
among sophisticated decision makers. However, existing experimental results are consistent
with their presence among experienced managers and business executives. Researchers have
found that 71% of the managers studied were risk-seeking for below-target levels of
wealth. Laughhunn, Payne & Crum, Managerial Risk Preference for Below-Target Returns,
26 Mgmt. Sci. 1238, 1243, 1245-46 (1980). Approximately 45<7o were risk-averting when

a prospect of a ruinous below-target level loss was presented in a gamble. Id. at 1243,

1245-46. Choice behavior did not significantly vary as between personal and business

decisions. Id. at 1243, 1245-46. Experiments varying target levels of managers while

maintaining the set of possible outcomes replicated Laughhunn, Payne and Crum's. Payne,

Laughhunn & Crum, Further Test of Aspiration Level Effects in Risky Choice Behavior,
11 Mgmt. Sci. 953 (1981). MacCrimmon and Wehrung found that over 70*^0 of the
business executives studied were risk-averse for business decisions involving potentially

large losses (one-half of their assigned capital budget); and that approximately 60% of
the executives were risk-seeking for decisions involving only gjiins. The result held for

both personal and business decisions. K. MacCrimmon 8l D. Wehrung, Taking Risks:
The Management of Uncertainty 111-22 (1986).

These results suggest that managers and business executives' utility functions are not

uniformly concave across all wealth or target levels. The lack of uniform concavity may
be due, at least in part, to framing effects. (Interestingly, MacCrimmon and Wehrung
found that the executives studied were more risk-seeking for gambles the authors describe

as threats than for gambles described as opportunities. Id. at 115. The authors label a

gamble with a .5 probability of loss a "threat." A gamble with a .9 probability of gain
is labeled an "opportunity." Altered choice behavior was given mean-equivalent, differently

labeled gambles not subject to the test.)



94 INDIANA LAW REVIEW [Vol. 24:65

outlined in Section IV then would be straightforward. The description
problem, however, would still exist. Alternative available descriptions of

the same prospective event presents a possibility of choice among them.
It is tempting to conclude that the description problem has no

solution. This, I think, is correct. It is also tempting to draw this distinct

and stronger conclusion: That the description problem in every application

has no solution and that, therefore, a choice among possible descriptions
is arbitrary. Such a conclusion should be resisted for two reasons. First,

the problem is not one of selecting among descriptions available to the
contracting parties. That is, it is not an agent-specific problem. Rather,

the problem is one of selecting among descriptions available at any
particular time or, more restrictedly, available to the participants in a
given market at a particular time. Foreseeability, not being foreseen by

a party, is the relevant standard. Second, the description problem is one

of selecting among available alternative descriptions such that an event
is foreseeable under at least one of them. It is not a problem of selecting

among descriptions simpliciter.
These two points can be applied as follows: The relevant set of

descriptions is, most restrictedly, the set consisting of descriptions con-

cerning a future event available to participants in a particular market.

The restriction is reasonable given the sophistication of contracting parties
in typical commercial settings. Only a subset of descriptions is also

relevant to the foreseeability of an event. Roughly, this is the subset

under which the event is explainable or predictable at a particular time.

("Roughly'* because a more precise criterion would allow nonexplanatory
or nonpredictive descriptions that are nomonologically or probabilistically

related to explaining or predicting descriptions.) **That x is an atomic

particle" may be a nonexplanatory or nonpredictive description. "That
X is uranium** may be a description, at a particular time, from which
harm to humans can be explained or predicted.

An event is foreseeable if it is explainable or predictable under a
description available at a particular time. Hence, an event is foreseeable

at a particular time if, under some description available to market

participants, it can be predicted or explained under that description. To
be sure, there may be several alternative explaining or predicting de-
scriptions available at any given time. However, this neither shows that

the event was not foreseeable at that time under each of those descrip-

tions, nor that the proposed selection is arbitrary. In fact, the selection

is nonarbitrary. If a contracting party*s expectations of a price increase

are to be determined, then only the available descriptions under which

the price increase is explainable or predictable are relevant. Hence, as

applied to contracting parties* expectations, the above proposal can count

as a solution to the description problem.

The compensation-expectation principle places a further constraint

on the set of available descriptions: only those predicting or explaining



1990] COMMERCIAL IMPRACTICABILITY 95

descriptions under which an event is foreseeable and for which com-

pensation is provided are admissible. On the above proposal, only avail-
able explaining or predicting descriptions are considered. An event is
foreseeable relative only to those descriptions. On the compensation-
expectation principle, only the expectations for which a party has been

compensated ex ante are to be considered. The set of compensated-for
expectations can be smaller than the set of expectations that have ex-

plaining or predicting descriptions. The reason is straightforward: not
all expectations are compensated ex ante. Only the intersection of the

two sets is to be considered. In other words, only those expectations

which have corresponding explaining or predicting descriptions and for

which compensation is provided are admissible. Just those expectations

satisfy the compensation-expectation principle. **X will break*' is perhaps

foreseeable under the description **X is a chair," in tandem with other

information, but compensation ex ante might not be provided for this

expectation. Presumably this is typically the case because all events can

be (at least) predictable under some description, however general. The
compensation-expectation principle requires something more: ex ante

compensation.

