18 JLegal Stud 105

DATE DOWNLOADED: Fri Aug 4 13:38:39 2023
SOURCE: Content Downloaded from HeinOnline
Citations:
Please note: citations are provided as a general guideline. Users should consult their preferred
citation format's style manual for proper citation formatting.
Bluebook 21st ed.

Richard A. Epstein, Beyond Foreseeability: Consequential Damages in the Law of
Contract, 18 J. LEGAL Stud. 105 (1989).
ALWD 7th ed.

Richard A. Epstein, Beyond Foreseeability: Consequential Damages in the Law of
Contract, 18 J. Legal Stud. 105 (1989).
APA 7th ed.

Epstein, R. A. (1989). Beyond foreseeability: consequential damages in the law of
contract. Journal of Legal Studies, 18(1), 105-138.
Chicago 17th ed.

Richard A. Epstein, "Beyond Foreseeability: Consequential Damages in the Law of
Contract," Journal of Legal Studies 18, no. 1 (January 1989): 105-138
McGill Guide 9th ed.

Richard A. Epstein, "Beyond Foreseeability: Consequential Damages in the Law of
Contract" (1989) 18:1 J Legal Stud 105.
AGLC 4th ed.

Richard A. Epstein, 'Beyond Foreseeability: Consequential Damages in the Law of
Contract' (1989) 18(1) Journal of Legal Studies 105
MLA 9th ed.

Epstein, Richard A. "Beyond Foreseeability: Consequential Damages in the Law of
Contract." Journal of Legal Studies, vol. 18, no. 1, January 1989, pp. 105-138.
HeinOnline.
OSCOLA 4th ed.

Richard A. Epstein, 'Beyond Foreseeability: Consequential Damages in the Law of
Contract' (1989) 18 J Legal Stud 105 Please note: citations are
provided as a general guideline. Users should consult their preferred citation
format's style manual for proper citation formatting.
-- Your use of this HeinOnline PDF indicates your acceptance of HeinOnline's Terms and
Conditions of the license agreement available at
https://heinonline.org/HOL/License
-- The search text of this PDF is generated from uncorrected OCR text.
-- To obtain permission to use this article beyond the scope of your license, please use:
Copyright Information
BEYOND FORESEEABILITY: CONSEQUENTIAL
DAMAGES IN THE LAW OF CONTRACT
RICHARD A. EPSTEIN*
I. INTRODUCTION
THE image of the Garden of Eden both before and after the Fall plays a
powerful role in religious and literary theory. It also has its precise, if
humbler, analogue in modem law and economics scholarship. Eden be-
fore the Fall is the complete contingent state contract: the relationship
between parties is so specified that nothing that has not been anticipated
can occur during the life of the contract. Each possible breach is known in
advance, as are the elements of the appropriate remedy. In such a world,
a common-law judge need only consult the sacred text of the contract in
order to resolve all doubts about the rights and duties of the parties.
The Fall from Eden is the world we live in, where contracts never cover
all the contingencies that might arise. This world necessarily arises
whenever the cost of contracting is positive, for now it no longer pays to
draft contracts to envision what will happen in all possible states of the
world, even if such were technically possible. Now contract interpreta-
tion becomes a second-best proposition that addresses the uncertainty
and ambiguity that explicit provisions could have resolved but did not.
Redemption after the Fall is only partial, and lies in the sound rules of
contract construction. Of necessity, the possible techniques are divided
into two basic types. 1 First, it is possible to read those portions of the
* James Parker Hall Distinguished Service Professor of Law, University of Chicago.

Earlier versions of this article were delivered at the Conference on Liability in Law and
Morals, held at Bowling Green State University, April 15-17, 1988, and at the Brooklyn
Law School on April 20, 1988. I should like to thank Richard Craswell and Alan Schwartz
for their most insightful comments on an earlier draft. Sean Smith and James Fiero provided
their helpful research assistance.
1 For a further defense of this consumer sovereignty model, see Alan Schwartz, Proposals
for Products Liability Reform: A Theoretical Synthesis, 97 Yale L. J. 353, 357 (1988).
[Journalof Legal Studies, vol. XVIII (January 1989)]
© 1989 by The University of Chicago. All rights reserved. 0047-2530/89/1801-0002$01.50
THE JOURNAL OF LEGAL STUDIES
contract that are explicit in order thereafter to decide how the contested
issue should be resolved in light of what the parties themselves have
decided. Second, it is possible to resort to some general theory of con-
tracting behavior that tries to reconstruct the "rational" bargain that self-
interested parties would have made in order, ex ante, to maximize their
joint gains from the agreement. This second technique requires courts to
understand not only the circumstances of any immediate dispute but also
the larger business and institutional context in which the contract was
formed. Invariably, these two methods of contract interpretation tend to
reinforce each other in practice. What parties have, in fact, agreed to is
strong evidence of what rational parties would have agreed to: indeed,
one function of bargaining theory is to understand why certain contract
provisions have been included when their relevant terms are clear and
unambiguous. Similarly, what rational parties would have agreed to is, in
turn, strong evidence of what these parties did, in fact, agree to where
there is silence or ambiguity. There is, accordingly, a complete congru-
ence between the "efficient" identification of the proper contract terms
and honoring what the parties did, or would have agreed to do, under
contract.2
The art of contract interpretation, then, requires the application of a
general theory of bargaining to particular contractual provisions. Accord-
ingly, one critical question is, What is the mix between the discrete in-
quiry into the particular facts and the reliance on the general theory of
contracting? In this context, the common law (of both England and the
United States) has exhibited something of an uneasy dualism. With regard
to the primary obligations of the parties-to buy, to ship, to work-the
tendency has been to allow the parties themselves to specify the subject
matter of the agreement, including any price term. This inquiry is case
specific. After breach, the opposite tendency has emerged for the selec-
tion of remedy. While the judicial practice is far from uniform, courts and
commentators often treat the required damage rules as though they were
generated by some normative theory external to the contract itself. From
this premise, the implicit understanding has grown up that damages are
resolved more by "rules of law" and less by default rules of construction.
This attitude is most evident in typical statements about the function
and purpose of damage awards for breach of contract. Here discussions of
the damage rules often begin from a social norm that the innocent party
should be made whole after breach. The question whether it is possible to
contract out of the rules is suppressed, so that it is often unclear how the
2 See Richard Craswell, Contract Remedies, Renegotiation, and the Theory of Efficient
Breach, 61 So. Calif. L. Rev. 630, 633 (1988).
CONSEQUENTIAL DAMAGES
issue of contracting out would be resolved if raised expressly. Samuel

Williston articulated the classical position as follows. "In fixing the
amount of these damages, the general purpose of the law is, and should
be, to give compensation: that is to put the plaintiff in as good a position
as he would have been in had the defendant kept his contract."- 3 The
statement takes a stance external to the will of the parties, for it is the
purpose the law "should" have that seems dispositive. Countless cases
take much the same view. Thus it has been rendered axiomatic that "the
function of the award of damages for a breach of contract is to put the
plaintiff in the same position he would have been in had there been no
breach." 4 Charles Fried has argued for the normative power of this posi-
tion on fairness grounds: "If I make a promise to you, I should do as I
promise; and if I fail to keep my promise, it is fair that I should be made to
hand over the equivalent of the promised performance. In contract doc-
trine this proposition appears as the expectation measure of damages for
breach."'
The dominance of the expectation interest is sometimes challenged, but
usually on the wrong ground. Lon Fuller, in his classic study, "The
Reliance Interest in Contract Damages," ' 6 takes it as "obvious" that the
expectation measure of damage is entitled to less protection than the re-
liance interest-that is, compensation that places the plaintiff in the same
position he would have enjoyed had there been no contract. In addition,
he further claims that the reliance interest is entitled to less protection
than the restitution interest, by which the defendant is made to surrender
the gains obtained under the contract. "It is as a matter of fact no easy
thing to explain why the normal rule of contract recovery should be that ' 7
which measures damages by the value of the promised performance.
Fuller's criticism rests on the supposed superiority of one legal norm over
another. His position does not challenge the (implicit) assumption that the
3 Samuel Williston, A Treatise on the Law of Contracts, § 1338 (1920). The statement is
picked up in much of the modem literature. See, for example, Craswell, supra note 2, at 636.
4 See, for example, Spang Industries, Inc. v. Aetna Casualty & Surety Corp., 512 F.2d
365 (2d Cir. 1975). See also Hawkins v. McGee, 84 N.H. 114, 117, 146 A. 641, 643 (1929),
quoting the Williston passage, supra note 3.
5 Charles Fried, Contract as Promise 17 (1981).
6 L. L. Fuller & William R. Perdue, Jr., The Reliance Interest in Contract Damages, 46
Yale L. J. 52, 56 (1936).
7 Id. at 57. Fuller's doubt rests on the alleged circularity of the expectation principle. The
contract has value only if the law decides to enforce it, so the contract cannot be enforced
only because it has value. Id. at 59-60. The best answer to this point is to note that, from a
system perspective, the choice of legal rule is made to maximize joint benefits of the parties
so that the evaluation of the individual case rests on the application of some prior accepted
social norm.
choice of proper remedial rules is largely a judicial function. Yet as a