VI. The Proposed Principle of Loss Distribution Applied

Section 2-615 is to be applied in conjunction with the compensation-

expectation principle as follows. Satisfaction of the statutorily specified

conditions of section 2-615 are first determined. This requires deter-

mination of the cause (causes) of an input price increase. Determination

of whether risk of its occurrence was allocated by the buyer and the

seller is also required. And the foreseeability of the price increase must
be established. The compensation-expectation principle is applied next.
Its application requires representing the foreseeability (expectation) of

the input price increase by a probabiUty density function. Additionally,

a determination of the seller's cost is required if the input price rise

falls within its range.

The seller's compensated-for expectation is the product of the mean
of the probability density function times the seller's cost within its range,^"^

as reflected in the contract price. This is the seller's risk premium. By
definition, it satisfies the compensation-expectation principle. Let p be

the proportionate variation in the contract price associated with the

seller's risk premium. That is, p = (risked variation in contract price

94. This is true only if the reconstructed probability density function describes a

normal distribution. The assumption is justified by the prevalence of its use, by its sampling

properties, and by the possibility that a court could adopt it as a presumption in applying

the compensation-expectation principle absent contrary evidence. For the latter part of

the justification, compare infra text following note 111.



96 INDIANA LAW REVIEW [Vol. 24:65

- contract price)/contract price. Graphically, it corresponds to the distance

of the end point of the probability distribution from the distribution's

mean. Then, if the seller breaches, the buyer's damages are calculated

as:

Damages = (p x contract price) + (incidental expenses +
consequential damages)^^

These are the damages the seller deserves to bear.

An example may be useful. The following is loosely based on the
facts presented to the court in Louisiana Power & Light Co. v. Allegheny
Ludlum Industries, Inc.^ Seller is a steel concern, and buyer is a public
utility operating several nuclear power plants. The parties execute a
contract according to which seller agrees to supply condenser tubing

needed to operate buyer's nuclear reactors. The contract price is $1,100,000.
Condenser tubing production requires three important components: elec-

trolytic nickel, low carbon ferrochrome, and labor. Prior to the stipulated

delivery date, the input prices of the three components increase. Seller's

total production cost increases to about thirty-nine percent over the

contract price. Seller refuses to deliver the tubing at the contract price,

and the buyer purchases the item from another supplier. The cover price

is $1,750,000. No incidental expenses or consequential damages are sus-
tained by buyer. The buyer seeks damages of $650,000, the sum calculated
under section 2-712(2) of the Uniform Commercial Code.

What are the damages that the seller deserves to bear? That is, what
portion of the $650,000 loss on the contract does the seller deserve to

bear? In order to allocate the deserved loss, the court must apply the

compensation-expectation principle. Doing so requires a determination

of the seller's risk premium. Suppose the court finds that the foreseeable

range of aggregate input price variation for electrolytic nickel, low carbon

ferrochrome, and labor is twenty percent. Suppose too that there exists

a positive (proportional) relation between these foreseeable input cost

increases and the total production cost of the condenser tubing. Finally,

suppose that it is found that seller's price for bearing the risk of a

twenty percent variation in aggregate input prices is ten percent of the

contract price. Because the contract price is $1,100,000, seller's production

cost includes a $100,000 charge to bear the risk of that variation. Seller

therefore is compensated to incur a risk pf a twenty percent increase

(decrease) in the three mentioned inputs. Also, because there is an

assumed positive proportional relation between input price increases and

total production cost, seller is compensated to bear a risk of a variation

95. The Code allows recovery of incidental expenses and consequential damages

under § 2-715(l)-(2). The damage measure above preserves this allowance.

96. 517 F. Supp. 1319 (E.D. La. 1981).



1990] COMMERCIAL IMPRACTICABILITY 97

in production costs between $880,000 and $1,320,000 ($1,100,000 x f-
.20 = $880,000, $1,320,000). The aggregate input price increase of
electrolytic nickel, low carbon ferrochrome, and labor is about thirty-

nine percent. Hence, by assumption,'^ the production cost of the con-

denser tubing also has increased by about thirty-nine percent, to almost

$1,530,000 ($1,100,000 X 1.39 « $1,530,000). Buyer's cost of cover was
$1,750,000, of which seller was compensated ex ante to bear the risk

of production costs to $1,320,000. Because the contract price was

$1,100,000, seller's compensated-for risk of loss therefore is $220,000

($1,320,000 - $1,100,000 = $220,000). Seller cannot reasonably complain
about bearing losses up to that amount. No incidental expenses or
consequential damages were sustained by buyer. Hence, seller deserves

to bear a loss of $220,000 on the contract. The same amount is given

by the general measure of buyer's recoverable damages presented above:

Damages = (.20 x $1,100,000) -»- = $220,000
because the value of p is .20 (($1,320,000 - $1,100,000)/ $1,100,000 =

.20). Buyer justifiably bears a loss of $430,0(X), the remainder of the

loss on the contract ($650,000 - $220,000 = $430,000).
The buyer's damages under section 2-712(2) of the Code are calculated

as:

Damages = (cover price - contract price) -I- (incidental
expenses -I- consequential damages)

Buyer's recoverable damages are $650,000 under this familiar measure

(($1,750,000 - $1,100,000) + = $650,000). The Code measure and
the measure of deserved loss above give the same damages only when

p = .59 (($1,750,000 - $1,100,000)/$1, 100,000 = .59). Seller in this
case would have been compensated to bear the risk of the entire price

increase. The two measures typically need not yield the same amount

of damages.'®

There is an objection to the above application. It concerns the

informational component of a court's institutional competence. In par-

ticular, it concerns the informational constraints a court faces. The

objection is that the above application requires courts to make a number
of complex factual findings.'' Issues of causation and foreseeability (p)

97. Specifically, the assumption is that no substitution among inputs is possible
in the short-run and, as a simpUfication, that all input prices for condenser tubing increase

by about 39*^o. The assumption, of course, may be violated in a given case. Whether it
is violated is an empirical matter to be estabUshed by proof.

98. In a case of partial excuse, in which incidental expenses and consequential

damages are present, courts obviously have to "make adjustments under the various

provisions of this Article [Two]" to the latter two variables of both measures in the text.

U.C.C. §2-615 comment 6 (1987).
99. See A. Schwartz & R. Scott, Commercial Transactions: Principles and



98 INDIANA LAW REVIEW [Vol. 24:65

are examples. Further, a court must disaggregate a contract price to

determine a seller's risk premium. It must do so in order to apply the

compensation-expectation principle. However, courts are unsuited to make
such findings. This is not a matter of deeming the findings to be issues

of law and not fact.^°° It is that, regardless of how the issues are
denominated, courts are unsuited to make the required findings. Because
courts are unsuited, the application exceeds a court's institutional com-

petence. And because the compensation-expectation principle requires
application, a court is not competent to implement the principle. The

objection concludes that the proposed principle should be rejected.

The objection fails. To see this, notice first the structure of the
objection. It moves from an assertion about complexity in informational

requirements to a claim about institutional unsuitability, to a conclusion

about a principle of loss distribution. Each move is unsound. Consider
each move in turn. To be sure, the factual findings required are often
complex. Input price increases, for instance, may be the causal product
of a number of distinct events. Determining a seller's expectations via

a reconstruction of a probability density function also is difficult. It,

in part, requires identifying the explanatory or predictive descriptions

under which an event is foreseeable at a particular time in the seller's

market. However, both types of finding are not, in principle, different

from those required in other adjudicatory contexts, nor are they nec-

essarily of greater complexity. Compare the findings required in tort
litigation in which fault is being determined. Causal issues here can be

no less complex. Here, too, questions of alternative or aggregate causes

may be raised. Issues concerning the foreseeability of particular events
can also be present. Determining the relevant set of explanatory or

predicting descriptions also is required here. That such issues in tort

contexts are treated as matters of fact is irrelevant. Deeming them matters

of fact identifies the decision-maker who is to decide them. It does not
show that the order of complexity in deciding such issues is any different.

Policies 417-18 (1982); Dawson, Judicial Revision of Frustrated Contracts: The United

States, 64 B.U.L. Rev. 1, 17, 37 (1984); Schwartz, supra note 13, at 11, 20.

100. Courts have been unclear as to which of § 2-6 15 's conditions are matters of

law and which are matters of fact. Compare Oosten v. Hay Haulers Dairy Employees &
Helpers Union, 45 Cal. 2d 784, 291 P.2d 17 (1955) and Housing Auth. v. East Tenn.

Light & Power Co., 183 W. Va. 64, 31 S.E.2d 273 (1944) (both applying common law
doctrine of impossibility, treating "basic assumption" as a matter of fact) with Koppers

Co. V. United States, 405 F.2d 554, 558 (Ct. CI. 1968) (issue as to whether performance

of contractual specifications is commercially impracticable is a question of law) and

Sunflower Elec. Coop. v. Tomlinson Oil Co., 7 Kan. App. 2d 131, 139, 638 P.2d 963,

969 (1981) (excuse of performance by the doctrine of impossibility is a determination of

law). See also Restatement (Second) of Contracts § 261 reporter's note (1979) ("basic

assumption" is a matter of law).



1990] COMMERCIAL IMPRACTICABILITY 99

Of course, deeming the matters ones of fact in tort does allocate the
determination to the jury, and juries may decide them on normatively
irrelevant grounds — say, by their distributional preferences. The same
is true of any issue, however complex, to be resolved by a jury. The
possibility is not created by the differential complexity of factual issues

in contract. Hence, the first move is unsound.
The second move is also unsound for the same reason. The fact of

complexity in findings does not show an institutional unsuitability to
determine them. Compare a tort litigation in which fault is being assigned
among multiple defendants. Under a pure comparative negligence rule,
a jury has to decide causal issues and determine what each defendant

could foresee. Each defendant's probability density function is determined
in practice, if not in name. Judge and jury are presented with the same
evidence. Again, the fact that different issues are distributed between

them for decision does not show that judges cannot suitably decide
issues reserved for a jury, at least not when suitability concerns factual
issues, as it does here. Of course, contract and tort contexts differ. In
contracts, the compensation-expectation principle requires disaggregation

of the contract to determine a seller's risk premium. However, the fact

that the value of a further variable has to be determined does not, by

itself, indicate a greater degree of complexity. Nor does it show that
courts are comparatively unsuited to make the requisite findings. The
objection, at most, shows that both courts and juries are unsuited to

determine complex issues of causation and foreseeability.'^' This may be
true, but it is not an objection peculiar to the proposals in Sections IV

and V.