matter of basic contract theory, there is, in principle, no abstract way to
resolve this battle of intuitions. Damage rules are no different from any
other terms of a contract. They should be understood solely as default
provisions subject to variation by contract. The operative rules should be
chosen by the parties for their own purposes, not by the law for its
purposes.
The received wisdom on contract damages is flawed, then, both in its
methodology and in its concomitant elevation of the expectation measure
of damages. Accordingly, the first section argues that in contract the
damages payable should be set by agreement and not by any abstract
principles of justice, however pleasing and symmetrical they might ap-
pear to judges. On matters of both liability and remedy, the principle of
freedom of contract should control so that all legal damage rules should be
understood routinely as default provisions in lieu of express provisions,
not as general rules of fairness or natural justice entitled to independent
respect.
The second section then gives reasons why the range of damage solu-
tions to which parties might agree is more varied than the canonical
treatment of the subject suggests. The familiar expectation, reliance, and
restitution all share a certain intellectual elegance, scope, generality, and
versatility that a liquidated damage provision calling for a $100 payment
cannot hope to match. 8 Together they seem to exhaust the logical pos-
sibilities. Nonetheless, in fact, these three measures do not begin to cap-
ture the full array of remedial provisions that are expressly incorporated
into standard form contracts. The disjunction between observed practice
and general theory is especially evident in the rules that govern conse-
quential damages, that is, damages for lost profits, or damages to property
or person. In this context, the expectation measure seems to call for full
compensation, subject to such tort-like limitations as speculation and
foresight, and mitigation of damages. Yet most contractual provisions
seem to provide otherwise. This pattern holds in three separate areas:
warranties for the sale of goods, liability for employees' accidents, and
damages for delayed shipment of goods. The second section discusses
reasons why the rule of expectation damages, good in some contexts,
need not carry over to consequential damages.
8 Charles J. Goetz & Robert E. Scott, Liquidated Damages, Penalties, and the Just
Compensation Principle: Some Notes on an Enforcement Model and a Theory of Efficient
Breach, 77 Colum. L. Rev. 554 (1977); Samuel A. Rea, Jr., Efficiency Implications of
Penalties and Liquidated Damages, 13 J. Legal Stud. 147 (1984).
The third section of the article then uses the insights gathered from the
examination of express damage provisions to evaluate the general default
rules of contract damage as they derive from Hadley v. Baxendale.9 To
the extent that Hadley leads courts to follow the classical "tacit" assump-
tion-of-risk test, it provides an accurate mirror for the express contractual
provisions observed in contracts. In many instances, however, the courts
act as though the same damage rules applicable to stranger cases in torts
should carry over without modification to the law of contract. At this
point, the emphasis shifts to some linguistic variation of the "foreseeabil-
ity"-of-damage test that leads to incorrect default rules that call for a
systematic overexpansion of contract damages, relative to the voluntary
norm. This mistaken chain of reasoning can be usefully-but not exclu-
sively-illustrated by cases where carriers make late delivery of goods. A
brief conclusion then assesses the dangers of the wrong approach toward
contract damages.
II. FREEDOM OF CONTRACT AND THE MEASURE OF DAMAGES
The provisions of a standard contract are not an arbitrary assemblage of

words and phrases with no business and social function. Anyone who
drafts a complex contract knows that provisions on delivery, payment,
product quality, warranties, and a thousand other terms matter. Provi-
sions directed toward remedies when things go wrong are every bit as
important as those terms that are directed toward routine performance.
Even if contract breach is, as a matter of relative frequency, the exception
instead of the rule, breach can be a very expensive exception. For parties
who draft contracts that must work in thousands or millions of standard
transactions, the wording of every provision is of critical importance. It is
just not credible to assume that commercial parties draft toward perfor-
mance but ignore breach in ways that justify more vigorous judicial inter-
vention on remedial questions. The extensive negotiation over explicit
limitations on consequential damages tells a different story. Let the terms
on damages or other remedy be too restrictive, and it will be difficult to
obtain business. Let them be too generous, and the business that is ob-
tained will be a losing proposition. The drafting of contract terms is a
constant search for getting the business that is wanted, and for wanting
the business that is gotten. Fine-tuning in drafting is the order of the day.
An important object lesson emerges from this discussion. The impor-
tance of the written contract itself should not be forgotten in the constant
9 9 Ex. 341, 156 Eng. Rep. 145 (1854).

quest for contract theories of greater abstraction. If the contract measure

of damages functions as a default provision, then the methodology for
handling these cases should be clear: first, look for express provisions on
damages, and only as a last resort adopt some general principle of law.
While the parties may not have entered into a complete contingent-state
contract, their fall from grace need not be complete if they have taken into
account the contingencies that did, in fact, occur.
Consistent with the heavy bias toward state control in this area, the
cases have often done the opposite.'° One influential illustration of the
wrong methodology is Kerr S.S. Co. v. Radio Corporationof America."
The defendant telegraph company failed to transmit a message delivered
in code to its Manila office by the plaintiff steamship company, in conse-
quence of which the plaintiff sought the recovery of some $6,675 for "the
freight that would have been earned if the message had been carried." 2
Judge Cardozo, speaking for a unanimous court, limited the plaintiff's
recovery to a refund of the return of the tolls, or some $26.78. His case
discussion was couched in terms of the general damage rules found in
Hadley v. Baxendale,' 3 with specific attention to the question whether the
defendant had sufficient knowledge from the cipher of the nature of the
plaintiff's business, which would make the recovery of lost freight appro-
priate. Cardozo denied that the defendant could get the needed informa-
tion from the unintelligible cipher but, on receipt of the message, "might
infer that the message had relation to business of some sort. Beyond that
it could infer nothing." ' 4 The message could have related to the carriage
of cargo, the employment of an agent, or to nameless transactions "as
divergent as the poles." ' 5 There was not enough information to impress
on the defendant the nature of the risk. Cardozo did not say what outcome
would have been correct had the telegram been in plain English, but it is
possible that he might have seen fit to accept some higher award, perhaps
the full $6,675. One cannot say.
To generalize his conclusion, however, Cardozo then launched into a
more abstract discussion about the distinction between general and spe-
cial damages. He noted that these categories were not fixed by nature but
'0 See also, Charles J. Goetz & Robert E. Scott, The Limits of Expanded Choice: An
Analysis of the Interactions between Express and Implied Contract Terms, 73 Calif. 261
(1985), chiding the courts for their unwillingness to accept the benefits of contracting out.
" 245 N.Y. 284, 157 N.E. 140 (1927).
12 245 N.Y. at 287, 157 N.E., at 141.
13 See discussion in Section IVA infra.
14 245 N.Y. 288, 157 N.E., at 141.
15 id.
depended heavily on the identity of the parties to the transaction. Mat-

ters, then, were "not absolute but relative."1 6 As between buyer and
seller, general damages normally include the difference between contract
and market price, but as between the sender of a telegram and a carrier,
even with reference to that same sale, the general damages on default are
"the cost of carriage and no more." 17 The functional reason for limit-
ing damages was understood by Cardozo. There was a risk of cross-
subsidization between senders of messages, where the telegraph company
was unable to charge different premia to different customers to reflect the
differences in anticipated consequential damages.18
What is so striking is that nowhere in his analysis does Cardozo once
rely on the explicit provision in this contract that, in fact, generally limited
damages to a return of the charges collected. 19 Indeed, the last clause of
the agreement identified the difference between ordinary messages and
ciphers, but it did not use the distinction in order to prevent the recovery
of lost profits for coded messages while allowing them otherwise, as the
logic of Cardozo's opinion seems to suggest. The last clause of the con-
tract provides explicitly "that this company shall not be liable in any case
16 id.
17 Id. at 289, 157 N.E. at 141.
18"The loss of a cipher message to load a vessel in the Philippines may mean to one the
loss of freight, to another an idle factory, to another a frustrated bargain for the sale or
leasing of cargo" (id.). To be more accurate, the true problem lies not in the different types
of loss but in their different magnitudes, which could vary as much within these classes as
across them. But even with this emendation, it is clear that Cardozo has identified a key
element in restricting damages. This is one of the reasons why the privity limitation in
products liability cases is important as well. It prevents careless plaintiffs from obtaining
subsidies from careful ones. See Richard A. Epstein, Products Liability as an Insurance
Market, 14 J. Legal Stud. 645 (1985). And Cardozo had strong intuitions about the point in
other cases. See, for example, Moch Co. v. Rensselaer Water Co., 247 N.Y. 160, 159 N.E.
896 (1928), which was defended on insurance grounds in Charles Gregory, Gratuitous
Undertakings and the Duty of Care, I DePaul L. Rev. 30 (1951).
'9 245 N.Y. at 293-94, 157 N.E. at 143. The clause reads as follows. "It is agreed between
the sender of the message on the face hereof and this company, that said company shall not
be liable for mistakes or delays in transmission or delivery, nor for nondelivery to the next
connecting telegraph company or to the addressee, of any unrepeated message, beyond the
amount of that portion of the tolls which shall accrue to this company; and that this company
shall not be liable for mistakes or delays in the transmission or delivery, nor for delay or
nondelivery to the next connecting telegraph company, or to the addressee, of any repeated
message beyond the usual tolls and extra sum received by this company from the sender for
transmitting and repeating such message; and that this company shall not be liable in any
case for delays arising from interruption in the working of its system, nor for errors in cipher
or obscure messages."
Note, too, that his cavalier dismissal of the agreement has been carried over into the
casebook tradition. See, for example, Friedrich Kessler, Grant Gilmore, & Anthony Kron-
man, Contracts: Cases and Materials 1152 (3d ed. 1986). The opinion is so edited to eliminate
both Cardozo's brief discussion of the contract language and the text of the actual provision.
for . ..errors in cipher or obscure messages." The contract therefore

appears to place the risk of mistransmission of these messages on the
sender, doubtless in recognition of the greater possibility for error. Yet
given that the message was not sent at all, the last clause, which governs
only errors in cipher or obscure messages, does not seem to relieve the
company from all liability for a coded message that is not sent at all;
rather, cases of nondelivery of coded messages are treated like nondeliv-
ery of uncoded ones. The problem of ciphers was doubtless a common
one because of the need to preserve secrecy for important business mes-
sages. The contract thus provides for an authoritative and sophisticated
answer to the question that Cardozo chose not face.2 °
In a sense, Cardozo reached the right result in the case after all, given
that his abstract solution comports with the standard voluntary contract
solution-in this instance at least, restitution damages. The result in Kerr
thus minimizes the cost of recontracting that the parties would otherwise
have to bear. But the decision rests in very large measure on a clever
interpretation of Hadley instead of on an understanding of the relevant
business practices. An adequate theory of contract analysis should begin
with transactions, not with doctrine. It is to some transactions with ex-
plicit contract provisions that the doctrine now turns.
III. DAMAGE SOLUTIONS
In order to understand how damages are selected, it is critical to aban-

don the ex post approach to the subject and to ask the single question,
Which default rule promises, under the known circumstances of the trans-
action, to maximize the joint gain of the parties? In order to make that
judgment, it is critical not only to look at the contracting process after the
breach has occurred but also at the full range of prospects and perils that
the contract itself addresses at formation. When all these circumstances
are taken into account, the expectation measure of damage makes per-
fectly good economic sense in some contexts but not in others.21 Where
the damage measures chosen by the law tend to meet the needs of the
parties in standard transactions, there will be little contracting out. But
where there is a tension between the rules set by the legal system and the
economic interests of the parties, contracting out will take place, at least
20 Note the plaintiff argued that "this provision (that is, the full clause) is inapplicable
where the telegraph company has omitted to transmit the telegram at all, and moreover that
it is unreasonable and oppressive in the limitation affixed to liability where the message is
repeated. We leave these questions open." 245 N.Y. at 294, 157 N.E. at 143. The contract
seems to resolve them in favor of the refund.
2 For the relevant formulas, see E. Allan Farnsworth, Contracts § 12.9-12.11 (1982).
when the anticipated gains exceed the costs of negotiating and drafting the
voluntary provisions. In this section, I consider a number of contexts to
which the damage rules apply and contrast, where applicable, the explicit
provisions observed in standard contracts with the presumed measures
created by operation of law.22
A. Seller's Remedies: Nonpayment of Price and Nonacceptance

of Goods
Suppose the seller has delivered, and the buyer has accepted, all the
promised conforming goods. The expectation measure calls for the buyer
to pay the price, as the Uniform Commercial Code (U.C.C.) now re-
quires.23 This result seems to make perfectly good sense for the parties,
even from the ex ante perspective. Any lesser measure of damage would
allow the buyer to renege on the contract and, in effect, to purchase the
goods for less than what had been promised while pocketing the dif-
ference between the original contract price and the lesser award of con-
tract damages. The system of credit would be undone by opportunistic
behavior.
Similarly, where a buyer does not accept the goods or repudiates the
contract before delivery, the expectation measure of damage also works
well. Here the seller's measure of damage "is the difference between the
market price at the time and place for tender and the unpaid contract price
together with any incidental damage provided in this Article, but less
expenses saved in consequence of the buyer's breach,"-2 4 subject to fur-
ther qualifications where the contract/market differential does not "put
the seller in as good a position as performance would have done," in
which case lost profits (including allowance for reasonable overhead) are
allowed.25
Whatever the differences between the two measures of damages used
for repudiation or nonacceptance, four features remain constant. First,
the seller is to receive only money, so any consequential losses attribut-
able to the want of money alone can generally be avoided by resort to the
capital markets. It follows that the price term is the upper bound on
damages, so the buyer can never complain that he is required to bear
losses far in excess of the amounts he has agreed to pay. The prospect of
22 For general discussions of contract remedies, see William Bishop, The Choice of
Remedy for Breach of Contract, 14 J. Legal Stud. 299 (1985); Lewis A. Kornhauser, An
Introduction to the Economic Analysis of Contract Remedies, 57 Colo. L. Rev. 683 (1986).
23 See U.C.C. § 2-709 (1).
24 Id. at § 2-708 (1).
25 Id. at § 2-708 (2).
mitigation eliminates the possibility of enormous consequential damages.

Second, the seller is required to take into account the benefits received
from breach, that is, the value of the goods in his hands, as a setoff against
any liability. Third, the insistence on the contract/market differential im-
poses on the seller perfect incentive to mitigate damages, for he assumes
the risk, good or bad, in any movement in prices after breach. He will thus
have, after breach, the incentive to act as a single owner intent on max-
imizing his own gain. Fourth, the administrative or litigation costs neces-
sary to determine contract-market differentials are relatively low, given
the ease of getting the relevant data and the relatively small size of the
damages. Calculations of this sort can be done routinely, say, in ordinary
exchanges or other active markets. In principle, these damage rules
should be subject to variation by contract, but as the fit between the
default rules and the optimal contract rules seems very close, very little
recontracting should be expected.
B. ConsequentialDamages
The expectation measure of damages for nonpayment or nonaccept-
ance of goods offers a useful foil for the cases of consequential damages of
primary interest here. The rules governing seller's remedies illustrate the
importance of taking into account the gains that the seller retains in the
event of breach. They also point to several key variables relevant to
the overall desirability of any given rule. At a minimum, it is necessary to
take into account (a) the incentives for the defendant to perform, (b) the
incentives for the plaintiff to perform and to mitigate the losses conse-
quent upon breach, and (c) the anticipated costs of resolving the differ-
ences between the parties through either litigation or settlement.
1. Buyer's Remedy for Seller's Breach of Promise. The difficulties
with the expectation measure of damages become more acute when we
shift from the seller's remedy for buyer's breach to the buyer's remedy
for seller's breach. The question of consequential damages can arise when
goods are not delivered on time, or when they misfunction once deliv-
ered. Now the action against the seller is not obviously limited to an
amount less than the price term of the contract. There is no ability to
prevent losses by obtaining money-a fungible good from capital mar-
kets. The cost of putting the buyer in exactly the same position he would
have been in had the contract been performed properly could be enor-
mous, whether we deal with lost profits, personal injury, or property
damage. To make matters more difficult, the consequential damages typi-
cally do not arise solely because of the seller's breach but usually depend
on a combination of factors, some of which are in control of the buyer.
With consequential damages, the superiority of the expectation mea-
sure of damages is far from self-evident. Whatever the abstract law of

contract, many ordinary sales contracts contain warranty provisions that
specify both the obligation of the supplier and the remedy in the event of
breach. The provisions widely adopted across different industries stipu-
late for liquidated damages or call for the repair or replacement of
damaged goods at the option of the seller.2 6 While the full expectation
measure of damage may require the seller to bear the cost of delivery to
the firm, contracts often leave that loss on the buyer, who likewise may
receive no compensation for any interim loss of use. Often the warranty is
voided where the product was damaged by excessive or impermissible
use by the buyer. 27 Most notably, all liability for personal injury or prop-
erty damages is excluded. Without question, these warranties call for a
risk-sharing arrangement, which leaves the buyer worse off when the
seller honors the warranty in full than he, the buyer, would have been if
the product had worked perfectly in the first place. Warranty restrictions
similar to these led the New Jersey Court in Henningsen v. Bloomfield
Motors, Inc. 28 to invalidate these warranties as limitations on recovery for
26 See, for example, the survey of warranty provisions undertaken by Bogert & Fink,
Business Practices Regarding Warranties in the Sale of Goods, 25 Ill. L. Rev. 400, 408-9
(1929).
Thus the standard tire warranty reads as follows. "Every pneumatic tire of our manufac-
ture bearing our name and serial number is warranted by us against defects in material and
workmanship during the life of the tire to the extent that if any tire fails because of such
defect, we will either repair the tire or make reasonable allowance on the purchase of a new
tire."
The standard automobile warranty provides:
WARRANT each new motor vehicle manufactured by us, whether passenger car or commercial
vehicle, to be free from defects in material and workmanship under normal use and service,
our obligation under this warranty being limited to making good at our factory any part or
parts thereof which shall, within ninety (90) days after delivery of such vehicle to the original
purchaser, be returned to us with transportation charges prepaid, and which our examina-
tion shall disclose to our satisfaction to have been thus defective; this warranty being ex-
pressly in lieu of all other warranties expressed or implied and of all other obligations or
liabilities on our part, and we neither assume nor authorize any other person to assume for
us any other liability in connection with the sale of our vehicles.
This warranty shall not apply to any vehicle which shall have been repaired or altered
outside of our factory in any way so as, in our judgment, to affect its stability, or reliability,
nor which has been subject to misuse, negligence or accident, nor to any commercial vehicle
made by us which shall have been operated at a speed exceeding the factory rated speed, or
loaded beyond the factory rated load capacity.
We make no warranty whatever in respect to tires, rims, ignition apparatus, horns or other
signalling devices, starting devices, generators, batteries, speedometers or other trade ac-
cessories inasmuch as they are usually warranted separately by their respective manufactur-
ers.
For a discussion of the functions of these warranties, see George L. Priest, A Theory of the
Consumer Product Warranty, 90 Yale L. J. 1297 (1981).
27 Priest, supra note 26.
28 32 N.J. 358, 161 A.2d 69 (1960).
personal injury and to usher in the modern age of product liability law by
imposing full tort liability for consequential damages.
Against this backdrop of express contractual provisions, there is ample
reason to doubt that the expectation measure of damage of the classical
common law maximizes the joint gains of the parties ex ante. If it did, we
should expect to observe it frequently in practice, which is decidedly not
the case. The failure to observe this standard in practice cannot easily be
attributed to the systematic ignorance of buyers and sellers in all product
markets, for someone must have the incentive to break the logjam if
making the plaintiff whole on breach is the ideal contract measure of
damages. The better approach, therefore, is to ask why it is that informed
parties might not choose to use this damage measure. A closer inspection
of the expectation measure of damages reveals some costs of its ap-
plication.
First, under the orthodox view, the basic rule provides for full conse-
quential damages, but (like the tort rules of contributory negligence it
parallels) 29 it then allows the defendant an affirmative defense, where the
plaintiff is in breach of some condition precedent. The two halves of this
rule thus raise two high-stakes issues (plaintiff's and defendant's breach)
on which enormous liabilities can turn. As a general matter, the parties'
investment in litigation increases with both the uncertainty of the out-
come and the size of the stakes.30 Ex ante, the parties wish to avoid this
cost, as it represents a deadweight loss to both sides. In addition, full
consequential damages raise the real risk that the plaintiff, while in the
better position to avoid the loss, will, in fact, not take the right steps to do
so. Any money that is spent on further loss reduction is his own, while the
money that is saved is the defendant's. The temptation to maximize pri-
vate gain results in the systematic externalization of losses: why should I
spend my money to reduce his damages? The law recognizes this and
imposes a duty of mitigation of damages to counter this all too-tenacious
human tendency. 3' But that duty is a very imprecise tool to use against so
persistent a business practice, for defendant's monitoring of plaintiff con-
29 See discussion of Evra Corp. v. Swiss Bank Corp., 673 F.2d 951 (7th Cir. 1982), at IVC
infra.
30 See, for example, George L. Priest & Benjamin Klein, The Selection of Disputes for
Litigation, 13 J. Legal Stud. 1 (1984); Donald Wittman, Dispute Resolution, Bargaining, and
the Selection of Cases for Trial: A Study of the Generation of Biased and Unbiased Data, 17
J. Legal Stud. 313 (1988).
31 See Susan Rose-Ackerman, Dikes, Dams and Vicious Hogs: Entitlement and
Efficiency in Tort Law, in this issue, for the importance of mitigation rules in tort law. Her
article highlights the difficulty of an explicit doctrinal incorporation of mitigation rules,
which are largely unavoidable in the tort context.
duct, whether under a misuse or a mitigation doctrine, is both costly and