The third move from institutional unsuitability to a defect in the
compensation-expectation principle also fails. Even if the first two moves
were sound, it still would not follow that the proposed principle should

be rejected. Additional devices can be introduced to nullify a court's

assumed fact-determining limitations. One such procedural device is the
appointment of a special master. '^^ An appointed master can engage in

101. C/ Brennan, Causal Chains and Statistical Links: The Role of Scientific
Uncertainty in Hazardous-Substance Litigation, 73 Cornell L. Rev. 469, 491-501 (1988);
Sugarman, The Need to Reform Personal Injury Law Leaving Scientific Disputes to
Scientists, 248 Sci. 823 (1990) (both urge the unsuitability of judicial fora for resolution

of disputes involving complex causal issues).

102. See Fed. R. Civ. P. 53(a), (b) (a judge is permitted to appoint a special master

empowered to make findings on particular issues). In an arbitral forum, arbitration rules
allow the appointment of an expert to make findings on issues to be determined by the
arbitration. Cf Arb. Rules of the U.N. Comm. on Int. Trade Law, Art. 27 (1976) ("The
arbitral tribunal may appoint one or more experts to report to it, in writing, on speciHc
issues to be determined by the tribunal."); ICC Rules of Conciliation and Arb., Art. 14

(1988) ("The arbitrator may appoint one or more experts, define their Terms of Reference,
receive their reports and/or hear them in person.").



100 INDIANA LAW REVIEW [Vol. 24:65

the factual inquiry necessary to apply the compensation-expectation prin-

ciple. Evidentiary devices also are available. In particular, assignments

of burdens of proof in demonstrating satisfaction of the proposed prin-

ciple can be introduced. Given an assignment of the burden of proof,

a court does not have to determine which events were foreseeable. It

does not have to determine the seller's risk premium, nor to establish

the cause (causes) of an input price increase. The court (or jury) need
only determine whether a party has successfully sustained its burden.

In fact, courts already impose the burden of proof on a party claiming

the excuse of commercial impracticability.'^^ The extension of the seller's

burden of proof to include the compensation-expectation principle itself

is modest. A standard for satisfying the assigned burden of proof could
be altered as well. Specifically, the standard could be one of substantial

or clear and convincing evidence. The point here is neither to advocate

a particular assignment of the burden of proof between seller and buyer,

nor to advocate a favored standard for satisfying the assigned burden

of proof. The point, instead, is simply that the proposed principle does

not stand or fall with a court's assumed informational limitations. This

is sufficient to show that the objection fails.

Unexceptional judicial practice demonstrates the effects of assigning

the burden of proof. Heat Exchangers v. Map Construction Corp.^^
involved a contract for the construction and installation of air condi-

tioners and plumbing. The seller failed to effect timely delivery, and

the buyer contracted with another supplier at a higher cost. The court
awarded the buyer damages. On appeal, the seller contended that its
failure to deliver was excused because delivery would have been com-

mercially impracticable. This contention was based on a set of supervening

events, including the unavailability of a supply of component parts. The

court rejected the seller's allegation. In doing so it placed the burden

of proof on the seller as to the impracticability of timely performance. '°^

The court found the seller presented no evidence as to the unavailability

of some component parts **beyond the broad generalities" of seller's

employee.'^ The Heat Exchangers court, by assigning the burden of

103. See Opera Co. of Boston, Inc. v. Wolf Trap Found, for Performing Arts,

817 F.2d 1094, 1102 (4th Cir. 1989); Alimenta (U.S.A.), Inc. v. Gibbs Nathaniel (Canada)

Ltd., 802 F.2d 1362, 1364 (11th Cir. 1986); Louisiana Power & Light Co. v. Allegheny
Ludlum Indus., Inc., 517 F. Supp. 1319, 1324 (E.D. La. 1981); Eastern Airlines Inc. v.

Gulf Oil Corp., 415 F. Supp. 429, 438, 440 (S.D. Ha. 1975); supra notes 85-86, 90, 93.

104. 34 Md. App. 679, 368 A.2d 1088 (1977).
105. Id. at 688, 368 A.2d at 1093; cf. Jennie-O Foods Inc. v. United States, 217

Ct. CI. 314, 329 (1978) (plaintiff, the party making an impracticability claim, has the

burden of proof as to the unavailability of alternative supplies).

106. Heat Exchangers, 34 Md. App. at 688, 368 A.2d at 1093.



1990] COMMERCIAL IMPRACTICABILITY 101

proof to seller, avoided the necessity of determining the available supply

of component parts. It therefore avoided making its decision dependent

on its informational limitations.

Iowa Electric Light & Power Co. v. Atlas Corp}^^ is well known
and less prosaic in effect. In this case, an assignment of the burden of

proof relieved the court from determining the extent to which input

price increases were foreseeable. Iowa Electric involved a four-year con-

tract for the supply of uranium concentrate. The contract contained a
price escalation clause allowing an increase of 3.75 cents per pound of
shipped uranium concentrate per month. The base price in 1975 was
"$7.10 per pound escalating to $8.45 per pound in 1978. "'^'^ By 1978,
the seller's production costs were $17.80^^ per pound and the market
price had gone to about $43.00 per pound. ^•^ The seller's loss on the
contract in its second year was estimated at approximately $1.8 million.'^'

The buyer sought specific performance of the contract, and the seller

counterclaimed for an equitable adjustment of price based on commercial

impracticability.