error prone. The expectation rules with affirmative defenses may not offer
the best prospect of minimizing the total costs of contractual failure.
The weaknesses of the expectation damage system are thrown into high
relief by comparing it with an alternative regime that gives fixed damages
without making any provision for separate affirmative defenses. This sys-
tem of damages could be superior for both parties because of the way in
which it reduces the joint costs of litigation while preserving the incen-
tives on both sides to perform as agreed. The plaintiff whose level of
recovery is fixed in advance has a powerful incentive to mitigate his loss.
Any rule of fixed damages, independent of subsequent events, makes him
face the identical incentives of a single owner of all relevant inputs. Every
dollar he spends in mitigation now results in a dollar's saving for himself
from the reduction in consequential damages. He will therefore behave
exactly as he would if he had been the sole and original author of his own
harm. While one should not expect perfection in the delicate task of
mitigation-the innocent party may be a complex firm with agency-cost
problems of its own-clearly there is nothing that the legal rule could do
to improve performance once it has eliminated this potential conflict of
interest posed by any ad hoc mitigation rule. There is no need to build
mitigation doctrines explicitly into legal rules in order to create the cor-
rect incentives to mitigate. Any fixed lump-sum damage award has just
that effect, regardless of the level at which it is set.
The amount of the fixed damages is critical, however, in influencing the
frequency of breach by the plaintiff. If damages were exceedingly high,
the level of precautions taken would be low, given the fixed rate of return
that the plaintiff could expect to receive. Similarly, the risk of fraudulent
or dubious claims of seller's breach would increase as well. Where the
seller is suing for the buyer's failure to accept goods or to pay for goods
accepted, the question of seller's breach is relatively unimportant, so that
all incentives can be directed toward the conduct of the single wrongdoer.
Matters are far more complicated with lost profits, personal injury, and
property damages, for here the possibility that both sides will be in breach
is far more likely. Setting damages below actual loss is far more important
in this context precisely because there is frequent need to restrain abuses
by defendants and plaintiffs simultaneously.
Any effort to constrain misbehavior by the buyer invites misbehavior
by the seller. But the critical question is the rate of substitution between
buyer and seller incentives for breach. Here there are at least two reasons
to think that the control of buyer's misconduct will often be the more
important issue, pointing to a lower level of fixed damages. First, reputa-
tional constraints tend to operate far more powerfully upon institutional
defendants than they do upon individual buyers. Major failures are per-
ceived by the market, resulting in a loss of future sales that can best be
avoided by maintaining product quality.32
Second, it is important to note that even low damage awards can exert a
considerable incentive on a defendant to perform his contracts. Consider
the standard repair and replacement warranties set out above.3 3 Here
defendant must lose from any individual transaction if required, say, to
refund the consideration received or to make repairs while still having to
bear other costs under the contract. In principle, the defendant will have
no incentive to supply defective products so long as his breach costs him
more than he gains. Even a low level of nonperformance is sufficient to
remove all the profits that the defendant derives from other contracts that
are performed successfully. 34 The combination of reputational and finan-
cial losses helps keep the defendants in check.
The last point concerns administrative costs. Fixing these damages
reduces the associated uncertainty and, therefore, the costs of administer-
ing the remedial provisions. Nonetheless, if there is any serious question
whether the defendant was in breach, the costs of litigation will increase
as the size of the damage award increases, even if damages are fixed.
These administrative costs will become far larger if the level of damages is
both large and uncertain and subject to reduction for plaintiff's miscon-
duct and failure to mitigate. All in all, the optimal contracting strategy
does not appear to call for the high consequential damages, subject to
defense rules, that courts have tended to adopt. Less clearly, within the
class of fixed damage awards, there is reason to expect these damages to
be kept relatively limited, which is what the express contracts have typi-
cally provided.
2. Workers' Compensation. The same general pattern of limited con-
tract damages was often observed in industrial accident cases in the era
prior to mandatory workers' compensation. 35 In this context, the expecta-
32 One finding from the event studies is that the stock price of public companies falls far
more than the estimated future liabilities that they have to bear. Bad news is bad for
business, so the reputational effects are quickly reflected in the financial markets. See
generally, Andrew Chalk, Market Forces and Aircraft Safety: The Case of the DC-10, 24
Econ. Inq. 43, 46 (1986).
33 See note 26 supra.
34 See discussion of Federal Express Contract, at IIIB3 infra.
3 For discussion, see Richard A. Epstein, The Historical Origins and Economic Struc-
ture of Workers' Compensation Law, 16 Ga. L. Rev. 775 (1982); Price V. Fishback, Liability
Rules and Accident Prevention in the Workplace: Empirical Evidence from the Early Twen-
tieth Century, 16 J. Legal Stud. 305 (1987). Fishback notes that there is no reason to believe
tion measure of damages suggests that compensation should be provided

for lost income, for pain and suffering, and for medical expenses. It also
invites the use of contributory negligence and some forms of assumption
of risk as affirmative defenses to control plaintiff's misconduct. 36 Yet the
voluntary systems often exhibited very different characteristics. Under
the basic structure, the defendant paid less than full damages, but for all
harms "arising out of and in the course of employment," not only for
those caused by the defendant's negligence. One consequence of the
voluntary compensation formula is to preserve incentives on the plaintiff
to take care, without having to incorporate an explicit defense of contrib-
utory negligence. The plaintiff who does not recover full tort damages
always has some incentive to take precautions against loss.
Consistent with this framework, another benefit of the compensation
formula is to exert a selection effect so that persons with extra severe
injury costs avoid physical labor with a high risk of injury. Fledgling piano
players will not work crushing machines in coal mines if they cannot
recover a dime for their lost musical opportunities when they smash a
finger. So long as the defendant pays damages that are in most cases
greater than his own cost of compliance with safety norms, then the
compensation rules chosen make each side bear part of the risk so as to
ensure, roughly speaking, that each side has some incentive to minimize
it. The damage provisions under these systems are geared to specially
negotiated formulas and schedules that make no effort to duplicate the
numbers that might come from some general expectation, reliance, or
restitution measure. The damages provided, while below those of the tort
that a single compensation system will be ideal for all firms within a single industry. Thus,
where coal is produced by individual miners working alone, a tort system with a large
assumption-of-risk defense may be appropriate, while a workers' compensation scheme
might be more desirable where mining is a team operation (id. at 314, 324).
36 Alan Schwartz has offered a qualified defense of the similar rules in product liability
cases, as best meeting the insurance and compensation objectives of the parties. He con-
cludes first that the full compensation ideal is generally compromised in cases of pain and
suffering but works reasonably well without wage losses and medical expenses. The defense
of contributory negligence then controls against plaintiff misconduct. Yet even with this
system the two goals are in necessary tension, for whenever the defenses are held, the
insurance goals cannot be met. Schwartz, supra note 1, at 368. His general discussion rightly
treats freedom of contract as the preferred regime, but then-incorrectly, in my view-
limits the choice of default regimes to negligence and strict liability for defective products,
with or without contributory negligence.
Schwartz relies in part on Fishback's study, supra note 35, to demonstrate the inferiority
of workers' compensation solutions. Yet Fishback studied mandatory and not voluntary
systems. If the mandatory systems set the level of compensation too high, then workers'
compensation may turn out to be inefficient, but the defect could dwell in the price term, not
in the basic structure. Schwartz, supra note 1, at 395 n.82.
system, are often substantial, especially where the employer is in an

effective position to avoid losses. But the need to handle both plaintiff's
misconduct and administrative costs still exerts its influence. It is only the
modern proposals for comprehensive public insurance that treat full com-
pensation as an ideal, 37 and these programs have been defective precisely
because they ignore the moral hazard question associated with plaintiff's
precautions. 38
3. Common Carriers. A similar pattern of argument helps explain the
contractual provisions governing the liability of common carriers for lost
profits attributable to the delayed shipment of goods. The expectation
measure of damage calls forth the question of extensive litigation over
both plaintiff's and defendant's conduct. In contrast, the use of a fixed
tariff regardless of circumstances helps advance the joint interests of the
parties. It gives the plaintiff the incentive to mitigate losses, while impos-
ing on the defendant some financial incentives to perform as promised,
which are doubtless augmented by the fear of reputational loss. The small
level of damages awarded reduces the costs of both litigation and set-
tlement.
In principle, it is possible that expectation-measure damages could
dominate over the fixed tariff, but there is evidence from the structure of
real contracts that suggests the opposite. The modern Federal Express
standard form, for example, calls for a return of the contract price when
the package is not delivered-a restitution form of damages. 39 But it also
contains a liquidated damage provision for loss or damage to goods that
limits recovery to $100 or actual losses, whichever is lower-even when
caused by the company's negligence-with opportunity to purchase addi-
tional coverage for stated rates. 40 The company then completes its list of
31 See, for example, Jane Stapleton, Disease and the Compensation Debate (1986); Paul
C. Weiler, Protecting the Worker from Disability: Challenges for the Eighties 75-76 (1983).
31 See M. Trebilcock, The Role of Conduct Variables in the Design of No-Fault Compen-
sation Systems 23, criticizing Weiler, and 33, criticizing Stapleton (unpublished manuscript,
J. Legal Stud. office).
31 "Our liability for delays not caused by your negligence shall be limited to a refund of
your delivery charges" (Federal Express Terms and Conditions).
40 "Except as otherwise provided in our tariff and our conditions of carriage, our liability
for loss or damage (even if caused by our negligence) is limited to the amount of actual loss
or U.S. $100, whichever is less, unless you fill in a higher "declared value" and pay an
additional charge. We do not carry cargo liability insurance, but you may pay 30¢ for each
additional $100 of declared value. If you declare a higher value and pay the additional
charge, our liability will be the lesser of your declared value or the actual value of your
package" (Federal Express Terms and Conditions). Nor does Federal Express conceal its
limited warranties. Its money-back guarantee is featured on the outside of its Overnight
Letter along with a similar money-back guarantee that it can "tell you the exact status of
your package within 30 minutes of your call" (Federal Express Overnight Letter). There is
no implied common-law warranty covering this benefit.
limitations with boldface disclaimers of all liability for incidental, conse-

quential, or special 4damages,
1
"whether or not we knew that damages
might be incurred."-
The form contains all the explicit limitations on damages that high-
minded judges and academics often deplore, but it is wholly admirable in
both its conception and execution. The difference in damages between
delay and loss of packages is about eight- or tenfold (if the $100 maximum
is awarded), which bears at least rough correlation to the differential
levels of losses. Moreover, while that $100 damage figure for loss or
damage may look puny, in fact it conveys very powerful information to
the consumer about the firm's reliability in making deliveries. If the net
profit per standard transaction is, say, 5 percent, then for a $10 transac-
tion, the gain is about fifty cents. A damage award of $100 thus translates
into a situation in which a failure rate of half of 1 percent strips the firm of
all its profits, even if there were (a) no reputational losses to the firm and
(b) no additional costs to trace lost packages or to process complaints and
the like. A rate of failure of less than one in a thousand is probably above
the break-even point for the firm and far higher than its actual nondelivery
rate. 42 With its sophisticated client base and the competitors snapping at
its heels, the limitations on damages found in the Federal Express agree-
ment cannot be attributed to some mysterious contract of adhesion; after
all, why stop at disclaiming expectation damages if, as dictator, you could
just knock out all liability by dictating contract terms? Instead, the terms
limiting liability are best understood as a way to maximize joint benefits to
the parties ex ante: they minimize the costs of administration and mitiga-
tion that otherwise arise with breach while insuring a certain high level of
contract performance by the firm.
IV. DEFAULT RULES
The provisions found in the full range of voluntary contracts in sales,

employment, and especially in common carrier contexts provide a frame-
work for evaluating the default rules of contract as they have evolved
from that most famous of carrier cases, Hadley v. Baxendale. Thus the
pattern is explicit that often rejects the expectation measure of damages
and its concomitant defenses in favor of far more restrictive measures of
41"In any event, we won't be liable for incidental damages (for example, alternative
carrier transportation costs), consequential damages (for example, loss of profits or income),
or special damages, whether or not we knew that such damages might be incurred" (Federal
Express Terms and Conditions).
42 On-time delivery rate in excess of 99 percent (Federal Express Public Relations office,
Memphis, Tenn.).
recovery. The original Hadley decision reached just that result, and for
much of the nineteenth century, the cases adhered to that result under the
tacit assumption of risk rule. Over time, however, this restrictive inter-
pretation of Hadley has given way to a more expansive view of the plain-
tiff's level of damage recovery. This section traces some high points in
that case development, with special reference to consequential damages
for late delivery under contracts of carriage.
A. Hadley v. Baxendale
As is well known, Hadley involved the defendant carrier, who delayed
the shipment of the plaintiff's millshaft for five days. The breach estab-
lished, the plaintiff claimed recovery for lost profits during the days that
the mill was shut down, but these were denied as a proper measure of
damage, limiting damages (or so it would appear) to a refund of the money
paid to the carrier for shipping the shaft. The familiar rule of decision in
that case contained two prongs. "Where two parties have made a contract
which one of them has broken, the damages which the other party ought
to receive in respect of such breach of contract should be such as may
fairly and reasonably be considered either arising naturally, i.e., accord-
ing to the usual course of things, or such as may reasonably be supposed
to have been in the contemplation of both parties at the4 3time they made
the contract as the probable result of the breach of it."
Both branches of the rule loosely resonate with the general principle of
expectation damages, that a person is responsible for the harm he has
caused. But, clearly, more is at stake in the case, for the next sentence
indicates that where the plaintiff has "knowledge of special circum-
stances," the damages that the defendant can contemplate will take into 44
account only those circumstances "so known and communicated."
Otherwise, only the damages "generally" flowing from the breach are to
be involved. While these rules here look as though they are directed to the
measure of damages, in fact they are not. Later on in his opinion, Baron
Alderson suggests that one function of the communication of circum-
stances is to allow the defendant to insist on a variation of the terms of the
contract as regards the damage issue. "[H]ad the special circumstances
been known, the parties might have specially provided for the breach of
45
contract by special terms as to the damages in that case."
The issue in this case is recurrent across contract law: where does the
43 9 Ex. at 354, 156 Eng. Rep. at 151.

44 id.
45 Id. at 355, 156 Eng. Rep. at 151.
burden of recontracting lie? In this context the narrower question is, Why
is the burden of recontracting placed on the defendant, even after he
receives communication of special information? It is also possible to hold
that the simple communication of notice to a defendant does not increase
the recoverable damages unless there is some clear evidence that the
defendant had agreed to pay the additional damages associated with that
special risk. The explicit disclaimer of all consequential, incidental, and
special damages on the Federal Express standard form shows how notice
often entails the repudiation, not the assumption, of risk. This result
comports with the theory of the "tacit assumption of risk" championed
by both Judge Willes 4 and Justice Holmes 47 in the nineteenth-century
elaboration of the principle.
Theory aside, Hadley is a case whose result is at war with its premises.
To read the decision, one might think that the liability imposed was very
substantial, thus creating a significant cleavage between the standard
terms that the law imposes and those found, for example, in the Federal
Express standard form. But appearances are deceiving, for Hadley, in
fact, imposed sharp restrictions on recovery. Baron Alderson's general
rule seems to make lost profits from delayed shipment of the crankshaft
the "natural consequences" of breach, at least if the only criterion in
question were causation in fact. The earlier delivery would have avoided
the loss that the later delivery did not. A fortiori the defendant seems hard
pressed to take advantage of the special circumstances referred to in the
second limb since the opinion (where it applies the principle) treats the
case on the assumption that the plaintiffs did communicate to the defen-
dant that the item shipped was a broken crankshaft, and that they were
the operators of the mill. Yet the court said that nonetheless the defen-
dants did not have sufficient notice to be on their guard:
But how do these circumstances show reasonably that the profits of the mill must
be stopped by an unreasonable delay in the delivery of the broken shaft by the
carrier to the third person? Suppose the plaintiffs had another shaft in their pos-
session put up or putting up at the time, and that they only wished to send back the
broken shaft to the engineer who made it; it is clear that this would be quite
consistent with the above circumstances, and yet the unreasonable delay in the
delivery would have no effect upon the intermediate profits of the mill. Or, again,
suppose that, at the time of the delivery to the carrier, the machinery of the mill
had been in other respects defective, then, the same results would follow. Here it
is true that the shaft was actually sent back to serve as a model for a new one, and
that the want of a new one was the only cause of the stoppage of the mill, and that
4 British Columbia Saw-Mill Co. v. Nettleship, 3 C.P. 499, 508-509 (1868).

47 Globe Refining Co. v. Landa Cotton Oil Co., 190 U.S. 540 (1903).
the loss of profits really arose from not sending down the new shaft in proper time,
and that this arose from the delay in delivering the broken one to serve as a model.
But it is obvious that, in the great multitude of cases of millers sending off broken
shafts to third persons by a carrier under ordinary circumstances, such conse-
quences would not, in all probability, have occurred; and these special circum-
stances were here never communicated by the plaintiffs to the defendants. It
follows, therefore that the loss of profits here cannot reasonably be considered
such a consequence of the breach of contract as could have been fairly and
reasonably contemplated by both the parties when they made this contract.48
Lost profits were disallowed under the special circumstances test. But
the result involves a sleight of hand, as Baron Alderson adopts the clever
strategy of mentioning two alternative scenarios (the extra crankshaft or
the mill that is otherwise run down) which, if true, imply that a delay in
shipment yields no lost profits to the firm. He hints that other circum-
stances might yield the same result. (The government could have put the
mill under lock and key to prevent civil disorder.) But at no time does he
make anything remotely like a probabilistic assessment of the relative
frequency of the various scenarios to determine that the stated probabili-
ties cover the great multitude of cases. In practice, goods are often ship-
ped posthaste long distances precisely because the want of a single part
keeps an entire plant from operation: hence the importance of the air
freight business. To call what happened in Hadley unforeseeable, or be-
yond the contemplation of the parties, is to make professional busi-
nessmen systematically ignorant of the commonplace. It is more accurate
to say that lost profits from delayed shipment of any equipment are the
natural and common consequences of the breach, even if the shipper
knows nothing about the particulars of any given situation. If anything
remotely like the capacious tort standards of foresight are used, then the
case is quite trivial-the other way. If a stranger had wrecked the crank-
shaft as it lay on the carrier's loading dock, lost profits are a required
element of damages under any reading of the foreseeability test. Foresight
here, like reasonableness in so many quarters of the law, utterly lacks the
descriptive content that allows it to be the principled basis for decision.
What is needed is not verbal jousting but functional accounts. The best
explanation of Hadley v. Baxendale is that the optimum contract struc-
ture need not, for the reasons noted above, use expectation damages as
48 9 Ex. at 355-56, 156 Eng. Rep. at 151. Note that the reference to "both" parties sounds
more restrictive than it is, for the plaintiff will always know those special circumstances.
The question is whether they have been communicated to the defendant whose knowledge is
decisive, as the modern law now holds. "[I]t is foreseeability only by the party in breach that
is determinative. This is now accepted, even though Baron Alderson spoke of the contem-
plation of 'both parties' " (Farnsworth, supra note 21, at 877).
the norm in order to maximize the joint gains of the party. The old,
restrictive tacit assumption of risk test, widely rejected today,4 9 could be
unsympathetically dismissed as resting on hypothetical and fictional con-
tracts. But that objection can be lodged against any test for default provi-
sions: what else can one do except make the best guesses to fit the broad
run of cases? The right question to ask is, Which set of rules is more in
keeping with the logic of mutual benefit that underlies the law of contract
generally, and with the observed limitations in express contractual pro-
visions that can be found today? The idea of "foreseeability" may be
widely accepted, but it is "maddeningly vague." 50 More to the point, it is
wholly unmotivated by any commercial or business consideration.
B. Ordinary Interest as the Expectation Measure of Damages

The total separation of the legal rule from its social function carries with
it substantial costs for legal doctrine. The early nineteenth-century cases
read Hadley as an assumption-of-risk case but could not explain why the
risks assumed were so limited. As the situation emerged, the tension
between the overall foreseeability limitations and the restricted recovery
allowed under the cases became evident. As the justification for the nar-
rower rules never was forthcoming, the content of the rule itself shifted,
leading to the adoption of broader liability in contract for damages that
were "not unlikely" to result from the breach. The upshot of this trans-
formation is that damages for the delayed shipment of goods became
presumptively the difference between market price at the actual and
scheduled time for arrival, even for those cases where the proper use of
the standard expectation measure damages points in favor of a more
limited measure of damages-one equal to interest for the period of delay
on the capital amount of the shipment. The twists and turns in this unfor-
tunate evolution of legal doctrine are well illustrated by two English con-
tract cases-The Parana5 1 and The Heron H1,52 both of which deal with
the liability for late delivery of nonperishable, fungible goods under any
ordinary contract for carriage. The Paranagot the right result, in part for
the wrong reasons. The Heron H, which mechanically applied Hadley, got
the wrong result for the wrong reasons. Both cases require some detailed
analysis.
49 See, for example, Farnsworth, supra note 21, at 875, which approves of abandoning the
test as "overly restrictive and doctrinally unsound." See U.C.C. 2-715 Comment 2, which
rejects this test as well. In neither case are any functional reasons offered for the result.
SO Richard A. Posner, Economic Analysis of Law 115 (3d ed. 1986).
5' 2 P. 118 (1877).