The court rejected the seller's counterclaim. Its reasoning is revealing.

Price increases in many of the inputs were conceded to be unforeseeable."^
The court admitted that **[i]t would be unfair to expect Atlas [seller]

to have prophesied the magnitude of the increases complained of . . . .""^

The seller's counterclaim was rejected because it failed to bear its burden
of proof as to how many of the input price increases were unforesee-
able. ""^ The court impUcitly acknowledged that the actual production
cost of uranium concentrate fell outside of the seller's probability function

for production cost increases. It simply lacked the information needed

to construct seller's ex ante probability function. Thus, the effect of the

assignment of the burden of proof is significant. Requiring seller to

provide the information makes the court's acquisition of the information

unnecessary. Either seller supplies the requisite data or the entire extent

of production cost increases is treated as foreseeable. In the former

instance, the court obtains the information from seller. In the latter

instance, acquisition of the information is superfluous: the court simply

107. 467 F. Supp. 129 (N.D. Iowa 1978).

108. Id. at 137.

109. Id. at 132.

110. Id. at 132, n.5.

111. Id. at 131, n.2.

112. Id. at 132, 137 & n.lO.
113. Id. at 135.

114. Id. at 132-33, 133 n.6, 135; cf. In re Ocean Air Tradeways, Inc., 480 F.2d

1112, 1117 (9th Cir. 1973) (the party advancing the defense of impossibility of performance

must prove that the events alleged to frustrate performance were not reasonably foreseeable).



102 INDIANA LAW REVIEW [Vol. 24:65

assumes that the entire input price increase is foreseeable. In both

instances, the court itself need not provide the requisite information.

Again, as in Heat Exchangers, the court avoids making its decision

dependent on its own informational limitations.
Consider, finally, a related objection. This objection is not just that

courts are subject to informational limitations. It is that informational

limitations induce unpredictability in case outcomes, and that induced

unpredictability in case outcomes impairs the stability of contracts. The
compensation-expectation principle, according to the objection, induces

unpredictability. This objection fails, too, because the proposed principle

does not, by itself, generate unpredictability in case outcomes. At best,
only the principle together with devices for its application would do so.

Let C = the cost to seller of performance satisfying the compensation-
expectation principle, D = the damage schedule a court assigns to the
seller if the seller breaches (D > 0), and p = the probability that the
seller is found liable for a particular level of damages. Seller will breach

if, and only if, C > E PjDj. The values of p and D are not determined
by the compensation-expectation principle itself. They are determined by
devices such as the assignment of burdens of proof and a court's adoption

of particular damage schedules given uncertainty about the buyer's dam-

ages. Both devices are independent of the compensation-expectation prin-

ciple.

Contractual stability is unimpaired if C < .E pjOj. Values of p can
be such that both the seller's and buyer's uncertainty about litigation

outcomes is reduced. This can occur as a result of assigning burdens

of proof. D's values can also be set by a court adopting damage schedules

when presented with uncertainty concerning the buyer's damages. For
instance, a court could adopt the following rule: Buyer's damages (D)

are equal to the cost of cover when the seller fails^ to establish the value
of C. If p is "high enough" and D > C, C < .E PjDj, then seller will
perform. Contractual stability is, therefore, preserved. The point here,

as in two paragraphs ago, is not to advocate the use of such devices.

It is simply that the compensation-expectation principle induces neither

unpredictability nor contractual instability. The proposed principle cannot

be rejected on this basis.

VII. Conclusion

Section 2-615 of the Uniform Commercial Code excuses contractual
performance under statutorily specified conditions. Difficulty in applying

the section is at least partly the result of the absence of a normative

principle of loss distribution. The proposal in Sections III through VI

provides that normative principle. Ex ante compensation is its basis: A
commercially sophisticated party cannot reasonably object to a loss, the



1990] COMMERCIAL IMPRACTICABILITY 103

risk of which it has been compensated to bear. This is the compensation-

expectation principle argued for in Section IV. Its prescriptive force

derives from the absence of any justifiable grounds to object. Given

compensation ex ante, it is because a particular loss cannot reasonably

be objected to that the loss can be imposed. The compensation-expectation
principle is * internal*' to the Code because it is supported by Code
provisions, comments, and case law (Section IV). Competing normative

principles of loss distribution violate the constraint of **internalness"

(Sections II - III).

The proposal in Sections III - VI uses the outlined principle of loss
distribution as follows. Contractual loss is to be distributed according

to the parties* desert. A contractual loss is deserved when a party cannot
justifiably complain about its imposition on that party. And a com-
mercially sophisticated party cannot justifiably object to a loss, the risk

of which it was compensated to bear (Section IV). Deserved loss is

coextensive with the foreseeable risks of loss for which a party has been

compensated. Hence, in applying section 2-615 of the Uniform Com-
mercial Code, courts are to determine the set of foreseeable risks for

which a party was compensated. Those risks, by definition, are a party's

risk-compensated set of expectations. A party's risk-compensated ex-
pectations of an event are the set of compensated expectations such that

there is at least one description of the event under which it is explained

or predicted (Section V). Courts are to identify those descriptions and

reconstruct a party's risk-compensated expectations. Section VI considers

and rejects objections based on institutional incapacity, informational

complexity, and induced contractual instabihty.

The proposal has comparative virtues. To see this, begin by distin-
guishing the following three elements: (1) the conditions under which

legal relief is granted; (2) the nature of the relief granted; and (3) the

extent of relief. A determination of the first element does not also
determine the other two elements. For instance, a court could find seller's

performance to be commercially impracticable while adjusting the contract

price and ordering specific performance. '^^ In such a case, statutorily

115. See, e.g., ALCOA v. Essex Group, Inc., 499 F. Supp. 53, 57 (W.D. Pa. 1980)
(contract terms equitably adjusted upon a finding of commercial impracticability); G.W.S.

Serv. Stations, Inc. v. Amoco Oil Co., 75 Misc. 2d 40, 346 N.Y.S.2d 132, 140-43 (N.Y.
Sup. Ct. 1973) (seller's delivery of a contractually specified quantity of gas found to be

commercially impracticable; court ordered specific performance of a judicially adjusted

quantity); McGinnis v. Clayton, 312 S.E.2d 765, 779 (W. Va. 1984) (Harshbarger, J.,

concurring) (a court can equitably adjust contract terms and order specific performance

upon a finding of commercially impracticability). But see Iowa Elec. Light & Power Co.
V. Atlas Corp., 467 F. Supp. 129, 139 (N.D. Iowa 1978) (equitable adjustment is not

possible under buyer's specific performance remedy, under § 2-716 of the Uniform Com-
mercial Code).



104 INDIANA LAW REVIEW [Vol. 24:65

specified conditions provide the foundation upon which seller's per-
formance is found to be commercially impracticable — this constitutes
element (1)."* Buyer's relief takes the form of an order for specific

performance of the adjusted contract, element (2), and the extent of

the buyer's relief is determined by the difference between the contract

price and the adjusted contract price of the seller's performance, element

(3). The court must make separate findings as to each element. Comment
6 of section 2-615 can be construed to recognize this need:

In situations in which neither sense nor justice is served by either

answer when the issue is posed in flat terms of **excuse" or
'*no excuse," adjustment under the various provisions of this

Article [Two] is necessary . . . [including] the general policy of

this Act to use equitable principles in furtherance of commercial

standards and good faith. '•^

A fully specified doctrine of commercial impracticability at least requires
that elements (1) and (3) be given precise content. (**At least" because

element (2) concerns the type of relief granted and, therefore, forms

part of a general account of remedies.) Hence, alternative accounts of

commercial impracticablity can be compared with respect to elements

(1) and (3).
The above proposal has comparative virtues, at least as compared

to accepted judicial practice. One such virtue is a well-defined principle
for determining when performance becomes commercially impracticable.
Compare in this regard In re Westinghouse Electric Corp. Uranium
Contracts Litigation^}^ to Louisiana Power. ^^^ At issue in In re West-

inghouse was whether losses of about $100 million for temporarily storing

and ultimately removing spent nuclear fuel rendered the seller's per-

formance impracticable. The district court determined performance to

be practicable. It did so by considering the proportion of loss to the

value to seller (Westinghouse) of the entire contract. ^^° The Court of

Appeals for the Fourth Circuit disagreed with the District Court's de-

termination. In a decision in which it partially reversed the District

Court's judgment, it employed a different formulation: the cost of the

116. See supra text accompanying notes 4-8.

117. U.C.C. § 2-615 comment 6 (1987).
118. Florida Power & Light Co. v. Westinghouse Elec. Corp., 597 F. Supp. 1456

(E.D. Va. 1984); In re Westinghouse, 517 F. Supp. 440 (E.D. Va. 1981).

119. Louisiana Power & Light Co. v. Alleghany Ludlum Indus., Inc., 517 F. Supp.
1319 (E.D. La. 1981).

120. See Florida Power & Light Co., 597 F. Supp. at 1477-78; In re Westinghouse,
517 F. Supp. at 453.



1990] COMMERCIAL IMPRACTICABILITY 105

contemplated performance to the cost of the alternative performance."^'

The Louisiana Power court used still a different proportion. There, a
finding that the seller's performance was practicable was made '*more
apparent*' when the proportion of loss to the seller's annual profit on
its manufacturing plant was considered. "^^ All three courts agree that

impracticability is to be determined by a comparison of cost.*^^ The

disagreement among the courts concerns the favored comparison. No
court provides a justification for treating any particular proportion as

indicating impracticability. Instead, appeal is made to proportionate cost
increases present in other cases. "^'^ Impracticability as indicated by ranges

of proportionate cost increases is not justified. Hence, the measures are

ill-defined. The compensation-expectation principle requires that a party

can reasonably complain about the imposition of loss in excess of the

foreseeable risks the party was paid to bear. It provides a justification

for treating particular losses as indicating impracticability. Hence, the

principle for determining impracticability is well-defined. It gives content

to the Code's oracular comment that impracticability results when **the
essential nature of the performance" is altered.""

Another comparative virtue is one of unification: the presence of

a well-defined principle of loss distribution given that performance is

found to be commercially impracticable. The same principle that deter-

mines when also determines how much to excuse, and it does so in both
instances by resort to bargain-specific facts. Therefore, the extent of

relief granted, element (3), is connected to the conditions under which

relief is granted, element (1). The justificatory relation between the two

elements is not adventitious.