52 3 All Eng. Rep. 686 (1967).
In The Parana,the defendant had taken for shipment cargoes of sugar

and hemp from Ilo Ilo to London. The expected length of time for the
journey was between sixty-five and seventy days, but owing to the defec-
tive condition of the ship, the journey was 127 days. During the journey,
much of the sugar had leaked out of the boat, and the hemp had declined
in price. The plaintiff held the hemp for some time after The Parana
arrived in London and then sold it for a loss of about £289. The registrar
and merchants, following standard business practice, allowed the owner
to recover for the full value of the sugar but not for the losses associated
with the sale of the hemp. The judge of the Admiralty Division restored
the award for the hemp, and that determination was reversed on appeal.
For the delay in the shipment of the hemp, the proper award was the
interest on the principal sum (represented by the value of the hemp) for
the period of the delay.
The decision of the registrar and merchants, as affirmed by the court of
appeals, is correct. The more troublesome question is the right way to
analyze the problem. Here Mellish, L.J., identified some, but not all, of
the relevant principles. His opinion began with a dutiful recitation of the
two key tests in Hadley v. Baxendale, and then argued that the losses
sustained by the plaintiff here were in essence speculative because the
particular circumstances of the contract were not made known to the
carrier. Doctrinally, he held as follows. "In order that damages may be
recovered, we must come to two conclusions-first, that it was reason-
ably certain that the goods would not be sold until they did arrive; and,
secondly, that it was reasonably certain that they would be sold im-
mediately after they arrived, and that that was known to the carrier at the
53
time when the bills of lading were signed."-
Earlier in his opinion, he had distinguished several cases where such
damages for late shipment might have been allowed.
If goods are sent by a carrier to be sold at a particular market; if, for instance,
beasts are sent by railway to be sold at Smithfield, or fish is sent to be sold at
Billingsgate, and, by reason of delay on the part of the carrier, they have not
arrived in time for the market, no doubt damages for the loss of market may be
recovered. So, if goods are sent for the purpose of being sold in a particular season
when they are sold at a higher price than they are at other times, and if, by reason
of breach of contract, they do not arrive in time, damages for loss of market may
be recovered. Or if it is known to both parties that the goods will sell at a better
price if they arrive at one time than if they arrive at a later time, that may be a
ground for giving damages for their arriving too late and selling for their lower
sum. But there is in this case no evidence of anything of that kind. As far as I can
53 Supra note 51, at 123.

discover, it is merely said that when goods arrived in November they were likely
54
to sell for less than if they had arrived in October, for the market was lower.
Mellish then observed that nothing constrained the owner of the hemp
(in fact, a mortgagee of the cargo with rights of owner) from selling the
goods while at sea if he thought that the price was advantageous. The
delay, therefore, did not require him to miss a favorable market if one was
available. Similarly, the decision not to sell the hemp immediately upon
arrival was not mandated by the breach but was an independent decision
based on the owner's estimation of the future movement of the market,
which he happened to misread. The losses were disallowed because it was
too "speculative and uncertain" to calculate when the hemp would have
been sold.
The specific institutional arguments made by Mellish contain a good
deal more force than the bland proposition that the losses in question
were not foreseeable under Hadley. It surely counts as a point in favor of
the carrier that the cargo owner had complete control over the timing of
the sale, but even if that had not been the case, interest on the capital sum
is the correct measure of damages, which corresponds with the expecta-
tion measure of damages. Revert once again to the model of the "single
owner," used to explain why fixed damages give the optimal incentives to
reduce loss given the defendant's breach. The key here is to ask the
maximum sum of money the owner of the cargo would pay in order to
induce the prompt shipment of the hemp to London. That sum of money
would ordinarily be equal to the amount of money that the single owner of
craft and cargo would have spent to avoid the delays in shipment. That
sum is, in fact, interest on the principal amount for the late shipment and
not more-at least absent the exceptional circumstances noted by Mel-
lish, L.J., above.
The conclusion is dependent on the structure of the relevant market.
The London market for hemp was a continuous market where the goods,
themselves nondepreciable, could be disposed of at the time of their
arrival. The present price in effect contains the best collective estimation
of the future price for the goods. The price of hemp could well go down,
but, by the same token, it could also go up. Counsel for the seller grasped
the relevant point intuitively when he noted the "insuperable" difficulties
posed by the loss of market rule for late shipment. He asked, "If sugar
had gone up whilst hemp had gown down, it would be impossible to allow
the shipper to recover for the hemp and not allow for the sugar, and yet
54 Id. at 121.
what is to be done if they belong to different shippers? Is the shipowner to

lose when the prices fall and not to gain when they rise?"" In essence, his
argument is that the uncorrelated price variation between two goods pro-
vides, ex post, compensation for the loss where there is a single owner-
negating his claim for damage payments. Why tolerate the anomaly of
having to pay damages for the downward price movement where the
hemp and sugar are owned by separate parties?
The argument can be taken a step further because it does not depend on
the offsets for two separate quantities being found as a matter of fact ex
post. It is quite sufficient to identify the expected price movements for
each good ex ante. If the best estimates of future prices for this hemp,
itself a durable good, were below the present market price, then the
market would be out of equilibrium so that the current price would
quickly fall, once again making the value of present and future goods
indistinguishable except for the interest component. In this regular world,
the owner of hemp knows that the expected price he will get in the future
is a function of the actual price that he can get today, even if the hemp
could not have been resold while at sea. The only adjustment the owner
need make is to take into account the delay in the receipt of the goods
or cash, and that itself is represented perfectly by an interest payment.
Stated otherwise, the breach by the carrier not only carries with it the
prospect of harm but also the likelihood of benefit, which now functions
as compensation for the breach. Both the anticipated harm and the antici-
pated benefit must be taken into account, as they fall, of necessity, to the
same person, here the owner. Where the distribution of future prices is
normally distributed about the mean figure, then this "in-kind" compen-
sation received in the event of breach exactly offsets the loss that the
breach itself has occasioned, save only for the time difference.
The economic measure of efficiency thus corresponds perfectly with
the "corrective justice" standard, given that the plaintiff's prospect of
gain, when coupled with interest payments, redresses any imbalance that
defendant might have created. In addition, risk aversion itself is not an
important factor in the analysis because there is a necessary dispersion in
future prices, whether the shipment itself takes sixty-five or 120 days. At
most, some allowance would have to be made for any increased disper-
sion of risk owing to the longer time period-a marginal calculation that is
just not worth the candle, given the relatively small amounts of money
involved. It follows, therefore, that if the cargo owner actually purchased
insurance from the shipper for the decline in the market price, he would,
in effect, be entering into a separate futures transaction where he pur-
55 2 P. at 119.
chased a put (that is, an option to sell the goods) for a price equal to that
which the market revealed on the expected time of arrival. There is no
reason for that futures contract to be embedded in a contract for carriage;
it could be as well handled in a separate futures contract or, as Mellish
noted, by resale of the goods while at sea. Ex ante the shipper's only
anticipated loss in The Paranawas the interest attributable to the delay in
shipment. The prospect of loss was, of course, reasonably foreseeable but
so, too, was the equal and opposite prospect for gain.
The situation is, of course, very different in the particular settings
mentioned by Mellish, L.J., as exceptions to the general rule. Suppose
that a late delivery of goods means that the owner will miss an organized
market that does not gather again for another year, and that storage of the
goods in the interim is not possible, owing to the substantial risk of spoil-
age. Now the single owner of both ship and cargo would invest greater
sums of money to see that the cargo arrives in good time, for the cost of a
later shipment is no longer mere delay but necessarily a loss of market,
where no prospect of future gain functions as the appropriate offset. The
case of the seasonal market is functionally equivalent to the case of the
leaky sugar, and so the customary admiralty rules provided. As there is
no compensating benefit from breach, the rules have to be tailored to take
into account the increased costs of delay, just as the single owner of ship
and cargo would now spend greater sums of money to insure earlier
arrivals. Loss of market becomes the proper measure of damages, not
interest. Admiralty practice again made the right distinction. The con-
tract/market rule governs whenever (time lag aside) expected future
prices do not have a distribution that is symmetrical around the current
price-a point that is true in all of Mellish's examples.
The analysis in The Parana identified some of the key elements in
support of the limited damage rule. The decision also gained strength
because it was consistent with the custom of the trade, which moves
strongly to the social optimum, even if the parties themselves do not
understand fully the economic determinants of a sound decision. The
Heron II, decided 90 years later, represents a conscious effort to ra-
tionalize the contract damage rules in accordance with abstract principle
by judges who were wholly ignorant of their internal economic logic. The
result was verbal fencing of no possible utility, the repudiation of The
Paranafor aesthetic reasons, and the adoption of the wrong default provi-
sion for cases of delayed shipment of durable goods.
In The Heron H, the plaintiffs chartered the defendants' vessel to carry
a cargo of 3,000 tons of sugar from Constanza to Basrah, with the option
in the charterer to have the cargo diverted to Jeddah. The option was not
exercised, but the owner of the boat was nine days late in arriving at
Basrah, having made an illegal diversion of the ship. The sugar on board
was immediately sold at a price below what it would have fetched had the
ship arrived on time some nine days before. The market for sugar at
Basrah was continuous, and just before The Heron II reached port, an-
other ship had unloaded a considerable cargo, which had depressed the
price of sugar. If The Parana represented good law, then the proper
measure of damages was the lost interest on the value of the freight for the
nine-day period. But that decision was rejected on the ground that Had-
ley, rightly understood, called for a damage rule that took into account
questions of foreseeability. Even though the remote possibilities of loss
were sufficient to trigger loss in tort, they could not do the same in
contract. Lord Reid tried to identify the right standard in contract cases as
follows:
It appears to me in the ordinary use of language there is a wide gulf between

saying that some event is not unlikely or quite likely to happen and saying merely
that it is a serious possibility, a real danger, or on the cards. Suppose one takes a
well-shuffled pack of cards, it is quite likely or not unlikely that the top card will
prove to be a diamond: the odds are only three to one against; but most people
would not say that it is quite likely to be the nine of diamonds, for the odds are
then fifty-one to one against. On the other hand I think that more people would say
that there is a serious possibility
56
or a real danger of its being turned up first, and,
of course, it is on the cards.
As a matter of ordinary English, there is every reason to doubt that the

probabilities are as Lord Reid would assign them. "Quite likely" might
refer to three-to-one odds against, but to speak of a "serious possibility"
or a "real danger" is to invoke shorter odds than the long shot (as it would
be in racing) odds of fifty-one to one. A serious risk of loss could easily be
10 percent or more in many contexts. But the terminological quibble in
the case is not central to the inquiry.57 What is critical is its economic
logic. Once three-to-one against becomes a case of compensation, then it
follows that lost profits will be necessarily the routine measure of dam-
56 The Heron II, 3 All Eng. Rep. 686, 694-95 (1967).