121. Florida Power & Light Co. v. Westinghouse Elec. Corp., 826 F.2d 239, 277
(4th Cir. 1987).

122. Louisiana Power & Light Co., 517 F. Supp. at 1324 n.4.
123. See supra notes 14, 114.

124. See Louisiana Power & Light Co., 517 F. Supp. at 1325-26 (citing Transatlantic
Fin. Corp. v. United States, 363 F.2d 312, 315 (D.C. Cir. 1966) and Iowa Elec. Light

& Power Co. v. Atlas Corp., 467 F. Supp. 129 O^.D. Iowa 1978), rev'd on other grounds,
603 F.2d 1301 (8th Cir. 1979)); cf. Iowa Elec. Light & Power Co., 467 F. Supp. at 140
(citing American Trading & Prod. Corp. v. Shell Int'l Marine Ltd., 453 F.2d 939, 942
(2d Cir. 1972) (citing Transatlantic Fin. Corp., 363 F.2d 312 and English precedents

involving proportionate cost increases)).

125. U.C.C. § 2-615 comment 4 (1987) ("Increased cost alone does not excuse
performance unless the rise in cost is due to some unforeseen contingency which alters

the essential nature of the performance.").

Whether the principle is to be used to determine impracticability of performance is

a distinct issue. Its resolution depends on evidentiary matters and the allocation of burdens

of proof. See supra text accompanying notes 107-112. Both items pertain to the application

of the compensation-expectation principle and not to its content.



106 INDIANA LAW REVIEW [Vol. 24:65

Consider in this respect Florida Power & Light v. Westinghouse
Electric Corp.^^^ again. There, the district court determined that the

seller's performance was practicable based on its favored comparison of

costs. '^^ It went on to find that part of the seller's loss was the result
of an unforeseeable contingency, the government's delay in honoring its

commitment either to build a nuclear fuel reprocessing plant or to
construct storage sites. ^^* In allocating the loss represented by the interim

storage costs, the court appealed to **it[s] own sense of fairness. "•^^ The
specific considerations included the benefit which the buyer's ratepayers

already received and the seller's tardiness in removing some spent fuel

to an off-site location. The court divided the interim storage costs equally.
The point here does not concern the distribution of loss upon which

the court decided. Nor does it concern the specific content of the court's
imphcit principle of loss distribution. Rather, the point concerns the

relation between the court's finding and the facts appealed to by the

court in distributing loss. It is that the relation remains mysterious

without a justificatory connection between a finding of an unforeseeable

contingency and partial or full excuse. The question is simply put: what
is it about unforeseeable contingencies which justifies distributing loss

in a particular way?

The facts the court cites do not justify its distribution between buyer
and seller because they are irrelevant to the terms of the bargain. The

ratepayers' benefit from the performance of the contract is an element

that is extrinsic to the transaction. It is relevant only to the value of

performance to the ratepayers. The terms of the contract remain the
same even if the ratepayers* benefit from the performance of the contract

would be different. An ability to bear a loss (cost) is, by itself, irrelevant
to the justification of imposing that loss (cost) in the first place. '^^

Similarly, the seller's tardiness in removing some spent fuel to an off-

site location is irrelevant. The court has already decided that seller's

contractual obligation simply was to dispose of spent fuel, either by

storing or reprocessing.^^' Removal of spent fuel to an off-site storage

location was not part of the contract.

126. 597 F. Supp. 1456 (E.D. Va. 1989).

127. Id. at 1478.

128. Id.

129. Id.

130. Cf. Asphalt Int'l, Inc. v. Enterprise Shipping Corp., S.A., 667 F.2d 261, 266

(2d Cir. 1981) ("The doctrine of commercial impracticability focuses on the reasonableness

of the expenditure at issue, not upon the ability of a party to pay the commercially

unreasonable expense."). Case support aside, the statement in the text is at least true

when commercially sophisticated actors are involved, as in Florida Power & Light Co.,
597 F. Supp. 1456.

131. In re Westinghouse, 517 F. Supp. at 446-47.



1990] COMMERCIAL IMPRACTICABILITY 107

The compensation-expectation principle appeals to facts pertinent to
the terms of the bargain. In particular, appeal is made to the set of
foreseeable risks present to a party at the time the contract was executed

and the compensation was received to bear those risks. Hence, the

proposed principle is bargain-specific. Contractual loss is distributed by

reference to those facts. It is because a party was compensated to bear

a set of foreseeable risks ex ante that ex post imposition loss upon the

party is justified. The amount of relief a court grants is justified by
the time when relief is granted.

A proviso needs to be noted. It may be that a court would do
better not to allocate loss according to the compensation-expectation

principle in a particular case. That is, the case may be one in which
distributing a loss according to desert is unjustified. The prospect of a
seller's or a buyer's consequent bankruptcy or perverse incentive effects

may provide such cases. Alternatively, the party may not be a com-
mercially sophisticated actor, in which case the compensation-expectation

principle itself may not be satisfied. Either possibility would not show
that the above proposal is wrong. Each only shows that the proposal

has a restricted scope. There is nothing inconsistent about a court making

the following three statements: (1) seller was compensated to bear a risk

of a price rise up to $100 on a contract price of $20 and, therefore,

seller can reasonably complain about bearing a loss in excess of $100;

(2) given a price increase to $200, seller deserves to bear $100 of the

loss; (3) but seller nonetheless ought to bear the entire $200 loss. Claims

of desert can be overridden by other normative claims. There is nothing

unusual here. Moreover, the Uniform Commercial Code allows the result
via section 1-103 which, subject to displacement by Code provisions,
makes general equitable principles applicable to a transaction. When
normative claims other than desert require that the compensation-ex-

pectation principle not be employed, further concerns are present. Their

presence does not defeat the above proposal.

There is good reason to assume that the compensation-expectation

principle will not be overriden via section 1-103. Overriding it requires

that application of section 1-103 yield a different result from the com-

pensation-expectation principle. This, in turn, requires that the general

equitable principles referred to in section 1-103 apply and, as applied,

impose a different distribution of loss. However, the requirements rarely

will be satisfied jointly. Identification of operative equitable principles

is needed because mere mention of their existence is insufficient. '^^ Section

132. See Hillman, Construction of the Uniform Commercial Code: UCC Section
1-103 and "Code" Methodology, 18 B.C. Indus. & Com. L. Rev. 655, 688-89 (criticizing
cases invoking operative equitable principles without specifying their content).



108 INDIANA LAW REVIEW [Vol. 24:65

1-103 omits even a partial enumeration of the content of such principles.'"

Those principles, properly identified, also must have their source in the

existing law of the relevant jurisdiction.'^* Most important, an application

of the compensation-expectation principle must yield an inequitable result

according to the operative equitable principle identified.

Typical cases of commercial impracticability do not involve misrep-

resentation, fraud, duress, coercion, mistake, or bankruptcy: '^^ each of

which is a subject matter of the general principles referred to in section
1-103.'^^ Cases of commercial impracticability that involve such features

will not satisfy the compensation-expectation principle. Recall that the

compensation-expectation principle in Section IV reads:

(3)" If a party expects (foresees) an event to occur with prob-

ability p; if the party is compensated for that event's occurrence

(to her); and if that party has the opportunity to refuse the

compensation or bargain for a different level of compensation,

either being acceptable alternatives; then if the event occurs, that

party cannot justifiably object. . . .

Instances of misrepresentation, fraud, or mistake are instances in

which the opportunity to refuse compensation or bargain for a different

level is diminished. Hence, that part of condition (3)*' is not satisfied.

Instances of duress, coercion, or bankruptcy are instances in which the

133. Section 1-103 simply provides: "Unless displaced by the particular provisions

of this Act, the principles of law and equity, including the law merchant and the law

relative to capacity to contract, principal and agent, estoppel, fraud, misrepresentation,

duress, coercion, mistake, bankruptcy, or other validating or invalidating cause shall

supplement its provisions." U.C.C. § 1-103 (1987). Cf. Summers, supra note 19, at 913-

19 (providing illustrations of the equitable principles mentioned in § 1-103).

134. See R. Hillmlan, J. McDonnell & S. Nickles, supra note 11, 1.04, at 1-5
to 1-9; Hillman, supra note 129; 1 N.Y. State Law Revision Comm'n, Study of the
Uniform Commercial Code, 1955 Report 167-69 (1955).

135. See cases cited supra notes 11-12, 14. Of course, this is not to say that parties
seeking excuse will not also allege grounds other than impracticability. Often they do.

Allegations of unilateral or mutual mistake sometimes are conjoined with claims of

impracticability. See, e.g., ALCOA v. Essex Group, Inc., 499 F. Supp. 53 (W.D. Pa.
1980); McGinnis v. Cayton, 312 S.E.2d 765 (W. Va. 1984); Famsworth, Brickell & Chawaga,
Relief for Mutual Mistake and Impracticability, 1 J.L. & Com. 1 (1981); Note, Northern
Indiana Public Service Company v. Carbon County Coal Company: Risk Assumption in
Claims of Impossibility, Impracticability and Frustration of Purpose, 50 Omo St. L.J.
163, 176 (1989). Typically, relief based on mistake is denied. See, e.g., Leasco Corp. v.

Taussig, 473 F.2d 777 (2d Cir. 1972); Hanover Petroleum Corp. v. Tenneco Inc., 521

So. 2d 1234, 1240 (La. Ct. App. 1988), cert, denied, 526 So. 2d 800 (La. 1988); Washington

State Hop Producers, Inc. Liquidation Trust v. Goschie Farms, Inc., 51 Wash. App. 484,
754 P.2d 139 (1988), aff'd, 112 Wash. 2d 694, 773 P.2d 70 (1989); McGinnis v. Cayton,

312 S.E.2d 765, 769 (W. Va. 1984).

136. See supra note 130.



19901 COMMERCIAL IMPRACTICABILITY 109

alternatives to refusing compensation or bargaining for a different level

arguably are unacceptable. Thus, the relevant part of condition (3)" is

not satisfied. '^^ In either instance, the compensation-expectation principle

is violated. Because it is violated, the principle is inapplicable. It mandates

no preferred distribution of loss between the contracting parties. Hence,

application of the compensation-expectation principle would not yield

inequitable results. Operative equitable principles would have nothing to

override.

137. Section 2-302 is applied independently of both § 2-615 and the compensation-

expectation principle. See supra text accompanying notes 62-63.