-7 In this terminological vein, Lord Reid also criticized the test in The Parana, holding
that damages were not recoverable so long as it was not "reasonably certain" that the goods
would be sold while at sea but would be sold immediately upon arrival. "[So strict a test has
long been obsolete; and if one substitutes for 'reasonably certain' the words 'not unlikely' or
some similar words denoting a much smaller degree of probability, then the whole argument
in the judgment collapses" (The Heron II, 3 All Eng. Rep. 696 (1967)). Note that the correct-
ness of The Parana in no way depends on the probability of sale at sea, but on the symmetri-
cal nature of price variations.
ages, where the anticipated distribution of prices at the time of delay is

identical to that which is found at the agreed-on time of arrival. Given the
continuous nature of the Basrah market, the chances are close to 50
percent that the future prices will be lower, far better than the three-to-
one stated odds. The same result follows if the question is, Does the
shipowner know of the possibility of a decline in market prices? Lord
Morris makes the case clear enough. "The appellant shipowner knew that
there was a sugar market at Basrah. When he contracted with the charter-
ers to carry their sugar to Basrah, though he did not know what were their
actual plans, he had all the information to enable him to appreciate that a
delay in arrival might in the ordinary course of things result in their
suffering some loss. He must have known that the price in a market may
fluctuate. He must have known if a price goes down someone whose
goods are late in arrival may be caused loss." 58
Thereafter follows a recitation of the standard expectation measure
formula for damage that leads to the inexorable conclusion that the prima
facie measure of damage is the difference between the market prices at
the scheduled and actual time of arrival. 59 Lord Morris's strategy attrib-
utes to the shipowner perfect information over the relevant variables. In
fact, both shipowner and cargo owner know more and could calculate
better than he assumes. Morris takes the view of Cassandra and dwells on
the downside. He ignores completely what was also common knowledge
to both parties, that the price could increase as well. If, for example, the
rival trader (whose arrival is both random and unknown) had put into port
the week before (instead of the week after) The Heron II was scheduled to
arrive, then the shipper in The Heron I would have benefited from the
delayed arrival. The greater gap between the two arrival dates would have
increased the price The Heron H's sugar would have fetched on its late
arrival. Here the charterer could pocket all the additional profits. The
choice of the wrong measure of damages thus places a false incentive on
the vessel to arrive on time and limits the opportunity to make diversions
that a single owner would otherwise regard it in his interest to make. The
inefficient damage rule thereby increases total costs for both sides to the
arrangement. The logic of the rule depends on the relationship of the price
on the date of scheduled arrival to the full distribution of future prices.
Both the cost and the benefit sides of the equation must be taken into
account. The Parana reached the correct result, even though it did not
understand all the dynamics of trade. The Heron H misfired because it
58 Id. at 701.
59 id.
thought that the first office of the law of contracts was to explicate the
language of Hadley, not to analyze the recurrent business problem that
gave rise to the underlying issue.'
C. UndercompensatingLosses from Delayed Shipment

The Parana and The Heron H are, in my view, relatively easy cases
because the implicit in-kind compensation from breach gives a very sim-
ple explanation as to why interest for delay gives well-nigh perfect com-
pensation to the buyer. In other cases of delay by common carriers, this
condition no longer holds, and now the analysis must revert to the very
different paradigm of Hadley, where some systematic undercompensation
for consequential losses is needed to induce the proper level of loss
minimization. An instructive case is the recent decision of Judge Posner
in Evra Corp. v. Swiss Bank Corp.,6 1 which adds an overlay of privity
problems to the basic question of the proper measure of contract dam-
ages.
In Evra, the plaintiff (Hyman-Michaels Co., whose name was changed
to Evra Corp.) had made a very profitable charter of a ship, where the
contract called for the forfeiture of the charter if the required rental fee
was not paid "in advance" to the shipowner's account in a Swiss bank.
The money in question had to be paid from Hyman-Michaels in Chicago,
via London, to the Swiss bank. Hyman-Michaels used a system of wire
transfers that were initiated at most two days before the payments were
due. On one previous occasion, the charter payments were late, but an
arbitration panel, by divided vote, nonetheless prevented the owner from
cancelling the charter party, given that Hyman-Michaels had taken steps
to expedite the payments once it found out about the mix-up. Six months
later, Hyman-Michaels again tried to wire its payment from Chicago to
London to Switzerland, again less than two days before it was due. But
there was a breakdown in the system of crediting by the Swiss bank, so
6o Lord Upjohn tried to distinguish The Parana from The Heron II by resorting to an
argument of changed conditions. "No doubt in the days of sail pure and simple, when ships
might be delayed by head winds for days, loss of market would not be within the contempla-
tion of the parties. In 1877, when The Paranawas decided, the steam engine was coming
into its own, but it was still the golden age of sail and over half the ships built in this country
were sailing ships at this time; these matters may well have influenced Mellish, L.J., when
he pointed out the differences between delay in delivery by carriers on land and cases of
carriage of goods on long voyages by sea" (id. at 720). But the logic of the expectation
measure damages depends on conditions that were identical in both The Parana and The
Heron II. The argument from changed conditions is used all too often when the underlying
structure of the problem is not understood. See, generally, Richard A. Epstein, The Static
Conception of the Common Law, 9 J. Legal Stud. 253 (1980).
6 See note 29 supra.
that the deposit was not made in time; this time the owner was able to
cancel the charter when arbitrators found that Hyman-Michaels did not
try to remedy the defect with all possible dispatch. The question was
whether the bank, which had been found negligent, could be held for the
increased costs that Hyman-Michaels had to pay to renew its old charter
at the far higher current levels.
Under the Swiss law, the bank could not be held liable to Hyman-
Michaels because the two parties were not in privity. Judge Posner re-
fused to dismiss the case under Swiss law. Rather than resolve the conflict
question, he concluded that Illinois law also denied recovery on the claim,
even though it does not adopt the privity rule. In essence, he argued that,
while this was a tort case, it was also preeminently a botched commercial
transaction, so that the applicable damage principles should be derived
from Hadley v. Baxendale. Accordingly, much of his opinion is devoted
to a dissection of what the bank knew at the time it received the transfer
from Hyman-Michaels and when it knew it. He concluded that the bank
had no special notice of the urgency for payment, or of the catastrophic
consequences of default, and so could not be charged with the loss.
If Posner had stopped here, there would have been little change in the
understanding of Hadley. But he did not. In addition, he tried to link the
remoteness of damage in Hadley to the tort doctrine of "avoidable conse-
quences," which comes into play when the plaintiff has the opportunity to
take steps to avoid an injury, say, by seeking medical attention after an
accident, or by buckling up the seat belt before the accident. Hyman-
Michaels's behavior clearly was a monumental high-stakes blunder, for in
the effort to save a few days of interest on a $27,000 principal payment,
the firm took the substantial risk of losing a favorable charter party worth
in excess of $2,000,000. The preferred pattern of cost avoidance was an
early transfer of the money to the Swiss account. Posner then identifies
62
the many sources of "imprudence" in Hyman-Michaels's behavior.
It is just here that he misfires. The Hadley rule, for all its imperfections,
does obviate the need for an in-depth examination of the plaintiff's con-
duct in this particular transaction. Posner's effort to transform Hadley
into a contributory negligence rule is wholly inconsistent with the facts of
the original case, where the mill owner's conduct was in no way responsi-
ble for the delay in shipment. In addition, as Illinois is now itself a com-
parative negligence state, 63 any use of negligence principles should lead
62 Id. at 955.
63 Alvis v. Ribar, 85 111. 2d 1, 421 N.E.2d 886 (1981). The transaction in Evra took place in
1972, and the trial (it appears) in 1981, so the case itself would not appear to be governed by
comparative negligence principles. But future cases that followed Evra would be decided
under the comparative negligence rules.
not to a denial of the claim, as happened here, but to an apportionment of

loss with the plaintiff recovering perhaps one-half or at least some sub-
stantial fraction of its losses. Yet this elaborate system is inconsistent
with the economic logic of the transaction, which uses the sharp limitation
on damage proper in order to avoid the contributory negligence inquiry,
with its Hand-formula baggage. While in Evra that claim is easy to estab-
lish, other cases can be readily imagined in which it would be more
controversial. Some unanticipated emergency might have required a
speedy transfer, or earlier payments could have been misplaced by the
receiving bank for a longer period of time. In both situations, the contrib-
utory negligence defense could raise dogged legal battles and might prove,
in the end, unavailable to the defendant. The special knowledge rule
avoids all these complications.
One can go further, for here it seems that the Swiss, as is so often the
case in financial matters, were correct in adopting the privity limitation. If
it is unlikely that the bank would have assumed the risk of enormous
losses to a party with whom it was in direct agreement, then there is no
reason for the tort law to allow customers to make an end run round the
web of contractual understandings solely by using an intermediate bank to
implement their transfers. As Posner himself correctly noted, "It seems
odd that the absence of a contract would enlarge rather than limit the
extent of liability." 64 The advantage of privity is that it uses contractual
provisions to govern the full set of liabilities on all the parties. If the Swiss
bank must pay limited damages to its London correspondent bank, then
the Swiss bank has a (limited) financial incentive to comply, even if the
plaintiff himself does not get a refund of the money that he paid out to the
continental bank. It therefore follows that the defendant has some incen-
tive to avoid the loss, just as the plaintiff (who does not have specific
knowledge of the circumstances) has the powerful incentive to set up
some alternative arrangement that guarantees prompt and early payment
to the shipowner. Respecting the privity limitation thereby leads to a
distribution of gains and losses that is far closer to the contractual ideal
than any tort doctrine of judicial invention. The outcome in Evra was
correct, but its reliance on special circumstances, and especially on con-
tributory negligence, was not.
D. Limitations on Expectation Damages outside the Common

CarrierCase
In closing, it is useful to consider one other class of cases that has
caused some difficulty under the damage rule of Hadley. Suppose the
64 673 F.2d at 956.

owner of the plaintiff has entered into some collateral arrangement-for

example, a resale at a favorable price-whereby he can obtain supranor-
mal profits should the defendant perform the contract as desired but not
otherwise. The typical situation often arises where the plaintiff is fortu-
nate enough to enjoy some quasi-monopolistic situation that offers the
prospect of a supercompetitive return or a very profitable resale. In Brit-
ish Columbia Saw-Mills v. Nettleship,6 5 the plaintiff sought to construct a
new sawmill in the Pacific Northwest from which the returns might have
been very high if it were the first mill in operation in the new location. In
Victoria Laundry (Windsor), Ltd. v. Newman Indus., Ltd.,66 the plaintiff
enjoyed "especially lucrative" contracts for laundry service in virtue of
being of the first to reenter the laundry business at the end of the Second
World War when demand was well-nigh insatiable. In each case, the
expectation is valuable to the plaintiff, although valuation itself is com-
plicated by its very high variance.
These cases, however, are in their business aspects sharply distin-
guishable from, and more difficult than, both The Paranaand The Heron
I1. In Nettleship and Victoria Laundries, the payment of interest no
longer renders the plaintiff indifferent between breach and performance,
as the expectation standard itself requires. The net anticipated losses to a
single owner would have been greater than interest, so that some addi-
tional precautions from prompt arrival would be taken. Given that these
are hard cases, the optimal strategy for the plaintiff to protect these gains
is to bargain for a higher damage award (coupled with a higher base price)
in the event of breach. The difficulty in valuation and the size of the stakes
increase the deadweight losses to both sides. It could well be that more
sufficient monitoring and cooperation on a day-to-day basis yields a more
sensible approach to contract damages. In the alternative, a specific pro-
vision for liquidated damages tied to the length of the delay could allow
the defendant to calculate its own potential exposure, without having to
make a detailed estimate of the plaintiff's future losses-which could well
be influenced not only by the defendant's breach, but by many other
factors as well, such as the conduct of other independent suppliers. The
usual standard of the difference between contract and market, or the cost
of replacement, minimizes many of these difficulties and might well be
preferred by the parties themselves.
The objection, therefore, to awarding special gains from unknown ar-
rangements does not stem from their want of foresight; a problem so
recurrent in the academic literature could hardly be unknown to ordinary
business firms. The answer is, rather, that it is not clear that a default
65 3 C.P. 499 (1868).

6 2 K.B. 528, 537 (1949).
provision that holds simply that these special damages are recoverable is
better than one that studiously refuses to take them into account. The fit
between any default rule and the contractual ideal will be worse when
extraordinary profits are at issue than it is when they are not. But the only
lesson that can be drawn from that melancholy observation is that the
gains for contractual individuation are greater for cases that lie at the tail
of the distribution than for those located close to its mean. Holding that
these losses are not foreseeable, or are beyond the contemplation of both
parties or the party in breach, is a tiny part of the complete understanding
of the situation.
The point can be briefly shown by looking in detail at one such case
involving the lost profits on resale. In re R & H. Hall, & W. H. Pim
(Junior) & Co.'s Arbitration,67 the original buyer purchased an un-
specified quantity of wheat from the seller at a given price. The buyers
then resold the rights under contract to a subbuyer at a higher price, after
which the contract right was resold again at a still higher price. Subse-
quently, the cargo from a certain ship was identified as being the subject
matter of the contract, in accordance with standard industry practice. The
buyer, once notified, immediately informed his own subbuyer. The mar-
ket price then fell when the designated ship arrived in port, but the seller
still failed to deliver. For its admitted breach of contract, the seller was
held liable to its own buyer, but the question was whether it should have
to compensate its buyer for its losses in failing to complete its own sub-
contract and any subsequent losses that the buyer might face in the un-
likely event it was sued by that subbuyer. Both elements of damages were
allowed because, looked at from the time of the original sale, it was rated
an "even chance," hence not unlikely, that the wheat would be resold
before its delivery in port, and that was enough to bring the buyer's case
within the ambit of the second branch of the Hadley rule.6 8
The result here may or may not have been right, but the correct answer
does not depend solely on the odds of the resale, which must surely have
been high, or simply on the movement of the market. The right question
to ask is whether the price paid by the buyer included within it compen-
sation at this raised measure, given the breach. It is probably of some
interest that the contract was for the sale of a specific cargo, which meant
that the buyer could not substitute wheat from any other vessel to per-
form his own contract obligation. The true winner, therefore, seems to
be the subbuyer who is let out from a losing contract. Given, therefore,
67 All Eng. Rep. 763 (1928).

' See, for example, id. at 767, per Viscount Dunedin, whose language was later approved
in The Heron II. 3 All Eng. Rep. 686, 693 (1967), per Lord Reid.
that the buyer has no other way to lock his subbuyer in, the court's award
of the special damages seems to make a good deal of sense. The result
looks even stronger since it seems likely that the seller's failure to deliver
(while not discussed in the case) may have arisen because he had resold
the same cargo to a second purchaser and thereby captured the increase
in price that his first buyer should have enjoyed. On this view, it could
well be that the seller is only made, roughly speaking, to disgorge the
profits from the alternative transaction. The result might have been ren-
dered more precise if the profits from resale had been turned over to the
buyer, itself a kind of restitution measure that is generally precluded by
cases holding that ideal equivalence demands a uniform expectation mea-
sure of damages.6 9 The use of the restitution measure necessarily strips
the seller of all gains from breach and thus tends to secure compliance and
thereby reduce the frequency and importance of suits .70 Again, the theory
of contract damages must address both the defendant's gain and the plain-
tiff's loss. By leaving out one side of the equation totally, the court in re
Hall may have come to the right result, but surely for the wrong reason,
when it relied on the standard of equal likelihood.
V. CONCLUSION
This article has sought to demonstrate that the question of contract

damages can be examined by the same techniques that are ordinarily used
to handle other issues of contract interpretation. Where there are express
provisions that govern the award of damages, these should normally be
followed, unless one can show the same form of fraud, duress, or incom-
petence sufficient to set aside the basic obligations under the same con-
tract. Where the express provisions of the contract are incomplete or
entirely omitted, then the best that can ever be done is to develop those
rules which, as a general matter, appear to work to the joint benefit of the
parties at the time of contract formation. In working through this analysis,
it often occurs that the proper measure of damage is far lower than that
which appears to be implied by the now-dominant expectation measure of
damages, which provides that the plaintiff is to be left after breach in the
same position that he would have enjoyed had the defendant performed in
full. In some instances, contract damages are reduced to reflect the need
69 See, for example, Acme Mills & Elevator Co. v. Johnson, 141 Ky. 718, 133 S.W. 784
(1911), where, however, the court held that the contract was not for the sale of particular
wheat.
70 See Richard A. Epstein, Inducement of Breach of Contract as a Problem of Ostensible
Ownership, 16 J. Legal Stud. 1 (1987); Daniel Friedmann, The Fallacy of Efficient Breach,
18 J. Legal Stud., in this issue.
to control plaintiff's misconduct and to economize on litigation and settle-

ment costs. In other cases, these are reduced in order to prevent the
cross-subsidization of some plaintiffs by others. In still other instances,
damages are lower because the defendant's breach confers on the plaintiff
some implicit compensating benefit that reduces the need for cash com-
pensation.
Once these basic patterns are understood, it follows that there will be
good reason to reorient basic judicial attitudes toward contract damages.
First, it seems clear that the verbal formulations of Hadley v. Baxendale
and its progeny, which emphasize either the likelihood of damage on
breach or the defendant's knowledge of special circumstances, should be
set aside in favor of the nineteenth-century view that depended on a
theory of tacit assumption of risk, which, in fact, can be made far more
precise and powerful than has been generally understood. Second, the
theory of contract damages has important implications for the techniques
of contract construction. The general view today is that limitations on
damages ought generally to be narrowly construed, in large measure be-
cause they offend the expectation ideal.7" Yet if these contracts work ex
ante to the benefit of both sides, this strategy of interpretation contra
proferentem will only favor the few lucky plaintiffs who maintain action
for breach while hurting plaintiffs and defendants as a class. The tempta-
tion to do justice in the individual case should, therefore, be curbed in
order to advance the long-term welfare of contracting parties generally. A
simpler rule of ordinary meaning is a more reliable guide to contract
construction. Similarly, there should be greater acceptance of freedom of
contract as a general principle, as many academics now advocate.7 2 In
general, the repudiation of explicit contract provisions on grounds of
public policy typically follows a detailed judicial examination of the con-
tract provisions. The object of that demonstration is to show that the
plaintiff is not perfectly compensated in the event of breach.7 3 Yet once it
is realized that perfect compensation for damages ex post may yield a
systematic increase in overall damages, then this temptation, too, should
be avoided. The traditional view of contracts was that custom and com-
mon practice were generally sound. Repeat players do not make mistakes
on provisions so critical to their personal welfare.
7' Farnsworth, supra note 21, at § 12.18.

72 See, for example, Schwartz, supra note 1. See also Goetz & Scott, supra note 10.
73 The most dramatic illustration is still Henningsen v. Bloomfield Motors, Inc., 32 N.J.
358, 161 A.2d 69 (1960), which ushered in the age of strict liability in tort for product cases
by its merciless dissection of the limited warranty provided by Chrysler. The court, how-
ever, ignored all the problems with plaintiff's conduct, cross-subsidization of risk, and
administrative expense that now haunt so much of modern product liability law.

18 JLegal Stud 105

Uploaded by

18 JLegal Stud 105

Uploaded by

DATE DOWNLOADED: Fri Aug 4 13:38:39 2023

SOURCE: Content Downloaded from HeinOnline

Bluebook 21st ed.

ALWD 7th ed.

APA 7th ed.

Chicago 17th ed.

McGill Guide 9th ed.

AGLC 4th ed.

MLA 9th ed.

OSCOLA 4th ed.

* James Parker Hall Distinguished Service Professor of Law, University of Chicago.

issue of contracting out would be resolved if raised expressly. Samuel

choice of proper remedial rules is largely a judicial function. Yet as a

II. FREEDOM OF CONTRACT AND THE MEASURE OF DAMAGES

The provisions of a standard contract are not an arbitrary assemblage of

9 9 Ex. 341, 156 Eng. Rep. 145 (1854).

quest for contract theories of greater abstraction. If the contract measure

depended heavily on the identity of the parties to the transaction. Mat-

for . ..errors in cipher or obscure messages." The contract therefore

III. DAMAGE SOLUTIONS

In order to understand how damages are selected, it is critical to aban-

A. Seller's Remedies: Nonpayment of Price and Nonacceptance

mitigation eliminates the possibility of enormous consequential damages.

sure of damages is far from self-evident. Whatever the abstract law of

duct, whether under a misuse or a mitigation doctrine, is both costly and

tion measure of damages suggests that compensation should be provided

system, are often substantial, especially where the employer is in an

limitations with boldface disclaimers of all liability for incidental, conse-

IV. DEFAULT RULES

The provisions found in the full range of voluntary contracts in sales,

43 9 Ex. at 354, 156 Eng. Rep. at 151.

4 British Columbia Saw-Mill Co. v. Nettleship, 3 C.P. 499, 508-509 (1868).

B. Ordinary Interest as the Expectation Measure of Damages

5' 2 P. 118 (1877).

In The Parana,the defendant had taken for shipment cargoes of sugar

53 Supra note 51, at 123.

what is to be done if they belong to different shippers? Is the shipowner to

It appears to me in the ordinary use of language there is a wide gulf between

As a matter of ordinary English, there is every reason to doubt that the

56 The Heron II, 3 All Eng. Rep. 686, 694-95 (1967).

ages, where the anticipated distribution of prices at the time of delay is

C. UndercompensatingLosses from Delayed Shipment

not to a denial of the claim, as happened here, but to an apportionment of

D. Limitations on Expectation Damages outside the Common

64 673 F.2d at 956.

owner of the plaintiff has entered into some collateral arrangement-for

65 3 C.P. 499 (1868).

67 All Eng. Rep. 763 (1928).

This article has sought to demonstrate that the question of contract

to control plaintiff's misconduct and to economize on litigation and settle-

7' Farnsworth, supra note 21, at § 12.18.

You might also like